AIED 2009: 14th

International

Conference on Artificial

Intelligence in Education

Workshops Proceedings

Editors and Co-Chairs:

Scotty D. Craig University of Memphis, USA

Darina Dicheva Winston-Salem State University, USA

July 6-7th, 2009

Brighton, UK

ii

Preface

The supplementary proceedings of the workshops held in conjunction with AIED 2009,

the fourteen International Conference on Artificial Intelligence in Education, July 6-7,

2009, Brighton, UK, are organized as a set of volumes - a separate one for each

workshop.

The set contains the proceedings of the following workshops:

• Volume 1: The 2nd Workshop on Question Generation Co-chairs: Vasile Rus & James Lester. University of Memphis, USA & North

Carolina State University, USA.

http://www.questiongeneration.org/AIED2009/

• Volume 2: SWEL'09: Ontologies and Social Semantic Web for Intelligent

Educational Systems Co-chairs: Niels Pinkwart, Darina Dicheva & Riichiro Mizoguchi. Clausthal

University of Technology, Germany; Winston-Salem State University, USA &

University of Osaka, Japan.

http://compsci.wssu.edu/iis/swel/SWEL09/index.html

• Volume 3: Intelligent Educational Games Co-chairs: H. Chad Lane, Amy Ogan & Valerie Shute. University of Southern

California, USA; Carnegie Mellon University, USA & Florida State

University, USA.

http://projects.ict.usc.edu/aied09-edgames/

• Volume 4: Scalability Issues in AIED Co-chairs: Lewis Johnson & Kurt VanLehn. Alelo, Inc., USA & Arizona State

University, USA.

http://alelo.com/aied2009/workshop.html

• Volume 5: Closing the Affective Loop in Intelligent Learning

Environments Co-chairs: Cristina Conati & Antonija Mitrovic. University of British

Columbia, Canada & University of Canterbury, New Zealand.

http://aspire.cosc.canterbury.ac.nz/AffectLoop.html

• Volume 6: Second Workshop on Culturally-Aware Tutoring Systems

(CATS2009): Socio-Cultural Issues in Artificial Intelligence in Education Co-chairs: Emmanuel G. Blanchard, H. Chad Lane & Danièle Allard. McGill

University, Canada; University of Southern California, USA & Dalhousie

University, Canada.

http://www.iro.umontreal.ca/~blanchae/CATS2009/

iii

• Volume 7: Enabling Creative Learning Design: How HCI, User Modelling and Human Factors Help Co-chairs: George Magoulas, Diana Laurillard, Kyparisia Papanikolaou &

Maria Grigoriadou. Birkbeck College, University of London, UK; Institute of

Education, UK; School of Pedagogical and Technological Education, Athens,

Greece & University of Athens, Greece.

https://sites.google.com/a/lkl.ac.uk/learning-design-workshop/Home

• Volume 8: Towards User Modeling and Adaptive Systems for All

(TUMAS-A 2009): Modeling and Evaluation of Accessible Intelligent

Learning Systems Co-chairs: Jesus G. Boticario, Olga C. Santos and Jorge Couchet, Ramon

Fabregat, Silvia Baldiris & German Moreno. Spanish National University for

Distance Education, Spain & Universitat de Girona, Spain.

https://adenu.ia.uned.es/web/es/projects/tumas-a/2009

• Volume 9: Intelligent Support for Exploratory Environments (ISEE’09) Co-chairs: Manolis Mavrikis, Sergio Gutierrez-Santos & Paul Mulholland.

London Knowledge Lab, Institute of Education/Birkbeck College, University

of London, UK & Knowledge Media Institute and Centre for Research in

Computing, The Open University, UK.

http://link.lkl.ac.uk/isee-aied09

• Volume 10: Natural Language Processing in Support of Learning:

Metrics, Feedback and Connectivity Co-chairs: Philippe Dessus, Stefan Trausan-Matu, Peter van Rosmalen &

Fridolin Wild. Grenoble University, France; Politehnica University of

Bucharest; Open University of the Netherlands & Vienna University of

Economics and Business Administration, Austria.

http://webu2.upmf-grenoble.fr/sciedu/nlpsl/

While the main conference program presents an overview of the latest mature work in

the field, the AIED2009 workshops are designed to provide an opportunity for in-depth

discussion of current and emerging topics of interest to the AIED community. The

workshops are intended to provide an informal interactive setting for participants to

address current technical and research issues related to the area of Artificial

Intelligence in Education and to present, discuss, and explore their new ideas and work

in progress.

All workshop papers have been reviewed by committees of leading international

researchers. We would like to thank each of the workshop organizers, including the

program committees and additional reviewers for their efforts in the preparation and

organization of the workshops.

July, 2009

Scotty D. Craig and Darina Dicheva

iv

v

AIED 2009 Workshops Proceedings

Volume 6

Second Workshop on Culturally-Aware

Tutoring Systems (CATS2009)

Workshop Co-Chairs:

Emmanuel G. Blanchard McGill University, Canada

H. Chad Lane University of Southern California, USA

Danièle Allard Dalhousie University, Canada

http://www.iro.umontreal.ca/~blanchae/CATS2009

vi

Preface

Learners with various cultural profiles and backgrounds are benefiting from diverse

applications and initiatives of the Artificial Intelligence in Education (AIED)

community. That said, research in education has shown that teaching methodologies

and instructional design cannot always be universally applied as their impact can

greatly vary from one culture to another. In other words, pedagogical strategies that are

effective for one cultural group may not be effective with a different one. As a result,

researchers are increasingly focusing on cultural factors in the conception of AIED

systems, be it from the point of view of learners’ varying cultural backgrounds or the

influence that these backgrounds have on the choice of underlying teaching

methodologies. Developing AIED systems with cultural discernment capabilities can

therefore -- among other possibilities -- contribute to lessening the potential for

misunderstanding learner behaviours as well as allow for customized learning

according to cultural needs. A greater cultural focus can also increase the flexibility of

the systems we build and promote their acceptance and wider spread use. Furthermore,

in a world in which interactions between culturally diverse people and groups are

becoming usual, there is a need to learn about culture itself, with its rich and multi-

faceted variability.

A first edition of the workshop took place in 2008 in conjunction with ITS 2008

(Intelligent Tutoring Systems). This second edition further aims to examine ways in

which culture can be represented within the overarching goal of imparting knowledge

via AIED systems. It wishes to stimulate discussion on the impact of culture on such

systems and reflect on those emerging technologies that need to be developed to more

fully integrate cultural considerations. Finally, it aims at sharing and expanding the

very knowledge we have about culture.

We are most grateful to the many individuals who have made this half-day

workshop possible. We thank the Program Committees of the International Conference

on Artificial Intelligence in Education, especially workshop chairs Scotty Craig and

Darina Dicheva for their advice and collaboration in the planning of the workshop. We

wholeheartedly thank the members of the CATS 2009 Program Committee for having

dedicated time to evaluate workshop submissions within a limited time frame. We

finally thank Dr. Erica Melis for having kindly accepted to give the opening keynote

presentation.

We are very pleased that this second workshop on Culturally-Aware Tutoring

Systems has been well-received by the AIED/ITS community as well as by researchers

exploring related eLearning issues. We believe that intelligent systems can propose

viable solutions to help deal with various intercultural challenges and we hope that this

second workshop is a reflection of how this theme is increasingly becoming a

recognized field of research.

Welcome to CATS 2009, the second workshop on Culturally-Aware Tutoring

Systems.

July, 2009

Emmanuel G. Blanchard, H. Chad Lane & Danièle Allard

vii

Program Committee

Co-Chair: Emmanuel G. Blanchard, University of Montréal, Canada

Co-Chair: H. Chad Lane, University of Southern California, USA

Co-Chair: Danièle Allard, Dalhousie University, Canada

Lora Aroyo, Free University Amsterdam, The Netherlands

Jacqueline Bourdeau, TELUQ, Canada

Elisabeth Delozanne, Paris VI University, France

Benedict Du Boulay, University of Sussex, UK

Claude Frasson, University of Montreal, Canada

Guy Gouardères, University of Pau, France

Monique Grandbastien, University of Nancy, France

W. Lewis Johnson, Alelo Inc, USA

Judy Kay, University of Sidney, Australia

Susanne P. Lajoie, McGill University, Canada

Riichiro Mizoguchi, Osaka University, Japan

Roger Nkambou, University of Québec at Montréal, Canada

Amy Ogan, Carnegie Mellon University, USA

Elaine Raybourn, University of New Mexico, USA

Matthias Rehm, Augsburg University, Germany

Katharina Reinecke, University of Zurich, Switzerland

Isabelle Savard, TELUQ, Canada

Robert Wray, SOAR Technology, USA

viii

Table of Contents

Keynote Presentation

Culturally Aware Mathematic Education 1

Erica Melis

Full Papers

Toward Ontological Modeling of a Communication Exchange Model 3

for Use in Computer Assisted Language Learning (CALL)

Danièle Allard, Riichiro Mizoguchi, Jacqueline Bourdeau

and Emmanuel G. Blanchard

Learner-Concerned AIED Systems: Affective Implications when Promoting 13

Cultural Awareness

Emmanuel G. Blanchard and Susanne Lajoie

Virtual Environments for Cultural Learning 25

H. Chad Lane and Amy Ogan

Pedagogical Experience Manipulation for Cultural Learning 35

Robert Wray, H. Chad Lane, Brian Stensrud, Mark Core, Laura Hamel

And Eric Forbell

The Relevance of the Culture-Based Model in Designing Culturally Aware 45

Tutoring Systems

Patricia A. Young

Opening Keynote

1

Culturally Aware Mathematics Education

Erica MELIS

German Research Institute for Artificial Intelligence (DFKI)

66123 Saarbruecken, Germany, melis@dfki.de

Abstract. For the enculturation of the European platform for mathematics learning,

ACTIVEMATH, a number of dimensions had to be adapted culturally: presentation

of system and learning material, terminology, selection and sequencing of learning

objects, interaction, and learning scenarios.

Enculturation is not only important for tutoring systems that teach cultural relation-

ship, geography, culture, and behaving appropriately in other countries, etc. such as the

cultural training system Alelo but also for mathematics learning environments. At first,

mathematics education may seem to be an unlikely field for enculturation. However,

• mathematics education often follows the terminology (notions) of a particu-

lar book dominant in the region’s education or of a field that applies mathe-

matics, e.g., electrical engineering

• in the region’s or field’s mathematics education more often than not specific

notations is used

• educationalists and teachers emphasize the cultural context and for teaching

mathematics they follow the curriculum pre-scribed for their region and type

of school

• educational psychologists and pedagogists emphasize the importance of

posing authentic mathematical problems which are relevant in the student’s

life environment. This aspect is typical for a more constructivist learning

paradigm [5,1].

Theoretical Background. In addition, at the level of cognitive processes Piaget’s the-

ory of learning implies that learning heavily depends on the context and experience of

the learner. Piaget’s influencial theory of learning introduces two fundamental cognitive

functional processes in learning: assimilation and accomodation [3,4]. Assimilation and

accommodation are two complementary processes of adaptation in the learner’s mind

through which awareness of the outside world is internalised. In assimilation, what is

perceived in the outside world (in the context) is incorporated into his/her mind with-

out changing its structure. In accomodation, the mind accommodates itself to the evi-

dence/context with which it is confronted and thus adapts to it.

An immediate implication of this theory is that the learner’s context including its

cultural determination is reflected in the learning process and influences its adjustment

efforts, which require more effort the more the external (context) and internal (mind)

features differ. The mind has to adjust (1) its mathematical structure and (2) its contextual

features (for mathematics rather superficial). Both types of adjustment yield cognitive

load, and the cognitive load for (2) is extraneous and may hinder learning.

2

It has also be defined which (different levels of) cultural groups are relevant for en-

culturations in (mathematics) learning in order to represent them in the student’s pro-

file. It turns out that it is not only the region/country/language that requires enculturation

but also communities of practice (CoP), which can be a group inside another or a group

orthogonal to a region/country/language group, e.g., chemistry or electrical engineering

students. Wenger [6] acknowledges that (learning) practices are influenced by context(s)

and that the internal dynamics of a CoP including its norms (part of a culture) are de-

termined by the practices such as ’negotiation of meaning’ (i.e., understanding of math-

ematical notions and notations), ’learning’, as well as ’community actions’ (e.g., book

publishing, conferences), and ’differentiation from other communities of practice’. Any

CoP produces abstractions, symbols, stories, terms, and concepts that reify something of

its practice in a congealed form.

Note that even cultural variations whose nature appears to be linguistic such as nota-

tions and notions have a cultural (often historical) background. Reasons for such differ-

ences can originate from a dominant group of mathematicians in history and cultural ties

to that group in certain countries, regions, and groups. A similar development (maybe not

yet for mathematics) is observable today for communities of practice, e.g. users of email

or SMS. For cultural differences present in language-defined communities. One reason

may be that in mathematics the history of schools, research, teacher education, popular

books which influence the education culture go back to a language-defined culture that

was present before todays countries existed.

Adaptation Dimensions. Our research has empirically collected, addressed, and real-

ized the needed adaptation techniques and developed the underlying architecture and

knowledge representations needed for enculturation. We consider cultural adaptations of

• mathematical notations

• mathematical notions

• input of formulaæ

• choices of learning objects

• sequencing of learning objects and curriculum-dependent learning paths

• last but not least, language-adapted user interface.

For more details see [2]

References

[1] C. Twomey Fosnot. Constructivism: A psychological theory of learning. In C. T. Fosnot, editor, Con-

structivism: Theory, Perspectives, and Practice, pages 8–31. Teachers College Press, New York, Lon-

don, 1996.

[2] E. Melis and G. Goguadze and P. Libbrecht and C. Ullrich. Culturally Adapted Mathematics Education

with ActiveMath. Artificial Intelligence and Society, Special Issue on Enculturating HCI, 2009.

[3] J. Piaget. La représentation du monde chez l’enfant. Presses Universitaires de France, réimpression

quadrige, 2005 edition, 1947.

[4] J. Piaget. Equilibration of Cognitive Structures. Viking, New York, 1977.

[5] D. Schifter. A constructivist perspective on teaching and learning mathematics. In C. T. Fosnot, editor,

Constructivism: Theory, Perspectives, and Practice, pages 73–80. Teachers College Press, New York,

London, 1996.

[6] E. Wenger. Communities of Practice: Learning, Meaning, and Identiy. Cambrigdge University Press,

1999.

Full Papers

3

Towards Ontological Modeling of a

Communication Exchange Model for Use

in Computer-Assisted Language Learning

(CALL)

Danièle Allard1, Jacqueline Bourdeau

2, Riichiro Mizoguchi

3, Emmanuel G.

Blanchard4

1French Department, Dalhousie University,

6135 University Ave., Halifax, NS, B3H 4P9, Canada. d.allard@dal.ca

2 LICEF, Télé-université,

100 Sherbrooke O., Montreal, QC, H2X 3P2, Canada. jacqueline.bourdeau@licef.ca

3 Institute of Scientific and Industrial Research (I.S.I.R.), Osaka University,

8-1 Mihogaoka, Ibaraki, Osaka, 567-0047 miz@ei.sanken.osaka-u.ac.jp

4 ATLAS Laboratory, McGill University,

3700 McTavish Street, Montreal, QC, H3A 1Y2, Canada emmanuel.blanchard@mcgill.ca

Abstract: In this paper, we present preliminary findings as to the modeling of

a communication exchange process, using ontological engineering

methodology. Our model is being developed for use in the context of acquiring

a second/foreign language, with a focus on native language interference due to

cultural reasons, in conversation. It could potentially prove useful for other

endeavors in which communication and cultural differences are at play.

Keywords: ontological engineering, upper ontology, communication, culture,

learning

1 Introduction

To use a language as a means of expression is to communicate. As such, it is

inextricably linked to culture in multiple and complex ways. When people

communicate using language they do so in relation to one another and in terms of

prior knowledge and experience. Their voice is not only individual, but also

collective, a reflection, among others, of attitudes, beliefs, perceived social

expectations and patterns.

The acquisition of another language implies not only becoming familiar with a

different linguistic system, but also a cultural reality. By its very nature, in a threefold

4

manner, language actually expresses, embodies and symbolizes cultural reality [1].

Our research is concerned with making aspects of this cultural reality apparent while

focusing on the issue of cross-linguistic influence, which is essentially the act of using

skills stemming from one’s native language when communicating in another. When

these skills are used in ways that lead to errors and possible misunderstanding, it is

called cross-linguistic influence, transfer or interference [2][3]. For the sake of

simplicity, we use the term interference in this article.

We have identified several situations in which interference takes place and have

closely analyzed some of these in the context of conversation so as to make apparent

the linguistic and cultural considerations that explain how and why they might occur

[4][5]. The goal is to identify particular language learning difficulties, and in

understanding how they occur, design teaching strategies to help overcome the

difficulties, that can then be made available via computer-assisted language learning

systems.

We have been mapping this knowledge using ontological engineering

methodology. Until recently, we proceeded by examining interference by focusing on

how a given interference manifested through language, and the cultural considerations

that could explain it. That said, it became apparent that aspects of a communication

process such as context, the relative status of participants, the strategies they use to

achieve a given communication goal, also need, among others, to be closely examined

to more fully grasp certain cases of interference. We have consequently begun

preliminary work on modeling a relatively generic communication process that would

allow us to consider interferences within the context of a communication exchange

between two parties, especially conversation.

In this paper, we will briefly describe the modeling of a communication exchange

between two parties of possibly different cultural backgrounds in terms of ontological

engineering. The seminal aspects of our work, while developed for use in the context

of language acquisition and conversation, may prove valuable to other research in

which cultural differences in the context of a communication exchange is to be

modeled for use in intelligent computer systems.

We begin by a definition of ontology, and show an excerpt of an initial effort at

modeling interference. We show how this initial model can benefit from being

expanded and present preliminary modeling of an essential communication exchange

from which we can ultimately proceed to a finer analysis of interference, so as to use

this knowledge in computer-assisted language applications. In the process, we

highlight further research that needs to be conducted.

2. Ontology

Ontology is a term that was first used in philosophy to describe the study of being or

existence. Broadly put, it is a theory of the nature of existence. In computer and

information science, to quote excerpts of explanations by Gruber, “an ontology

defines a set of representational primitives with which to model a domain of

knowledge or discourse. The representational primitives are typically classes (or

sets), attributes (or properties), and relationships (or relations among class members)

[…]. Ontologies are typically specified in languages that allow abstraction away from

5

data structures and implementation strategies; in practice, the language of ontologies

are closer in expressive power to first-order logic than languages used to model

databases. For this reason, ontologies are said to be at the “semantic” level, whereas

database schema are models of data at the “logical” or “physical” level […];

ontologies are called out as an explicit layer [6].

More simply put, an ontology is a specification of a conceptualization. A

conceptualization is an abstract, simplified view of the world that we wish to

represent for some purpose, and the specification takes the form of the definitions of

representational vocabulary (classes, relations, etc.), which provide meanings for the

vocabulary and formal constraints on its coherent use [7].

Ontologies are designed for an agent or a community of agents so that knowledge

can be shared with and among these agents. Agents can be computer systems, and

humans, since ontologies also use human-readable text.

The Standard Upper Ontology Working Group states that “an ontology is similar to

a dictionary or a glossary, but with greater detail and structure that enables computers

to process its content. An ontology consists of a set of concepts, axioms, and

relationships that describe a domain of interest” [8].

There are several ontology building methodologies. Our own work is grounded in

ontology building as set forth by Mizoguchi Laboratory. Some of the reasons for

doing so is that, as Mizoguchi explains in relation to the YATO Upper Ontology, this

methodology carefully investigates quality description and the issue of representation,

and it differentiates between objects, processes and events; it is furthermore the only

initiative based on a model of roles [9] [10].

3. Initial modeling of interference

We initially analyzed interferences related to cultural differences in terms of how they

manifest at the language level in view of the cultural reasons underlying them, using

common forms of errors made by Japanese students learning English. These were

circumscribed using reports from seasoned instructors of language, as well as

published literature on the topic. The aim has been to formulate teaching strategies

based on our findings, and make these available via CALL systems. Furthermore, an

important goal is to make language instructors aware of the nature of interferences

and their causes. Instructors of English who do not know Japanese or are not familiar

with the culture are not likely to adopt teaching strategies to help prevent and

overcome interferences. Similarly, Japanese instructors of English, who often do not

master the English language nor have spent much time outside Japan, are not

necessarily equipped to help prevent interferences. To address interferences can be

especially valuable in view of preventing language patterns to become fossilized, in

other words, become hard to correct as time passes [2].

Figure 1 shows an excerpt of an example of interference. In this particular case,

we were examining the topic of suggestions and advice. To summarize, when giving

advice, interference can manifest in the use Japanese students make of modals and

imperatives. The underlying cultural reasons for this particular type of interference

are, among others, issues of tolerance (or intolerance) to advice, as well as factors

related to living in a collective society, in contrast to a more individualistic one. To

6

give a specific example, consider the sentence “You had better go to Osaka Castle to

see the beautiful cherry blossoms,” meant to be a friendly form of suggestion. To a

native English speaker who may have just arrived in Japan, however, this can

potentially be jarring, especially with the effect of articulation – a Japanese native

may stress “had better.” The choice of the modal (had better) is somewhat

inappropriate – “could” would be a better choice, if using modals. Part of the reason

for using “had better” over “could” lies in how dictionaries translate modals from

Japanese into English. The Japanese expression that translates as “had better”

implies, in Japanese, strong advice, concern for the hearer’s welfare, and highlighting

the possibility of a missed opportunity (in this case, it would be unfortunate not to see

the blossoms at the castle). Since “had better” is used in English for strong advice

and the possibility of a negative consequence, the translation does make sense. Part

of the issue revolves around what “strong” advice implies, itself related to cultural

differences. Simply put, in Japan, concern for the group’s welfare over the individual

is essentially more prevalent than in Canada or the United States, for instance. As

such, behaving according to expectations and codes of prescribed behavior in Japan is

especially important to help maintain group harmony. This results in “strong advice”

being rather common. As a consequence, Japanese people, comparatively speaking,

appear to be more accustomed to various forms of “strong advice,” and consequently

more tolerant to receiving it. Though in the more individualistic countries of Canada

and the USA there are also have expected codes of behavior, advice, if it strikes one

as “unsollicited,” has greater potential of not being accepted as readily as it probably

would in Japan (for a more detailed analysis, see [ 4]).

Figure 1, shows an essential ontology excerpt of what has just been explained. p/o

stands for “part of,” a/o stands for “attribute of.” The writing above the empty

squares indicate a role, that in turn is constrained within a certain class (often the

same as the role), indicated in the square right of the role. What is not shown in the

figure is that in this ontology, different types of interference come under a concept

called “L1-related interference” (native language related), itself falling under “English

language learning topics” We have classified interferences in terms of a general

English language curriculum for first/second year university Japanese students (not

specializing in language). Ideally, an instructor inputting a given topic of instruction

into a CALL system could then not only access teaching strategies, but be made

aware of possible interference in view of topics to be approached in the classroom,

more specifically its manifestation (in this case, modals), as well as cultural factors at

play. The instructor could also be directed to more detailed explanation as to the

nature and reason for the interference, in terms of linguistic and cultural

considerations. Mapping knowledge using ontological engineering makes such

retrieval possible, while allowing the knowledge to be accessed in different systems.

While this initial modeling allows to draw out essential elements related to a given

form of interference, it became apparent that it was also not quite elaborate enough to

take into account other relevant factors. For instance, when giving advice, the context

in which the advice is given can be of impact, as can the perceived social distance

between hearer and speaker. It seemed increasingly relevant to set the analysis of

interference in the context of a greater communication process, and for this purpose,

to elaborate a generic model of a communication exchange, that would highlight other

relevant elements in addition to language and cultural manifestations.

7

Figure 1: Modeling cultural differences in relation to interference when making suggestions/giving advice

4. Preliminary Modeling of a Communication Exchange

The aim of analyzing interference within the context of a communication exchange,

and modeling the process, is to make explicit some of the key elements to be taken

into consideration, as well as to ground the analysis of cultural differences

underscoring interference in terms of the various factors potentially at play.

Following are some of the main issues for which the model should be able to account.

We acknowledge that communication is a complex, scientific field of inquiry in and

of itself, and is further actively researched within various other disciplines. Our lens

is in view of the goals of our research, and what is summarized in the following table,

8

though inspired from essential elements derived from communication studies [11], is

not exhaustive; it is also simplified in terms of labeling, to ease understanding.

Table 1: Main issues our preliminary communication model should address

Participants hearer and speaker; native language spoken and language being

learned; cultural profile; hearer/speaker interpretation and

generation of meaning; other relevant characteristics pertaining

to participants

Extrinsic factors time, location, context

Communication

goal

goal (implicit/explicit)

Com. structure structure (macro and micro); turns, topics and sub-topics;

sequencing and nature of utterances (eg. a question)

Com. function Illocutionary acts (what one does when communicating to

another; eg. compliment); perlocutionary response

(acknowledgment of the impact of the illocutionary act)

(micro and macro levels)

Com. medium the medium (direct speech; email etc.)

Locus of

interference

potential interference as a result of cross-linguistic influence,

its manifestation and reasons (cultural and other)

We have begun work on a preliminary model, shown in Figure 2. We are

attempting to capture, in essence, how a communication exchange can be understood,

identifying the major constituents to be considered in the process, in view of our

research goals and ontological engineering methodology. The model can then serve

as a framework for more specific models to be elaborated for situations in which

language interference can potentially occur between speakers of different native

languages that are underscored by various cultural factors. It is therefore the

reflection of the core of a communication exchange between two parties, providing

upper-level concepts and showing the relations among them. This model inherits

characteristics from higher level concepts defined elsewhere in the ontology, which,

for the sake of clarity, we do not expand upon.

We now proceed to explaining the components of our preliminary model. In the

future it will need to be further elaborated and put to the test with various examples of

interference; it also needs to be compared to, and possibly enriched by, existing

communication models of conversation and language exchange, namely those

developed by research in natural language processing and computational linguistics.

9

Figure 2: Preliminary upper-level model of a communication exchange for use in modeling language

interference

10

Time, location and context are first listed as constituents to the process. Speaker

and hearer each have a native language and a cultural profile, which are key elements

of our research. Linked to the cultural profile are what we label “cultural

interpretation-generation rules.” We hypothesize that both the speaker and hearer are

subject to such rules, which impact the communication process between them.

Should we wish to eventually take individual differences into consideration, we could

add an “individual mental profile” as a characteristic of speaker and hearer, with its

own set of interpretation-generation rules. For the purposes of examining

interferences underscored by cultural differences, however, the concept of cultural

interpretation-generation is currently sufficient.

Relations (the lines linking concepts seen on the right of the figure) show how the

speaker is the generator of a “representation,” a generic upper ontology label that

allows us to account for various forms of expression; YATO, Mizoguchi’s Upper

Ontology, defines this as a “content-bearing thing” [9]. “Representation” is

characterized by both “form” and “content.” We could eventually consider physical

gestures under “representation,” if deemed necessary. “Representation” is considered

to be an output of the speaker, as a result of “cultural interpretation-generation rules,”

shown by the “interpret” relation, with the “content” of the intended communication

being the input.

“content” is represented twice in the model, and is in fact the same (represented by

the relation “same as”). In addition to being a part of “representation,” content is an

integral part of a communication process in and of itself; it implies the function of

communication, or, to put it another way, the type of communication act the speaker

intends to perform (ex.: to give advice). Similarly, “form,” also a part of (p/o)

“representation,” is the “same as” “form,” a component of the communication

process. Relations show how the “representation” is “received by” the hearer, who

“interprets” it in view of his own “cultural interpretation-generation rules,” with, as an

output, a “result of understanding,” specified as a separate label, again to account for

it within the communication process.

The “goal of communication,” for its part, represents the overarching goal, which

can be implicit and/or explicit. Possible work in this area may be to draw out intention

hierarchies (a simple example of this would be how inviting someone for dinner

(explicit goal) can also be to express interest in the other (implicit goal).

The concept “way to communicate” represents the medium of communication. To

this effect, further work in terms of an in-depth analysis of ways in which to

communicate, and their relevant sub-parts, needs to be conducted. Ways of

communication should be approached by drawing out decomposition trees, in an

attempt to highlight granularity inherent in these ways. To do so will provide handles

with which to zoom in on the alternative ways that allow a given communication to be

expressed. To briefly illustrate: a way to communicate can be a “language way,” or,

for the sake of comparison, a “gesture way.” The “language way” can be achieved by

“delivering sentences.” Delivering sentences implies, among others, building content,

composing sentences, ordering them. These may vary in terms of a “face-to-face

way,” a “one-to-many way,” “an email way.” In other words, in our ontological

modeling, we will elaborate on two core aspects of the communication process: what

the process entails and is intended to produce, and how it is carried out. In other

words, “way to communicate” (how), though inherent to a communication process,

11

will also be modeled separately. Without separating what from how, it becomes

difficult to analyze and systemize various ways in which a single what can be

achieved. More specifically, language interference does not express itself

consistently; written and oral form, to be basic, will bring about different forms of

interference. Though we are currently focusing on conversation, we also want our

model to eventually allow for an analysis of interference in other exchange contexts.

A powerful example of how what and how viewpoints work together, in addition to

what decomposition entails, is that of the Omnibus project, focusing on theory-

awareness in Instructional Design. [12].

Cultural analysis underpinning the phenomena of interference will be further

performed, ontologically, under “cultural profile” and “cultural interpretation-

generation rules.” “cultural profile” will be defined within the greater ontology we

are building. To this effect, the work of Hofstede, GLOBE and Schwartz on cultural

dimensions and universal values contribute to shaping this definition while providing

some validated “strength factors” (a given dimension, or sub-dimension, is

stronger/weaker in one culture in comparison to another; for example collectivism has

a higher score in Japan than in Canada) [13] [14] [15]. Cultural expectations and

behavior in terms of a specific communication can be detailed under “cultural

interpretation- generation rules.” The “form” of communication – what is actually

being expressed, is where language interference is detailed, in other words, under

which our previous ontological modeling, as shown in Figure 1, can essentially be

unfolded. The class constraint for the manifestation of interference has been

temporarily labeled “content_2” to differentiate it from “content” (the content of a

communication). “content_2” is a wide-meaning class constraint to indicate that

manifestations of interference are apparent in various forms.

Finally, the hearer’s “result of understanding” can be contrasted to the speaker’s

intended content, and thus allow us to explain the possible impact of an interference.

This can be informed by the detailing of what takes place in the communication

dialogue and exchange, which would be expanded under “com dialogue/exchange.”

More specifically, it implies examining a conversation in terms of its structure, at both

a macro and micro level, namely in terms of turn-taking and sequencing, the nature of

utterances, in conjunction with topics and sub-topics of an exchange, given the

language being used. Further work as to how ontologically represent knowledge we

have of a communication dialogue/exchange is required, and analyzing models from

other fields of research, such as NLP, will prove useful.

To briefly compare with Fig. 1, this preliminary communication model allows us to

specify native language, cultural profiles and cultural expectations in terms of hearer

and speaker, which the previous model did not. While it addresses the manifestation

of interference, in view of language related issues and cultural differences, it further

allows the consideration of the time, location and context of utterances. Implicit and

explicit communication goals can be underscored, as can different ways in which to

communicate – which can affect the nature of the interference. In highlighting these

additional factors, as well as accounting for possible results of understanding on the

part of the hearer, this model accounts for a more comprehensive understanding of

interference in view of various relevant factors, while allowing greater

contextualization of how cultural differences interplay.

12

4. Conclusion

The preliminary communication model we have presented reflects a work in progress,

whose goal is to provide a reasonable ontological engineering approximation of a

communication exchange that would allow for the examination of language

interference within this process when involving participants of different cultural

backgrounds. It has been elaborated in view of analyzing conversation, yet remains

sufficiently generic so as to be able to apply it to other contexts relevant to language

acquisition, for instance written communication. While an examination of cultural

factors at play in view of any given interference can lead to significant understanding

of this interference, which we initially modeled, specific aspects of the

communication process -- the goal, the context, the medium, among others – can

further shed light on factors explaining why certain interferences take place.

Consequently, working at the elaboration of a communication model that would take

these various factors into account appeared necessary. As it is refined, compared to

other models, and further developed, namely in granularity, this model should prove

useful to our research on language interferences whose aim is to provide, within

CALL systems, teaching/learning strategies that help overcome interferences. The

model may also be useful to other research endeavors that examine cultural

differences in the context of a communication dialogue/exchange.

References

[1] Kramsch, C.: Language and Culture. Oxford UP, Oxford (1998).

[2] ZhaoHong, H., Ed.: Fossilization in adult second language acquisition. Multilingual Matters,

Clevedon (UK) (2004).

[3] Arabski, J., Ed.: Cross-linguistic Influences in the Second Language Lexicon. Multilingual Matters,

Clevedon (UK) (2006).

[4] Allard, D., Bourdeau, J., Mizoguchi, R. (2008) Towards Modeling Knowledge of Cultural

Differences and Cross-Linguistic Influence in Computer-Assisted Language Learning (CALL).

Proceedings of the Culturally Aware Tutoring Systems Workshop, held in conjunction with ITS

2008. http://www.iro.umontreal.ca/~blanchae/CATS2008/

[5] Allard, D., Mizoguchi, R., Bourdeau, J. Can Language and Culture Go Hand in Hand? Cross-

Linguistic Influence in the L2 Acquisition Process. Proceedings of the JALT Conference, October

2005. pp. 382-393. JALT, Tokyo (2005)

[6] Gruber, T.: Ontology. http://tomgruber.org/writing/ontology-definition-2007.htm

[7] Gruber, T.: “What is an ontology?” http://www-ksl.stanford.edu/kst/what-is-an-ontology.html [8] IEEE Standard Upper Ontology Working Group. http://suo.ieee.org/

[9] Mizoguchi, R. Yet Another Top-level Ontology: YATO. http://www.ei.sanken.osaka-

u.ac.jp/hozo/onto_library/upperOnto.htm

[10] Mizoguchi, R., Sunagawa, E., Kozaki, K., Kitamura, Y. (2007). A model of roles within an ontology

development tool: HOZO. Journal of Applied Ontology, 2 (2), pp. 159-179

[11] Smith, Mary John. Contemporary communication research methods. Wadsworth, Belmont (1988).

[12] Omnibus project: an ontology of learning, instruction and instructional design.

http://edont.qee.jp/omnibus/doku.php

[13] Hofstede, G.: Cultures consequences, 2nd edition. Sage, Thousand Oaks CA (2001).

[14] House, R.J., Hanges, P.J., Javidan, M., Dorfman, P., Gupta, V. (2004). Culture, leadership and

organizations: The Globe study of 62 societies. Thousand Oaks, Ca: Sage Publications.

[15] Schwartz, S.: A theory of cultural values and some implications for work. Applied pychology: An

international review, 48 (1). pp. 23-47 (1999).

13

Learner-Concerned AIED Systems:

Affective Implications when Promoting

Cultural Awareness

Emmanuel G. Blanchard, Susanne P. Lajoie

Advanced Technologies for Learning in Authentic Settings Laboratory, Department of

Educational and Counselling Psychology,

McGill University, Canada

{emmanuel.blanchard, Susanne.lajoie}@mcgill.ca

Abstract. There is a growing consensus among Information Technology scholars

that cultural considerations can promote the quality of interaction with users, as

well as provide new kinds of services. This paper explores how issues of cultural

awareness and affect can be jointly incorporated into AIED systems. Our analysis

reveals several elements where affective issues will necessarily emerge from

cultural considerations, as well as affective aspects that need cultural awareness in

order to be processed properly.

Keywords. Cultural adaptation, cultural variability of affects, cognition, human

computer interaction.

1. Introduction

There is a growing consensus among Information Technology scholars that cultural

considerations can promote the quality of interaction with users, as well as provide new

kinds of services. In the domain of Artificial Intelligence in Education (AIED),

research has started to address these areas in two ways: (1) to develop educational

systems that promote learners’ intercultural understanding [14, 18, 23], and (2) systems

that have the ability to adapt to cultural differences in order to enhance educational

efficiency [2, 28].

AIED researchers who are developing automatic culturally adaptive systems take

two main approaches. One approach is to design the most pedagogically relevant

resources based on cultural specificity as well as the domain knowledge that needs to

be taught and learned. However, research has shown that local educational techniques

are frequently the most efficient ones, whereas approaches related to a non local culture

may lead to misinterpretations even if the latter technique is more endorsed around the

world. For instance, Biggs [1] has shown that “Hong Kong”-based approaches led to a

higher rate of success with Hong Kong Chinese students than classic Commonwealth-

spread educational strategies. Similarly, “culturally-loaded” resources selected for

illustrating courses in another cultural context may not convey the intended

information,. This is one of the reasons why UNESCO presses the use of local

educational contents to promote the quality of intercultural education [33].

14

The second approach is to avoid culturally inappropriate elements that could

negatively affect learners’ perception of the activity. Indeed, systems that are perceived

as caring about their students’ specificities and success, are likely to be better rated by

learners [5]. For the purpose of this paper we call these systems “learner-concerned”.

Several theories stress the relationship between personal appraisals of an activity and

achievement motivation towards it [9]. Hence it is likely that perception of a system as

learner-concerned could raise learner’s willingness to perseverate or to get involved in

future similar learning activities.

This paper explores how issues of cultural and affective awareness can be jointly

incorporated into AIED research designs. First we define affective and cultural

concepts that pertain to this paper and then we examine intra and inter individual

culture-related affective variations. without omitting to consider the impact of living

conditions. After this review we discuss potential implications of cultural variations of

the affective domain for the development of AIED systems with cultural awareness.

2. Definitions

One shared problem of the affective and cultural domains is the ambiguity of the

language used to describe related concepts. The subjective nature of folk terminology is

not sufficient for scientific discussion where agreeing on a corpus of terms and their

clear definitions is necessary. The following section provides the definitions we refer to

in the continuation of this paper.

2.1. The affective domain

Our conceptualization of the affective domain stems from the meta-analysis performed

by Klaus R. Scherer [30] (table 1).

Table 1. Scherer’s list of identified affective phenomena and their respective definitions (deduced from [26])

AFFECTIVE

PHENOMENA

DEFINITIONS

Emotions - "An episode of interrelated, synchronized changes in the states of all or most of the five organismic

subsystems in response to the evaluation of an external or internal stimulus event as relevant to

major concerns of the organism"

- Scherer emphasizes the distinction between emotions i.e. “the total multimodal component

process” and feelings i.e. “a single component [of any affective phenomena] denoting the subjective

experience process”

Moods - "Diffuse affect states, characterized by a relative enduring predominance of certain types of

subjective feelings that affect the experience and behaviour of a person"

- “Often emerge without apparent causes”

- “Generally of low intensity”

Preferences - "Relatively stable evaluative judgments in the sense of liking or disliking a stimulus, or preferring it

or not over other objects or stimuli"

Attitudes - "Relatively enduring beliefs and predispositions towards specific objects"

- "Can be labeled with terms such as hating, valuing or desiring"

Affect

Dispositions

- "Tendency of a person to experience certain moods more frequently or to be prone to react with

certain types of emotions"

- Examples: trait anger, trait positive/negative affect

Interpersonal

Stance

- "Affective style that spontaneously develops, or is strategically employed in the interaction with the

person or a group of persons"

- “Examples: being polite, distant, cold, warm, supportive, contemptuous"

- “Often triggered by events (encounters of a person), but less shaped by spontaneous appraisal than

by affect dispositions, interpersonal attitudes, and most importantly strategic intentions”

15

As depicted in Table 1, Scherer’s framework of the affective domain consists of

classes of multi-componential processes, called affects or affective phenomena, which

have cognitive, bio-physiological and behavioural functions, and react sequentially

and/or in parallel [30]. Scherer distinguishes six classes of affective phenomena,

namely Emotion, Mood, Preference, Attitude, Affect Disposition (or trait-like affect

[31]), and Interpersonal Stance.

For a discussion of our use of this framework in computational research and AIED,

please refer to [5].

2.2. The cultural domain

Several disciplines have studied the influence of culture and have led to different

conceptualizations [7]. Comparisons between culture-related research are weakened by

the variety of framework used in cross-cultural studies: most of the time,

interoperability between those framework was never considered and is thus very low.

To deal with this issue, an upper ontology of culture is being developed [3, 4]. This

previous work aims at identifying major constituents to be considered when dealing

with any kind of cultural issue by eliciting their structure and interrelations without

having to endorse a particular culture representation framework. In this paper, we refer

to some of these elicited concepts that we now briefly introduce.

Culture can be defined by the cultural elements that the “owner” of the culture

(its related cultural group) has produced, or endorsed and adapted (through historical

processes such as conquests). Such cultural elements can be artefacts (agent-made

elements), practices (such as rituals, languages, common behaviours), ideational

elements (such as norms, scientific knowledge), and specific world representations

that emerge mainly at the group level (such as national stereotypes). This cultural

definition stems from the bi-faceted definition of culture provided by Kashima [15] as

(a) “a process of production and reproduction of meanings in particular actors’

concrete practices (or actions or activities) in particular contexts in time and space”,

and (b) “a relatively stable system of shared meanings, a repository of meaningful

symbols, which provides structure to experience”

A culture has to be distinguished from its cultural group (the population of related

cultured agents) even if the existence of one necessarily implies the existence of the

other. All human beings are indeed cultured agents in the sense that they grew in

cultural contexts, which led their behaviours and cognitions to reflect one or more

cultural influences (i.e. to make it simple, one can see a cultural profile as a patchwork

of cultural influences).

Some ideational elements and specific world representations are frequently

referred to as shared cognition, and describe information emerging at group level, that

is shared by group members even if each individual member does not necessarily

handle the whole “package” [29]. The latter is a major issue when trying to relate

group-level findings to individual cognitions and tendencies. Shared cognition

incorporates the notion of “core cultural ideas” [21] as “culturally defined and

promulgated issues of concern”.

A cultural conceptualization [29] designs an atomic part of a shared cognition.

For instance, the mentorship relation occurring between a senpai (an advanced student)

and a kohai (a beginner student) is an example of a japanese conceptualization of social

relations that may develop between students.

16

The cultural load refers to the specific cultural conceptualizations associated to a

cultural element. Appreciation of an element’s cultural load may change from one

individual to another, their respective cultural background making different

conceptualizations more salient than others. For instance, “Bérézina”, which is the

name of a russian river, is naturally associated to the catastrophic napoleonian retreat of

Russia in France, and is nowadays used accordingly in French folk language to refer to

negative events, which is obviously not the case in Russia, where it refers mainly to the

river.

Cultures may emerge from several kinds of cultural groups such as civic nations

(countries), ethnic nations, populations living in a specific area (or city), companies,

communities, groups of individuals sharing something (such as specific interests, belief

systems, mid-to-long term experiences, or practices). As mentioned earlier, cultured

agents are likely to be part of several kinds of groups, and as such, have blended

cultural influences (without underestimating the importance of personal factors), which

makes the development of artificial cultural awareness even more challenging.

3. Motivation for considering both affect and culture in AIED: A medical scenario

Development of cognitive tools for medical education is an important domain of

application for AIED research. Bioworld [16, 17] is an example of such systems that

helps medical students to improve their diagnostic skills. Bioworld has been designed

in a Canadian institution and, as such, is strongly influenced by the way medical care is

practiced and taught in Canada, and by classic interactions with Canadian patients.

However, should we wish to export Bioworld to other countries, cultural issues or

country-specific elements should be considered such as what medical procedures,

diagnostic tests, drugs and equipment are available based on the medical infrastructure

available. Furthermore, there may be different general approaches to medical education

from one nation to another or from one teaching hospital to another that need to be

considered. Although medical schools generally cover the same content, there may be

more attention on diseases that are prevalent due to conditions i.e., climate, specific to

their region. Hence, culturally specific issues would need to be considered in deploying

Bioworld to a new cultural environment that considers the needs and learning habits of

local students.

We are currently investigating studies that would expand the use and design of

Bioworld by considering issues pertaining to affect management. Physicians need to

manage their own affect as well as that of others. For example, giving bad news to a

patient is difficult for both the provider and receiver of such information. One might

consider affect management on the part of the physician as being a component of

emotional intelligence that medical students need to acquire in their role as healers [8].

Once again, emotional intelligence and affect management should consider the cultural

aspects of the specific situation. Canada is a multicultural society that encompasses a

multitude of communities of people who have originated from various parts of the

world. This intercultural setting is itself an opportunity for affective situations to arise.

Cross-cultural communication problems are obvious causes for the deterioration of an

affective climate. For instance, patients from other cultures may have religious beliefs

that influence social behaviours that the physician would need to be cognisant of so

that perceptions of appropriate physician-patient contact remain positive. A certain

level of cultural intelligence (ability to deal with people from different cultures) is thus

17

needed to facilitate communication and to avoid difficult affective situations that

physicians may face daily. The patient may also need assistance in learning how to

interact with physicians, knowing what is expected of them in terms of answering

personal questions about their conditions. Both physician and patient education that

illustrates proper cultural communication and interaction could enhance the quality of

the diagnosis and treatment.

Raising the cultural awareness of Bioworld would finally be particularly relevant

to target physicians that have to be deployed in foreign countries (cooperating with

organizations such as the Red Cross or Medecins without Borders for instances). In the

next section, we review more thoroughly several situations where issues pertaining to

both culture and affect are relevant.

4. Culture-related affective variations

Mesquita and her colleagues [21] have noticed in the literature on affect and culture

that affective antecedents, subjective experience, appraisals, behavioural responses, and

even physiological changes related to an affective experience are reported to differ

across cultures. Some affective phenomena also appear to occur more frequently in

some cultural groups than in others. Such findings led them to conclude that it has been

“convincingly demonstrated that there are cultural differences in the ecology of

emotions”. The following section briefly introduces some major considerations related

to affective variations depending on cultural settings. Research discussed below

expresses differences in behavioral and cognitive tendencies from one cultural group to

another. We are not positing a simplistic transcription of these results to the individual

level. However these results could inform the design of stochastic systems with

multicultural affective management capabilities.

4.1. Living condition considerations

Even if scholars agree that affective phenomena are in-context processes [21], this

contextual nature has been given little attention in affect-related AIED research [24].

In this section we define living conditions as elements of the surrounding world of an

individual world that may have influences on him or her.

Specific living conditions raise the probability of confrontations to certain events,

objects, or activities, making local individuals more familiar to them than other human

beings. Cultural load for these elements may thus emerge from these frequent

confrontations. Appraisal-focused theories stipulate that the evaluation of an element,

in terms of its “significance for the goals, needs, and well being of the organism is a

critical determinant in the elicitation and differentiation of its emotions” [31].

Furthermore, it is more and more acknowledged that appraisal of novelty is a

fundamental operation in the whole process of elements evaluation. Novelty can refer

to (a) “suddenness or abruptness of onset, often coupled with high stimulation

intensity”, (b) “(un)familiarity with the object or event, generally based on schema

matching”, and (c) “(un)predictability as based on past observations of regularities

and probabilities” [31].

As said before, habituation to specific living conditions influences the degrees of

familiarity of individuals to cultural elements, as well as their ability to predict

evolutions of such elements. As such it may produce major changes of appraisals,

18

leading to cultural variations in affective reactions to stimuli from an individual to

another.

Numerous studies have also shown that the cultural profile of individuals tends to

correlate with the frequency of daily-life positive or negative affective experiences

(mainly mood experiences)). For instance, lower levels of life satisfaction, lower

frequency of pleasant affects, and greater frequency of negative affects tend to be

reported by individuals from Asian cultures compared to North Americans [32]. This

can be understood as cultural variations in affect dispositions of individuals [31].

4.2. Intra-individual considerations

Internal differentiations have emerged from social and genetic evolution and can also

partly explain cultural variations in affect disposition of individuals. This can also be

seen as the result of pressures exerted by living conditions over a certain period of time

(thus this section is closely tied with the previous one). Indeed, research has shown the

existence of differences in physiological reactions between distinct populations and

cultural groups [21]. For instance, the stereotype of “reputedly ‘hot blooded’” south

Europeans is somewhat supported by reports indicating more blood pressure changes,

and less stomach sensations and muscles symptoms as reactions to the experience of

several basic emotions than northern Europeans [25]. The specific genetic evolution of

this human population may be a potential explanation for this.

There are also several cues that suggest that affect-related cognitive differences

exist between individuals from different cultural groups because, among other things,

their mind is full of cultural conceptualizations about their surrounding world that

directly impact on the cognitive components of affective phenomena. As expressed by

Mesquita et al [21], “an emotionology is a culture’s lore about emotions. Supposedly

included in these emotionologies are rules about how to feel”. For instance, it has been

reported that Hispanics and European American “consider pleasant feelings to be much

more desirable and appropriate than unpleasant feelings”, whereas “Asian cultures

emphasize pleasant and unpleasant feelings nearly equally” [32].

Cognitive differences can also emerge as different processing tendencies. Indeed,

valence, desirability, and subjective distance between affective phenomena are subject

to changes according to concerns inspired by core cultural ideas [21] For instance,

“Indians are more likely to view shame and happiness as more similar than shame and

anger, because anger is an emotion that is divisive and separates people” [19, 32].

Certain affective antecedents are also more likely to lead to belief changes in some

cultures than in others, due to variations of appraisal and consideration of cultural

conceptualizations [20]. Finally, the importance of appraisal dimensions [16], as well

as strategies used for memory recall of affective experiences [32] may also vary across

cultural groups.

4.3. Inter-individual considerations

Mesquita [20] sees affective phenomena “as cultural practices that promote important

cultural ideas”. Among the important cultural conceptualizations that differentiate

cultures from each other are the questions of the role of the individual in the society, as

well as norms and standards he/she will have to rely on. Representations of such

concerns have been of major importance in many frameworks that try to categorize

cultures. For instance, in Hofstede’s framework [12] of national cultures dimensions,

19

three dimensions (individualism/collectivism, masculinity/feminity, uncertainty

avoidance) have implications in the expression of such different conceptual

orientations between cultures (similar dimensions have been proposed in other studies).

Indeed, shared cognition varies between cultural groups, obedience and tradition

being salient in some cultures, whereas autonomy and originality are of main

importance in others [21]. Scherer and Brosch [31] have noted that the “focus of a

society on certain shared values is related to emotional experiences to a considerable

degree, particularly for certain emotions that depend on the compatibility of behaviour

with external and internal standards such as shame and guilt”. For instance,

considering differences between collectivist and individualist contexts, Mesquita [20]

has elicited several major affective differences between those two social orientations

(table 2).

To summarize table 2, meanings given to affective experiences vary according to

the social orientations of cultural contexts with affects in collectivist contexts being

characterized as relational whereas in individualist contexts, affective phenomena are

mainly perceived as intrapersonal and subjective. Furthermore, in collectivist contexts,

the affective aspect of an event tends to emerge not only from implications for the

individual, but also from its impact on the various social relationships he/she has [20].

Table 2. Key differences between individualist and collectivist affects (as mentioned in [20])

COMPONENTS OF AFFECTS COLLECTIVIST AFFECT INDIVIDUALIST AFFECT

Concerns One’s own social worth and the

worth of the in-group.

Focus on personal concerns only.

Appraisal Attention to the impact of other

people’s behaviour on relative

social positions (intentionality).

Less focus on the impact on

relative social positions.

Source of Appraisal The meaning of situations

appears as given (obvious).

Awareness of the subjectivity of

emotional appraisals.

Action readiness Focus on relationships and

therefore more action readiness.

Focus on bounded self and

therefore less action readiness.

Nature of shared emotions Social sharing involves ensuring

that others share in the concern

and behave accordingly.

Social sharing involves the

sharing of information.

Emotions as meanings Emotions signal changes of

reality: changes of beliefs about

self, others, and the relationship

between self and others.

Emotions signals internal,

subjective feelings: few

implications for beliefs.

Inter-individuals communication favour cultural differences. Even when different

cultures share a similar verbal language at first sight (for instance, French is the official

language of several countries), cultural load of some terms can induce variability, some

terms being familiar only to a specific cultural group, whereas others have a varying

meaning from one cultural group to another. Improper use of culturally loaded terms in

an inter-individual communication can thus raise misunderstandings and affect

comprehension, which can lead to affective reactions.

Such risks become even worse when languages are different because terms that are

translated do not always convey a similar meaning. Affect terms are a documented

example of such a thing. Even if scholars have determined sets of basic emotions from

the similarities found in emotion categories across different cultures and languages

[26], research also has shown that “there are no universal emotion concepts lexicalised

in all languages of the world” [21, 34], and that “not only may the meaning of lexically

equivalent words differ to some extent, but emotion taxonomies of given languages

20

contain emotion words that, in other languages, are not matched by even remotely

similar emotion concepts” [11, 21].

Awareness of proper social standards and communication norms (or lack of it) is

another important factor that induces affective experiences. Politeness theory [6] for

instance mentions the existence of positive and negative faces, where faces are standard

attitudes that promote positive or negative self esteem in private and public situations

and can be conveyed by many culturally-variable means such as verbal or body

languages,. Similarly, in global business, endorsement of the local etiquette and

business manners [13] can have an important role in determining the success or failure

of the interaction.

Finally, awareness of cultural stereotypes as well as the ability to properly manage

them in varying cultural settings is another important skill when someone wishes to

maintain positive interpersonal stances.

5. Potential implications for AIED

As previous mentioned findings demonstrate, caution must prevail when considering

affects in AIED systems since affective management may vary according to cultural

settings. Affect-related measurements and strategies have thus to be confirmed in the

cultural setting in which learning technologies will be used. But existing research

initiatives mixing affect and culture could also inform the design of culturally aware

AIED systems in many ways.

5.1. Informing the proper use of affective elements in AIED systems

Affective expressions of pedagogical agents should get inspiration from research

on body language and its variation in different cultural settings. Cultural desirability of

affective phenomena has to be considered before deciding to use affective induction

techniques. For instance, pride may be a desirable affective phenomenon in certain

cultural groups, but it should be avoided in others. Affect-related interaction modes

could lead to belief changes in some cultural groups more than in others. For example,

valence appraisal of a pedagogical agent may be more subject to change after it

reproached a learner [20].

Cultural awareness could also enhance team creation and management in

collaborative learning, by ensuring that all members will properly understand

implications of affective antecedents that may arise for their peers. Considerations of

cultural conceptualizations endorsed by potential members could also help to avoid

conflicting situations, or optimize chances of optimal critical thinking exchanges within

a team, or in a social learning setting

5.2. Improving student modelling.

Learner models should reflect the cultural specificities of learners. It should

convey information on which cultural settings the learner has become familiar with

during his/her personal history. Furthermore it should reveal the learners’ strength of

habituation to those settings in order to suggest (1) their level of habituation to cultural

elements, (2) the spectrum of cultural load that can be detected on such elements and

the affective reactions it may induce, as well as (3) the norms that might influence

21

human-human and human-computer interaction in the context of technology-enhanced

learning.

The development of the representation of the shared cognition of a cultural group

in addition to the individual learning model has to be explored [2]. The use of the data

grid paradigm and of technologies related to the notion of cognitive emergence are

promising fields of investigation for such purpose.

Linking shared cognition to an individual cognitive model of pedagogical agents is

also an interesting issue to address.

5.3. Promoting the positive perception of AIED systems

Lastly, improving cultural awareness should promote the transmission of a positive

ethos to its learners by facilitating the design of systems that are respectful of cultural

norms and standards. Ethos can be defined as the feeling of credibility transmitted by a

speaker to his audience, and has been extended to Human-Computer Interaction: the

computer’s “voice” has to be perceived as credible by the user [22]. Projected ethos is

of central importance for a speaker (here an AIED system) that wants its message to be

conveyed and endorsed by its audience (its learners), and consequently must meet

cultural norms or standards. Negative ethos results in the emergence of negative

affective attitudes towards the system that may for instance be considered as useless for

learning achievement, careless of learners’ success, or trustless in terms of the

reliability of the information it conveys. Negative ethos could result in a dramatic

decrease in learners’ motivation to perseverate in a technology-enhanced learning

activity, or to be involved in similar ones in the future. On the other hand, positive

ethos would lead to positive affective attitudes towards the system, by raising the

learners’ motivation to focus on systems’ messages, as well as the congruent perception

of system’s use with learners’ needs and the achievement of personal goals, thus

raising learners’ level of intrinsic motivation towards the activity [27].

6. Conclusion

In this paper, we focused on the enhancement of learning by suggesting ways to adapt

AIED to be sensitive to affective and cultural issues that influence learning. This

endeavour led us to posit the interconnections between affect and culture, affective

phenomena being subject to cultural variations, whereas cultural elements are

frequently eliciting affective reactions. This important fact has been neglected in AIED

research until now.

Indeed, there are numerous challenges induced by the highly complex task of

addressing those two ill-defined domains simultaneously to consider the interplay

between them on respective issues. We think that a precise formalism would allow the

integration and interoperability of research findings from both domains, and would thus

be of great interest to technology enhanced learning. Hence we are now integrating

affective considerations in the upper ontology of culture we are developing.

In this paper, we tried to show that a proper consideration of affect in AIED

necessary requires a better understanding of several cultural aspects. Similarly,

considering cultural adaptation without clearly addressing affective aspects will be a

non sense in many situations.

22

In future works, we want to better consider the cognitive implications of affective

regulation in cultural settings. This will necessitate a better understanding of internal

processes that are induced by cultural and affective experiences, as well as the

development of a model of interactions between cognitive structures and schemes, and

modules such as long and short term memories, Among other things, this endeavour

aims at informing the design of new cognitive models for pedagogical agents in order

to make them more “humanistic”.

References

[1] Biggs, J. B., Teaching across cultures. In F. Salili, Chiu, C., Hong, Y (Eds), Student Motivation : The

culture and context of Learning, New York: Plenum Press (2001), 293-308.

[2] Blanchard, E. G., Adaptation-Oriented Culturally-Aware Tutoring Systems: When Adaptive

Instructional Technologies Meet Intercultural Education, to appear in H. Song and T. Kidd (eds),

Handbook of Research on Human Performance and Instructional Technology, Hershey: PA,

Information Science Reference (2009).

[3] Blanchard, E. G., Mizoguchi, R., Designing Culturally-Aware Tutoring Systems: Towards an Upper

Ontology of Culture, 1st workshop on Culturally-Aware Tutoring Systems: CATS2008, held in

conjunction with International Conference on Intelligent Tutoring Systems: ITS2008, Montréal,

Canada (2008).

[4] Blanchard, E. G., Mizoguchi, R., Addressing the Interplay of Culture and Affect in HCI: An

Ontological Approach, 13th International Conference on Human-Computer Interaction: HCII2009,

San Diego: Ca, USA (2009).

[5] Blanchard E. G., Volfson, B., Hong, Y.-J., Lajoie, S. P., Affective Artificial Intelligence in Education:

From Detection to Adaptation. International Conference on Artificial Intelligence in Education:

AIED2009, Brighton, U.K. (2009).

[6] Brown, P., Levinson, S. C. Politeness: Some universals in language usage. Cambridge: Cambridge

University Press (1987).

[7] Cooper, C. R., Denner, J., Theory linking culture and psychology: Universal and community-specific

processes, Annual Revue of Psychology, 49 (1998), 559-584.

[8] Dillon, I. L. Emotions in the Intensive Care Unit, in D. W. Crippen: End of Life Communication in the

ICU, New York: Springer Verlag (2008), 107-113.

[9] Eccles, J. S., Wigfield A., Development of achievement motivation. San Diego, CA: Academic Press

(2002).

[10] Ekman, P., Friesen, W. V., The Facial Action Coding System. Palo Alto, CA: Consulting Psychologists

Press (1978).

[11] Gerber, E. R., Rage and obligation: Samoan emotion in conflict. In G. M. White & G. Kirkpatrick

(Eds), Person, self and experience: Exploring pacifc ethnopsychologies, Berkeley: University of

California Press (1985), 121-167.

[12] Hofstede, G., Culture’s Consequences: Comparing Values, Behaviors, Institutions, and Organizations

Across Nations, 2nd edition, London: Sage (2001).

[13] International Business Etiquette and Manners website. http://www.cyborlink.com/ . Last consultation:

May 8, 2009.

[14] Johnson, W. L., Serious use of a serious game for language learning. International Conference on

Artificial Intelligence in Education, Marina Del Rey, USA. (2007), 67-74.

[15] Kashima, Y., Conceptions of culture and person for psychology. Journal of Cross-cultural Psychology,

31:1 (2000), 14-32.

[16] Lajoie, S. P., Azevedo, R., Cognitive tools for medical informatics. In S. P. Lajoie (Ed.), Computers as

cognitive tools (Vol.2): No more walls (pp. 247-271). Mahwah, NJ: Erlbaum (2000), 247-271

[17] Lajoie, S. P., Developing professional expertise with a cognitive apprenticeship model: Examples from

Avionics and Medicine. In K. A. Ericsson, K. A. (Ed.). The Development of Professional

Performance: Approaches to Objective Measurement and Designed Learning Environments.

Cambridge University Press (2009), 61-83.

[18] Lane, C. H, Hays, M. J., Getting Down to Business: Teaching Cross-Cultural Social Interaction Skills

in a Serious Game. 1st workshop on Culturally-Aware Tutoring Systems: CATS2008, held in

conjunction with International Conference on Intelligent Tutoring Systems: ITS2008, Montréal,

Canada (2008).

23

[19] Menon, U., Schweder, R. A., Kali’s tongue: Cultural psychology and the power of shame in Orissa,

India. In S. Kitayama & H. R. Markus (Eds), Emotion and culture: Empirical studies of mutual

influence, Washington D.C.: American Psychological Association (1994), 241-284.

[20] Mesquita, B., Emotions in collectivist and individualist contexts. Journal of Personality and Social

Psychology, 80:1 (2001), 68-74.

[21] Mesquita, B., Frijda, N. H., Scherer, K. R., Culture and Emotion, In P. Dasen & T. S. Saraswathi (Eds.),

Handbook of Cross-Cultural Psychology vol.2: Basic processes and Developmental Psychology,

Boston: Allyn & Bacon (1997), 255-297.

[22] Miller, C. R., Expertise and Agency: Transformations of Ethos in Human-Computer Interaction, in M.

Hyde (Ed.), The Ethos of Rhetoric, University of South Carolina Press (2004), 197–218.

[23] Ogan, A., Jones, C., Aleven, V.,. Pause, predict and ponder: use of narrative videos to improve cultural

discussion and learning. CHI2008. Florence, Italy (2008).

[24] Porayska-Pomsta, K., Mavrikis, M., Pain, H., Diagnosing and acting on student affect: the tutor’s

perspective, User Modeling and User-Adapted Interaction 18:1-2 (2008), 125-173.

[25] Rimé, B., Giovannini, D., The physiological patterns of reported emotional states. In K. R. Scherer, H.

G. Wallbott, & A. B. Summerfields (Eds), Experiencing Emotion: A Cross Cultural Study,

Cambridge: Cambridge University Press (1988), 84-97.

[26] Russell, J. Culture and the categorization of emotions, Psychological Bulletin, 110 (1991), 426-450.

[27] Ryan, R. M., Deci, E. L., Overview of Self Determination Theory: An Organismic Dialectical

Perspective. E. L. Deci & R. M. Ryan (eds). Handbook of Self Determination Research. Rochester,

NY: The University of Rochester Press (2002).

[28] Savard, I., Bourdeau, J., Paquette, G. Cultural Variables in the Building of Pedagogical Scenarios, 1st

workshop on Culturally-Aware Tutoring Systems: CATS2008, held in conjunction with International

Conference on Intelligent Tutoring Systems: ITS2008, Montréal, Canada (2008).

[29] Scharifian, F., On cultural conceptualisations, Journal of Cognition and Culture, 3:3 (2003], 187-207.

[30] Scherer, K. R., What are emotions? And how can they be measured? Social Science Information, 44:4

(2005), 695-729.

[31] Scherer, K. R., Brosch, T., Culture-specific appraisal biases contribute to emotion dispositions.

European Journal of Personality, 23 (2009), 265-288.

[32] Scollon, C. N., Diener, E., Oishi, S., Biswas-Diener, R., Emotions across Cultures and Methods,

Journal of Cross-Cultural Psychology. 35:3 (2004), 304-326.

[33] UNESCO guidelines for intercultural education. Available online at:

http://unesdoc.unesco.org/images/0014/001478/147878e.pdf (english version). Last consultation:

May 8, 2009.

[34] Wierzbicka, A. Talking about emotions: Semantics, culture, and cognition. Cognition and Emotion, 6

(1992).

24

25

Virtual Environments for Cultural Learning

H. Chad LANEa,1 and Amy E. OGAN b

a University of Southern California, Institute for Creative Technologies, USA b

Carnegie Mellon University, Human Computer Interaction Institute, USA

Abstract. We review six virtual learning environments built to support the

acquisition of cultural knowledge and communication skills: Croquelandia, ATL,

Second China, TLCTS, BiLAT, and VECTOR. Each leverages modern 3D video

game engine technology which allows high-fidelity simulation of new cultural

settings. This includes representations of buildings, streets, art work, dress, voice,

gestures, and more. To bring more realism to simulated cultural interactions,

several of the systems include artificial intelligence (AI) models of culture,

communication, and emotion. Additionally, several rely on narrative-based

techniques to place the target culture in context and enhance motivation of those

using the systems. We conclude with a discussion of the role virtual environments

might play in broader intercultural education programs and suggest that new

approaches for their evaluation are critical if they are to become more widespread

in their adoption.

Keywords. intercultural education, virtual learning environments, educational

games, narrative-based learning environments, cultural modeling

Introduction

Immersive learning environments provide new and unique ways in which to acquire

cultural knowledge and develop intercultural communication skills. High-fidelity

graphics, sound, and animation make it possible to simulate many tangible aspects of a

specific culture, such as buildings, streets, art, dress, speech, gestures, and more. This

enables the provision of more authentic computer-based practice environments than may otherwise be feasible using traditional live role-play and media-based approaches.

Further, recent advances in artificial intelligence (AI) and cognitive modeling now

permit rich modeling of emotions, language, tasks, and more [21]. When built with

cultural accuracy, these models – and the virtual humans who utilize them – open new

avenues for teaching the cognitive and interpersonal aspects of other cultures.

In this paper, we review six immersive cultural learning environments and discuss

them in the broader context of intercultural education. There is not only significant

potential to use systems such as these to enhance intercultural education programs, but

also to formatively track intercultural development and for summative learner

evaluation. That said, far more empirical studies are still necessary to fully gauge the

efficacy of virtual environments to promote cultural learning and intercultural development. We conclude with a discussion of the future of virtual learning

environments in intercultural training and suggest several areas of possible interest.

1 Corresponding Author: H. Chad Lane, USC Institute for Creative Technologies, 13274 Fiji Way,

Marina del Rey, CA 90292 USA, email: lane@ict.usc.edu

26

1. Cultural and intercultural learning

Cultural training programs have evolved substantially in the last six decades. The

earliest examples began to surface after World War II when international travel and

collaboration became more prevalent in business and government work. As the need for these programs became more evident, scientific interest in creating theories of

intercultural growth, identifying underlying cognitive processes, and demonstrating

their effectiveness also grew. The field of intercultural training has attracted researchers

from a variety of disciplines, including anthropology, cognitive psychology, social

science, business, and more. The usual structure of intercultural training programs

includes a blend of didactic and experiential components, including methods such as

lectures, discussion, film, case study, and role playing [13]. Typically, the goal is to

induce changes in knowledge, skills, and/or attitudes. Evaluations of intercultural

training programs also tend to focus on these three dimensions. A recent meta-analysis

suggests that many training programs have been found to be effective at teaching

cultural knowledge, but generally fall short in skill acquisition and attitude change [16]. There is widespread agreement among intercultural training researchers that

intercultural growth occurs in stages. Whether it be a student studying abroad, or a

business person starting a new branch in a foreign country, the assumption that people

acclimatize gradually is both intuitive and generally supported by psychometric

measures of cognitive, affective, and emotional change. One such model is Bennett’s

Developmental Model of Intercultural Sensitivity (DMIS) [3]. An underlying

assumption of the DMIS is that as one’s ability to construe cultural differences evolves,

intercultural competence also increases. According to Bennett, “it is the construction of

reality as increasingly capable of accommodating cultural difference that constitutes

development” [3, p.24]. The DMIS posits two broad worldview orientations:

ethnocentrism and ethnorelativism, and includes three sub-stages within each.

It is our contention that virtual learning environments may represent a significant leap for intercultural training programs. We believe that they could (1) promote

positive movement through stage-based models, possibly helping learners achieve

greater levels of intercultural competence than traditional methods alone, and (2) act as

an assessment tool to gauge movement through these stages. There is a long list of

historical successes of intelligent tutoring systems to teach cognitive skills [1], and

such systems offer affordances for repeated practice opportunities, environmental

control, and high-fidelity interaction. We disagree strongly with Fowler and Blohm

who say: “…if the goal is intercultural effectiveness, performance in multicultural

settings, sensitivity, and establishing cross-cultural relationships, these cannot be done

by a computer.” [7, p.40]. Recent research on virtual humans and how people interact

with them suggest otherwise. It is certainly reasonable now to investigate whether virtual learning environments, with virtual characters from other cultures, combined

with authentic intercultural situations, hold potential for intercultural training.

2. Virtual cultural learning environments

In this section we summarize six virtual cultural learning environments. We analyze

them in terms of their focus on knowledge, skills, and/or attitudes, describe their

interactions, learning activities, and underlying models, and finally report any empirical

findings about their effectiveness.

27

Figure 1. Adaptive Thinking and Leadership System for intercultural communication and leadership training;

developed at Sandia National Labs [19]

2.1. Adaptive Thinking and Leadership System (ATL)

The Adaptive Thinking and Leadership (ATL) system is a networked, first-person, 3D

role-playing environment that focuses on teamwork, intercultural communication, and

adaptive thinking [19]. The primary focus is on improving the players’ abilities to

make decisions under stress, balance lethal and nonlethal aspects of conflict, and apply

communication skills. Learners play the role of a U.S. Army Special Forces Soldier, an

indigenous Iraqi citizen, or an invisible evaluator. All players are given backstories and

goals to achieve while role-playing in the virtual environment (observers are given information about both sides). A screenshot of characters communicating with one

another in ATL appears in figure 1.

ATL enables higher fidelity role playing than would normally be possible in a

typical classroom simulation. Like Tactical Language and Culture Training System

(2.4), careful attention is paid to tangible cultural elements, such as appearance,

buildings, dress, and so on. Players interact with one another via speech and can move

around freely in the virtual world. An instructor station allows for control over the

simulation, such as the triggering of story events and the introduction of “curveballs”

that are useful for creating additional stress and surprise (such as an explosive device

going off). Intercultural communication knowledge is derived primarily from scenarios

and backstories that are given to the learners before they begin. As students interact with one another, they are expected to take the perspective of their character, adopt

appropriate desires and attitudes, and act in ways that they believe the fictional

character would. Perspective taking is a well-established approach in the field of

intercultural communication [12].

The instructor is also responsible for assigning evaluation tasks to observers so that

they can assess the performance of their peers’ interactions within the virtual world.

Typically, learners are rotated in and out of this role so that that after a session, they

have both role played and evaluated the abilities of others in their class. A study has

been conducted with 51 officers who completed self-report scales on the usability of

the system and their perceived learning [19]. Raybourn et al. found that the officers felt

the scenario was realistic and that they believed they learned about their own strengths and weaknesses by participating in the simulation.

28

Figure 2. BiLAT simulation for negotiation with cultural awareness;

developed at the USC Institute for Creative Technologies [11]

2.2. BiLAT

BiLAT is a game-based immersive environment that teaches the preparation, execution,

and understanding of bi-lateral meetings in a cultural context [11]. The focus is on both

knowledge and skills in Arab culture, but with emphasis on the culture of business

meetings and negotiation skills and no coverage of the Arabic language. BiLAT is

supported by an hour-long video that demonstrates the principles being taught. Most

scenarios in the simulation place the learner in the position of a high-ranking U.S.

Army officer who must solve a problem in a small Iraqi community, although other

scenarios have been authored, such as for military transition team training. A

screenshot of the BiLAT meeting screen appears in figure 2.

Success in BiLAT requires the application of several different skills. First, learners

must study the background story and select characters to visit who can help solve the problems. Second, learners must prepare for the meeting by collecting information and

filling out a preparation sheet that identifies important pieces of information, goals for

the meeting, potential impasses, and more. Finally, learners must conduct face-to-face

meetings with the characters to achieve their objectives, such as convincing an Iraqi

police captain to patrol a market. Meetings require adherence to Iraqi cultural and

social interaction rules, as well as the application of integrative negotiation tactics.

An intelligent tutoring system provides guidance and feedback during face-to-face

meetings in BiLAT [14,15]. The ITS supports reflection with an after-action review

(AAR) for each meeting [14]. It also assesses all communicative actions taken in a

meeting as having a positive, mixed, or negative impact on the character, then evaluates

the actions as a group and walks the learner through the AAR. Evaluations of BiLAT

have shown that:

• learners with minimal prior negotiation experience seem to benefit the most

from using it [6]

• learners who spontaneously adopt social goals show the greatest gains in

understanding Arab culture [17]

• feedback from the ITS during meetings is effective at helping learners be

aware of culturally-relevant time periods in meetings and the Arabic business

practice of “following the lead of your host” [15].

Studies of learning from BiLAT fit squarely into the broader movement to understand

how people treat, react to, and learn from virtual humans.

29

Figure 3. Croquelandia for second language learning; developed at the University of Minnesota [22]

2.3. Croquelandia

Croquelandia is a 3D virtual environment designed specifically for learning Spanish

pragmatics, defined as the various ways in which meaning is communicated and

interpreted in interaction [23]. In particular, Croquelania deals with making requests. It

is intended to be used as part of an instructional sequence with college learners of Spanish. Learners collaborate and interact in three virtual spaces linked by an

interactive map: their host family’s house, a central plaza and marketplace, and a

professor’s office at the university. While in Croquelandia, players complete quests

related to specific requesting contexts in these spaces (e.g., requesting an extension on

a paper, asking a host father to have a party at the house).

Completing these quests successfully may require interaction with nonplayer

characters (NPCs)—computer-generated avatars present in each of the spaces. NPC

interactions were created from several role-play conversations with native speakers,

varying the semantics of the requests to expose different pragmatic meanings. These

conversations were then used as the basis for conversation trees that the learner chooses

from as the main mode of interaction in the game. The NPCs then react to the learners’

choices based on how the native speakers responded, resulting in various levels of success for quest completion. Additionally, learners are able to collaborate

synchronously with NPCs using voice or written chat, as well as with others from their

class. Beyond synchronous communication within Croquelandia, learners can also

create their own game content to share, and can leave messages for classmates on a

discussion board. Assistance is provided in the form of various tips and ideas that are

hidden throughout the environment. These tips were designed to help learners self-

monitor their progress and successfully complete each of the requests.

To evaluate the environment, a first study was completed with 53 participants, of

which 25 completed a pre- and posttest [22]. The discourse completion task used as the

assessment included six scenarios which varied in terms of the social distance between

interlocutors. In addition to the data from the Croquelandia environment and the pre- and posttests, one-on-one interviews were also conducted with the subgroup of 25

participants. Sykes reports that the data from the request scenarios show little change

from pre- to posttest. However, she reports that both the interview data and a set of

lengthy in-class presentations indicate an improved awareness of pragmatic issues.

30

Figure 4. Second China, an island in Second Life for Chinese cultural learning;

developed at the University of Florida [8]

2.4. Second China

Second China is a project with two components: a web-based text and multimedia

repository, and an island in Second Life which has been designed to mimic cultural and

visual aspects of China. The project focuses on “preparing the learner to behave with a

level of sophistication that communicates respect and understanding of the target

culture” [8]. The target audience is government personnel who will be conducting

operations in a foreign country. Students may choose to begin with the web-based or

the 3D content, and links are provided throughout both to access relevant content in the

other medium. A main pedagogical principle of the system is to let the learner

determine a personally relevant path through the available content and activities. When the learner arrives on the Second China island, he or she is met by a greeter

named Jiang who offers to guide the avatar to different locations around the island. The

learner may accept or may explore the environment independently. Exploration of the

island may include observing or performing activities that are culturally significant

(e.g., watching or doing taichi in the park), initiated using scripted animations. It may

also include walking around noting architectural variations and other experiential

aspects or participating in quests for information in a given locale. In addition to

unguided exploration, there are embedded scenarios that deliver important cultural

experiences. These scenarios are facilitated by a bot playing a culturally appropriate

role. The bot detects the presence of the avatar and guides the avatar through a scenario

that mirrors real life. To provide support, the learner receives questions throughout the guided learning experiences. Questions are delivered just-in-time where the particular

concept being checked is necessary for the continuation of the scenario, or where it

reinforces prior learning. While browsing, learners can access videos, or click on items

linking to websites that provide a selection of references. If accessed from within the

2D web-based portal, each scenario or locale in the environment is introduced by

outlining the target content areas and describing possible learning objectives.

Evaluation of the environment has begun in the form of a peer review committee

comprised of experts in various related areas. In a first peer review activity, committee

members were guided around Second China and participated in one of the planned

learning scenarios. Henderson and colleagues report that initial feedback, while

positive, highlights the need for continued research and attention to instructional design

such as clearer objectives, and evaluation metrics.

31

Figure 5. Tactical Dari for Arabic language and culture learning; developed by Alelo, Inc. [9]

2.5. Tactical Language and Culture Training Systems

The goal of the Tactical Language and Culture Training System (TLCTS) is to teach

functional skills in various foreign languages and cultures [9]. There are four versions

of TLCTS currently: Iraqi, Dari, Pashto, and French. Each provides three kinds of

interactive activities: part-task lessons that focus on core communication skills (called

“skillbuilders”), “arcade” games that provide more opportunities for basic practice, and

“mission” games that allow players to practice in realistic contexts. Mission games take

place in a high-fidelity 3D virtual world with the learner in a third person perspective (but fully controlling their avatar). The emphasis is on spoken communication: learners

must learn to speak the foreign language to complete the lessons and play the games.

Two views of the mission game in Tactical Dari are shown in figure 5.

The three modes of interaction in TLCTS provide different kinds of learning

opportunities. Skill builders focus on basic elements like vocabulary, pronunciation,

grammar and nonverbal behaviors. For example, one skill builder teaches the learner an

appropriate “goodbye” in several different contexts. Arcade games have a similar goal,

but give the learner artificial goals to achieve, such as listening to directions in the

target language and executing the steps to move out of a maze. The mission

environment is the primary vehicle in TLCTS for experiential learning where the

learner must engage in everyday activities, such as navigating a village and communicating in culturally appropriate ways. One example of an objective is to gain

permission to search a home for a person of interest. Characters in the mission

environment are autonomous and driven by AI models of speech, language, and

emotions [9]. The speech recognizer is trained using novice data so errors can be better

detected. Such errors are dealt with directly in the skill builder and arcade games, while

they are integrated naturally into the mission game (i.e., characters don’t understand

you). Further, using game performance and quiz results as evidence, TLCTS maintains

a probabilistic student model that tracks learning throughout use of the system.

Several evaluations of TLCTS have been conducted. A broad study on the

effectiveness of Tactical Iraqi on military participants reported significant gains in

learning of Iraqi Arabic and cultural knowledge [20]. Tactical Iraqi was used to investigate the “politeness effect” [24]. It was found that feedback messages that

helped the learner “save face” led to better learning and motivation. Finally, military

participants with high motivation and time in service, as well as those who spent the

most time with the Skill Builder tool, demonstrated the largest learning gains [10].

32

Figure 6. VECTOR for intercultural communication and peacekeeping training;

developed by CHI Systems [5].

2.6. Virtual Environment Cultural Training for Operational Readiness

The Virtual Environment Cultural Training for Operational Readiness system

(VECTOR) shares many similarities with Tactical Iraqi and BiLAT. The goals of the

system are to improve learners’ knowledge of Arabic culture and intercultural

communication skills, but with a focus on peacekeeping [5]. It situates learners in a

virtual foreign town and requires them to interact with locals, synthesize information,

and take actions to solve problems. A first-person perspective is used and users have free movement around the 3D world. This enables specific categories of cultural errors

to be made, such addressing locals in an improper order. An example of a game goal in

VECTOR is to find the identity of a bomber and stop him from attacking his next target.

Characters in VECTOR are controlled by adjustable cognitive models that encode

cultural rules, communicative action understanding, as well as emotional components.

A screenshot of VECTOR appears in figure 6.

VECTOR scenarios can be created with a set of associated authoring tools that

permit non-experts to create character models, user actions, training objectives,

dialogue trees, and underlying narratives [2]. Character behaviors are derived from

common superclasses that encode cultural norms, and can be tweaked such that they

react differently according to personality differences. Learners communicate with

characters through contextual, menu-based selections. VECTOR includes a “synthetic instructor” character that conducts an after-action review with the learner that

highlights right and wrong answers. Correct answers are reinforced with explanations

for why the actions had a positive effect on the characters, whereas mistakes are dealt

with by informing the learner what actions would have been more appropriate.

Feedback messages are created in the authoring tool and are associated with training

objectives [2] and displayed on the screen for the user to see during meetings.

3. Discussion

We have described six systems for intercultural training that cover different cultures

(Spanish, Chinese, Iraqi, Dari, Pashto, and French). They each utilize immersive

technologies such as Second Life or the Unreal Tournament Engine, while some are

integrated with other activities such as multimedia resources or skill building arcade-

33

style games. As mentioned earlier, Mendenhall’s review of intercultural training

programs suggests that behavior and attitudes are areas for improvement for

intercultural training [16]. It is generally believed that games have the potential to exact

changes in students’ knowledge, skills, and attitudes; however, the main focus of games so far, including those we have reviewed here, has been on knowledge and skills.

The immersive, affective power of games might go a long way towards enabling us to

change students’ attitudes towards other cultures. For example, dispositions like

empathy, ethnorelativism, etc., are a critical part of intercultural competence, and

narrative-based techniques are appearing more frequently in commercial video games

to evoke these kinds of emotions in players (just as movies do to draw audiences in and

keep their attention). These are fertile connections for future research.

Immersive cultural learning environments to date have focused primarily on

culture-specific training; that is, in any given system, the goal is to learn one particular

culture and a limited set of skills or knowledge within that culture. We perceive great

potential to use these systems to encourage students to transfer knowledge from one culture to another – to develop a set of culture-general skills. Byram [4] describes such

a set of skills, including savoir comprendre, learning how to interpret and explain

cultural practices and compare them with aspects of one's own culture. Taking a

culture-general approach to teaching in immersive environments might allow students

to more easily transfer to new skills within the same culture, or even to new cultures.

One question that remains is whether these simulations, like any simulation that

utilizes a necessarily simplified model of the world, present a valid representation of

the culture. While it is impossible and perhaps undesirable to model a culture with

complete accuracy, the cultural model might be deemed sufficient for learning goals if

it enables the student to acquire knowledge and skills in the target culture. With respect

to accuracy, it is still an open question how to validate such simulations. The models in

the systems discussed in this paper were developed in various ways, from role-playing with native speakers, to expert interviews, to the use of photographs and artifacts. A

future method of validation may be to return to independent natives of the culture to

assess their reactions to the content, and perhaps their ability to proceed successfully

through the simulation.

The systems described here are all relatively new, but are beginning to enter more

widespread use, especially with military audiences. The next major step is to more

rigorously evaluate their effectiveness against less expensive alternatives. Research

with most of these systems has only just begun to look at fundamental questions of

learning and acquiring intercultural competence. While evaluation of cultural skills

(and language skills) is difficult, we can turn to the intercultural competence

community who have developed a wide range of instruments to examine these questions [18]. These instruments range from self-report of attitudes to choosing

cultural explanations. Determining which of these instruments is appropriate for the

evaluation of immersive cultural learning environments is a first step, although this will

often depend on the context of use of the systems and most likely involves the use of

multiple methods to triangulate students’ developmental trajectories. Once the

effectiveness of these environments has been shown, they have the potential to be used

as testbeds for research into what leads to successful learning in these ill-defined,

interpersonal domains. The seeds for this type of research are appearing in studies of

BiLAT and TLCTS, but many more will be necessary to bring the fields of intercultural

training, language training, the learning sciences, and artificial intelligence closer

together.

34

Acknowledgements

The project described here has been sponsored by the U.S. Army Research,

Development, and Engineering Command (RDECOM). Statements and opinions

expressed do not necessarily reflect the position or the policy of the United States Government, and no official endorsement should be inferred.

References

[1] J.A. Anderson, A.T. Corbett, K. Koedinger, R. Pelletier, Cognitive tutors: Lessons learned. The Journal

of the Learning Sciences 4,2 (1995), 167-207.

[2] C.A. Barba, T.P. Santarelli, F.A. Glenn, D.C. Bogert, J. Belanich, Authoring tools of game-based training,

Proceedings of the Conference of the Society for Applied Learning Technologies (2006).

[3] M. J. Bennett. Towards ethnorelativism: A developmental model of intercultural sensitivity. In R. M.

Paige, Ed., Education for the Intercultural Experience, 21-71. Intercultural Press, Yarmouth, ME, 1993.

[4] Byram, M., Teaching and Assessing Intercultural Communicative Competence. Multilingual Matters,

Clevedon, 1997.

[5] J.E. Deaton, C. Barba, T. Santarelli, et al., Virtual environment cultural training for operational readiness

(VECTOR), Virtual Reality 8, 3 (2005), 156-167.

[6] P. Durlach, T. Wansbury, J. Wilkinson, Cultural awareness and negotiation skills training: Evaluation of

a prototype semi-immersive system, Proceedings of the Army Science Conference (2008).

[7] S.M. Fowler, J.M. Blohm, An analysis of methods for intercultural training, in Landis (2004), 37-84.

[8] J. Henderson, P. Fishwick, E. Fresh, F. Futterknecht, & B.D. Hamilton, Immersive Learning Simulation

Environment for Chinese Culture, Proceedings of Interservice/Industry Training, Simulation, and

Education Conference (2008), paper 8334.

[9] W.L. Johnson, A. Valente, Tactical Language and Culture Training Systems: Using artificial intelligence

to teach foreign languages and cultures, Proc of Innovative Applications of Artificial Intelligence (2008).

[10] W. L. Johnson, S. Wu, Assessing aptitude for learning with a serious game for foreign language and

culture. In Proceedings of Intelligent Tutoring Systems (2008).

[11] J. Kim, R.W. Hill, P. Durlach, et al., BiLAT: A game-based environment for practicing negotiation in a

cultural context. International Journal of Artificial Intelligence in Education (in-press).

[12] Kramsch, C., Context and Culture in Language Teaching, Oxford University Press, Oxford, 1993.

[13] D.Landis, J.M. Bennett, M.J. Bennett (Eds.), Hbk of Intercultural Training, Sage, Thousand Oaks, 2004.

[14] H.C. Lane, M.G. Core, D. Gomboc, A. Karnavat, & M. Rosenberg, Intelligent tutoring for interpersonal

and intercultural skills. Proceedings of Interservice/Industry Training, Simulation, and Education

Conference (2007), paper 1514.

[15] H.C. Lane, M. J. Hays, M.G. Core, D. Gomboc, E. Forbell, D. Auerbach, & M. Rosenberg, Coaching

intercultural communication in a serious game. Proceedings of the 16th International Conference on

Computers in Education (2008), 35-42.

[16] M.E. Mendenhall, G.K. Stahl, I. Ehnert, et al., Evaluation studies of cross-cultural training programs: A

review of the literature from 1988 to 2000. In Landis (2004), 129-144.

[17] A. Ogan, J. Kim, V. Aleven, C. Jones, Explicit social goals and learning in a game for cross-cultural

negotiation, To appear in Proceedings of the Intelligent Educational Games Workshop, Artificial

Intelligence in Education 2009 (in-press).

[18] R.M. Paige, Instrumentation in intercultural training, In Landis (2004), 85-128.

[19] E.M. Raybourn, E. Deagle, K. Mendini, J. Heneghan, Adaptive Thinking & Leadership

simulation game training for Special Forces Officers. Proceedings, Interservice/ Industry

Training, Simulation and Education Conference Proceedings (2005).

[20] E. Surface, E.Dierdorff, Special Operations Language Training Software Measurement of Effectiveness

Study: Tactical Iraqi Study Final Report, Special Operations Forces Language Office, 2007.

[21] W. Swartout, J. Gratch, R. Hill, et al., Toward virtual humans. AI Magazine 27, 2 (2006), 96–108.

[22] J. Sykes. Learner requests in Spanish: Examining the potential of multiuser virtual environments for L2

pragmatic acquisition, CALICO (in preparation).

[23] J. Sykes, A. Oskoz, A., S. Thorne, Web 2.0, synthetic immersive environments, and mobile resources

for language education. CALICO Journal, 25 (2008), 528-546.

[24] N. Wang, W.L. Johnson, The politeness effect in an intelligent foreign language tutoring system,

Proceedings of the International Conference on Intelligent Tutoring Systems (2008).

35

Pedagogical Experience Manipulation

for Cultural Learning

Robert WRAY b, H. Chad LANE

a,1,

Brian STENSRUD b, Mark CORE

a, Laura HAMEL

b and Eric FORBELL

a

aInstitute for Creative Technologies, University of Southern California, USA bSoar Technology, Ann Arbor, MI, USA

Abstract. Acquiring intercultural competence is challenging. Although intelligent

learning environments developed to enhance cultural learning can be effective,

there is limited evidence regarding how best to dynamically manipulate these environments in support of learning. Further, the space of potential manipulations

is extremely large and sometimes tangled with the implementation details of

particular learning systems in particular domains. This paper offers a framework for organizing approaches to such dynamic tailoring of the learning experience.

The framework is hypothesized to be useful as a general tool for the community to

organize and present alternative approaches to tailoring. To highlight the use of the framework, we examine one potential tailoring option in detail in the context of an

existing simulation for learning intercultural competence.

Keywords. Pedagogical experience manipulation, dynamic tailoring, intercultural competence

Introduction

The National Academy of Engineering has identified “personalized learning” and

“advancing virtual reality” as grand challenges for the new millennium [1]. Models of

human social and cognitive behaviors are enabling simulations to provide practice

opportunities in soft-skill and interpersonal domains, including cultural awareness [2,

3] among many others. Some simulations have attempted to adapt to individual learner

needs by tracking learning and presenting scenarios of appropriate difficulty. However,

little has been done to manipulate character behaviors and environmental events

dynamically, in a fine-grained way, with the goal of promoting learning. In this paper,

we report on a work-in-progress prototype that seeks to promote cultural learning via

tailoring of on-going dialogues of virtual human characters in a game-based learning

environment, which provides more direct individualization and personalization as

envisioned by NAE.

The technical approach falls into the broad category of experience manipulation,

which we define simply as adjustment of the usual behavior of simulation elements to

achieve an objective. If we assume that the goal of a simulation is to behave in the most

realistic way possible (i.e., fidelity is the most important dimension), then a decision to

1 Corresponding Author: H. Chad Lane, USC Institute for Creative Technologies, 13274 Fiji Way,

Marina del Rey, CA 90292 USA, email: lane@ict.usc.edu

36

manipulate the experience means that a goal other than fidelity has been given higher

priority. Examples of experience manipulation are evident in commercial video games

where entertainment value trumps realism. For example, a popular golf video game

will sometimes produce the sound of an explosion as the ball comes into contact with

the club. Although the player receives the realistic feedback from the simulation for

hitting a good shot (e.g., that the ball traveled 320 yards), the additional “explosion”

feedback is believed, at least by the game designers, to enhance the experience.

We are particularly interested in the use of experience manipulation to further and

to achieve pedagogical goals. Such pedagogical experience manipulation [4] concerns

how one can intrinsically adjust a learning environment and simulation to promote

learning. We assume pedagogical experience manipulation will generally be used in

tandem with extrinsic learner supports (i.e., those which come from “outside” the

simulation), such as instructional components and intelligent tutoring.

There are at least two opportunities for an intelligent system to adjust simulation

behaviors in order to promote learning (see Table 1). The first concerns “configuration”

where the system selects or generates situations/scenarios that are appropriate for a

learner at any given time. The second category of experience manipulation, and one

explored in this paper, is dynamically tailoring. We define dynamic tailoring as a

process that continuously adapts the environment and simulation to meet the needs of a

learner through his or her interactions with an on-going learning scenario.

As suggested by the table, manipulations for both configuration and dynamic

tailoring can be conceived to address the affective states of the learner, as well as direct

domain learning. While, as we highlight later, dynamic tailoring could have a clear role

for maintaining or improving learner motivation, in this paper, we primarily focus on

tailoring for domain learning, with intercultural communication the primary domain.

Table 1: Dimensions of pedagogical experience manipulation

� ��������������� ���� ������� ���

�� ����� ��������������������� �� �

�� ��� �������������

������������������ �� �

�����������������������������

����� ��� ��� �������� ������� �������������

���� ��� �������������

����� ������� ������������� ��

�����������������������������

Dynamic tailoring comprises not only more familiar methods of directly

supporting learning objectives, but also emerging methods that employ narrative

adaptation and user-interface manipulations [5-7] to engage and motivate. This paper

presents a framework that organizes a design space including all of these different

potential tailoring elements, aids design-time evaluation of options, and, long-term,

enables rigorous comparison and empirical evaluation of the impacts of different

tailoring approaches on learning.

1. Cultural and social learning with virtual humans

1.1. Acquiring cultural knowledge

Learning and adapting to a new culture is a significant challenge. In different cultural

contexts, interpersonal and communicative behaviors that seem natural may produce

unexpected results. For example, simple habits such as nodding and other forms of

backchannel feedback can lead to unintended agreements that may, in turn, negatively

37

affect trust, reputation, and so on. Although it is generally agreed that it takes years of

first-hand experience to fully acclimate (i.e., living in-country), it is certainly important

for someone who will be spending time in a new cultural context to prepare. This is the

problem cultural training programs attempt to solve.

To develop intercultural competence in a general way, training should be built

around identifiable stages of development [8]. For example, if one is rushed to the

point of behavior adjustment with limited or no understanding of the underlying

cultural reasons behind observed behaviors, it could greatly hinder their overall

development. Also, it is common for intercultural training programs to build off the

notion of cultural difference [8] and support learners through perspective taking

exercises [9]. Intercultural communication is a common focus of such approaches since

face-to-face contact is so prominent. Our work is focused on teaching intercultural

communication skills and our aim is to enhance a learner’s ability to identify cultural

differences, understand the underlying cultural concepts, and practice communicative

strategies in a computer-based learning environment.

1.2. The Cultural Meeting Trainer

The Cultural Meeting Trainer (CMT) is a simplified and scaled-down version of the

ELECT BiLAT system, developed at the Institute for Creative Technologies with the

support of their U.S. Army partners [10]. BiLAT is an immersive, virtual learning

environment that teaches intercultural negotiation skills. It includes instructional

materials as well as a virtual practice environment for conducting background research,

preparation for business meetings, simulation of social and business interactions with

virtual counterparts, and an after-action review.

The CMT focuses primarily on the cultural aspects of “small talk” and trust-

building that occur during social periods of meetings. Using the BiLAT game

mechanics and scenario data, the CMT enables interactions between a player/trainee

and several different simulated Iraqi non-player-characters (NPCs). The goal of each

interaction is to “chat” with an individual Iraqi and gain his/her trust sufficiently to

move into a formal negotiation.

Importantly, the CMT is focused on supporting research and experimentation. For

example, ELECT BiLAT represents the domain in a 3D virtual environment; the CMT

uses much simpler simulation technology (Flash, still images). The simpler world

representation and web-based interfaces make it a more portable and lightweight

system for implementing and evaluating different tailoring options. Because CMT

presently incorporates ELECT BiLAT’s learning objectives and supporting content

related to intercultural competence, evaluation results should be translatable to BiLAT

(as long as the results are not dependent on the lack of immersion in CMT). It may

even be possible to fold effective tailoring options into the BiLAT system as they are

identified and refined in the CMT.

2. Organizing the Design Space for Tailoring: A Framework

One of the motivations for the research we are pursuing is that while ELECT BiLAT

includes instructional support and an intelligent tutor that offers feedback, guidance,

and hints, a number of opportunities exist to further enhance the system’s ability to

convey the cultural knowledge it seeks to teach. This observation raises the questions

38

of what kinds of pedagogical experience manipulations are feasible and worthwhile in a

domain such as intercultural communication. Although CMT is a relatively simple

environment (in comparison to ELECT BiLAT), the number of possible interventions

is still very large. In order to understand the options and prioritize them, we have

developed a framework that organizes the design space for tailoring. The framework

enumerates the general reasons one might use tailoring strategies and the classes of

tailoring strategies one might use in simulation- or game-based learning environments.

The framework is designed to be general enough to describe potential interventions

across a wide-range of domains. It is not intended to be specific to the CMT or to the

domain of cultural learning. Others should be able to use the framework to organize

and situate their own tailoring research and extend and adapt the framework to their

needs. For example, this framework could be combined with an ontology of cultural

learning [11] to connect inter-cultural learning objectives to specific tailoring strategies.

Further, our hypothesis is that this framework will facilitate the separation of

evaluations of tailoring manipulations from evaluations of specific learning

environments offered for particular domains. For example, one might ask about the

relative effectiveness of tailoring via manipulating character utterances vs. highlighting

options as a scaffolding strategy, or compare the impact of outcome manipulation

across different purposes, timings, and frequencies independently of the domains in

which these strategies were implemented.

The following subsections describe the current framework, using examples from

CMT to highlight potential design options. To be explicit, we have not implemented

(nor intend to implement) all of these strategies and approaches outlined here. The

current purpose in this section is to begin to suggest the large design space of

potentially effective tailoring options and the way in which we have organized the

space to inform implementation choices.

2.1. Purposes of tailoring

Dynamic tailoring can be employed for five functional purposes:

Scaffolding is the process of adapting practice-environment content to support learning

and practice directly, via supports that enhance the development of conceptual or skill

knowledge, especially for novice or struggling students [12, 13]. Scaffolding enables

the system to manipulate the practice context to directly support observed deficits in

student knowledge/learning. For example, a tailoring scaffold might be to elaborate an

utterance by an NPC in response to incorrect action with an explicitly guiding phrase

("Please do not show me pictures of your wife; this is not appropriate in my culture.")

Fading: Scaffolding is often accompanied by processes that allow “fading,” the

gradual removal of scaffolds that are no longer necessary as students move to higher

levels in terms of the zone of proximal development within the learning space [14].

With respect to dynamic tailoring, fading strategies are ones that gradually remove

scaffolds as the student demonstrates increasing skill. For example, if there was a

scaffolding strategy to only offer task-salient options to a novice trainee, a fading

strategy for tailoring might be to increase the number of available options, including

distracters, with increasing skill.

Challenging: Challenges are dynamic content adaptations that make fine-grained, turn-

by-turn components of practice more difficult. They can be used to exercise specific

39

skills of the learner and keep him or her engaged and in the zone of proximal

development [15]. Challenges include manipulations that increase complexity or

difficulty, either for progression to higher levels or for assessment (e.g., “is the student

now capable of handling this situation?”). In the context of dynamic tailoring for

cultural training, challenging is assumed to be a decision that can be made for specific

interactions in a dialog (a touchy subject or surprising reaction) rather than at the

scenario level (a difficult character) and that can be customized and adapted to specific

learning objectives. An example of the latter option might be providing a student with

little direct feedback in response to a provocative question, leaving it to the student to

decide whether or not the question was appropriate.

Scaffolding, fading and challenging all involve manipulation of the level of

difficulty perceived by the student. However, they can be distinct processes. In

cultural training for example, challenging could mean manipulating character responses

to be more negative. However, if student performance decreases, scaffolding could be

used to increase the pedagogical content of character responses (i.e., the character

explains the rationale for their negative reactions). If student performance improves,

fading would decrease the pedagogical content of character responses, removing the

scaffolding.

Engaging: Student errors may derive not only from a lack of knowledge or expertise

but also from issues related to attention, stress, and motivation. These demands are

especially important for cultural training, as part of the goal of the practice environment

should be to cause (manageable) overloads on attention and stress to attempt to suggest

the real-world conditions for interaction. Psychophysiological sensors, such as EEG,

galvanic-skin response (GSR), and even eye tracking can provide insights into

cognitive state to enhance diagnostics and enable better tailoring [16]. Adaptations can

be prescribed that are designed to address errors or test student skill in response to

observations of cognitive state. Thus, adaptations may include both scaffolding

adaptations (e.g., when a novice student was observed making errors under high stress,

the system could introduce an event that reduced complexity). Others may be geared

more toward directly engaging the learner in the content (e.g., narrative adaptations).

An example might be forcing a strong reaction from a character in an on-going

conversation when attention to the conversation was observed to be lagging.

Individualizing: Trainees with motivations, mindsets, or technical illiteracies that do

not match the simulation experience may not engage or learn as well as those who do

enjoy / learn well from the experience [17]. To improve the effectiveness of a practice

environment and better reflect the preferences of a diverse population of trainees,

practice environments also need to adapt how the simulation actually works to the

learner. An example might be removing time pressure for a learner who is internally

motivated and thus likely prefers a more exploratory (test-evaluate) learning process.

2.2. Classes of tailoring actions

Regardless of the purpose of the tailoring manipulation, another key dimension

influencing the potential role of tailoring is to consider the ways in which any kind of

tailoring could be accomplished. For current simulation- and game-based learning

environments, tailoring manipulations can be grouped into five categories.

40

Outcome manipulation: The effects of the student's action are modified by the

tailoring system. For example, a scaffolding outcome manipulation strategy might be to

dampen the negative effects of errors when the student is a novice. Such a strategy may

reduce frustration and provide some sense of accomplishment for the novice learner.

Character utterance or gesture: A character generates some communicative

performative, mediated by tailoring. Tailored utterances can be made in response to a

student action or can also be anticipatory, providing context or even guidance for some

upcoming student decision. The responses generated by the character can be spoken

(including textual output) and also indirect (via gesture). Character response is a

subcategory of outcome manipulation but we bring it out as a distinct class for two

reasons. First, face-to-face interaction is particularly important in the CMT domain and

most of the feedback a player receives will come thru other characters. Second, in

current simulation environments, there is typically a strong encapsulation of the

implementation of character behaviors and other events in the simulation at the

implementation level. Thus, the implementation of any tailoring strategies for

utterances would likely differ substantially from more general outcome manipulations.

Player choice manipulation: Another way to tailor experience is to manipulate the

options and actions available to a character. Such manipulations would include direct

modifications of the actions available to the student at any one time. For example, in a

simulation (like CMT) where actions are chosen via menu, a scaffolding strategy might

be to highlight or filter the actions that are salient in the current situation. (Such

moding may work well in menu-based environments, but have limited applicability

elsewhere.) Player choice manipulations could also include narrative-style adaptations,

where routes of exploration are blocked or objects in the environment are not fully

functional until some learning or experiential objectives are met.

Simulation/Event manipulation: Experience can also be tailored by dynamically

adapting the situation to which a student is presented. This category is similar to

outcome manipulation, with the distinction that the event manipulation need not be

made in direct response to a student action. In general, this class includes any non-

reactive event introduced to support pedagogy rather than realism. For example, one

could introduce biases that decreased complexity and vicissitude in character

interactions for novice students and biases to increase complexity and unpredictability

in those interactions as the core concepts of interaction are mastered. Other examples

would be to introduce environment events that reinforced pedagogical feedback or to

bring the learner’s attention to the environment.

Gameplay manipulation: Tailoring may also include manipulations that change

the way the simulation is experienced. Examples might include adding or removing an

explicit representation of the progress and status, depending on the learner’s source(s)

of motivation, enabling access to external (non-game) content to amplify or aid the

pursuit of learning objectives, and changing the interface of the game to accommodate

the preferences and technical literacy of individual learners.

Table 2 introduces general example(s) for each class of action and then provides a

specific example(s) of these manipulations in CMT. The purpose motivating each

tailoring strategy is highlighted in bold.

41

Table 2: Examples of different types of tailoring actions available in simulated environments.

Example Tailoring Strategy CMT Example

Ou

tco

me

ma

nip

ula

tion

Early incorrect actions can make

progress difficult & frustrating. A

tailoring scaffold can reduce the

negative impact of incorrect actions so

it is possible for novices to advance in a scenario.

When a player jumps immediately to

business after introductions, the normal

negative effect of this “brash” action is

dampened, roughly simulating that this

character “knows” how Americans typically approach business.

Ch

ara

cter

utt

era

nce

A tailoring scaffold results in a

character providing direct guidance

about a learning objective in its response.

In response to the player showing a picture

of his wife, the character is directed to say:

“Showing such pictures is not a good thing to do in my culture.”

Ch

ara

cter

ges

ture

Tailoring can scaffold or fade with

facial expressions and body language to accompany character speech.

When a novice student shows an

inappropriate picture, the character displays

a look of surprise or disgust, in order to

amplify any uttered response.

Inp

ut

ma

nip

ula

tion

Tailoring can scaffold actions via

filtering inappropriate or highlighting

appropriate actions, fade, by changing

when filtering/highlighting occurs and

its extent, and challenge by

introducing “distracter” actions.

To challenge a student who previously

talked to a character who was a soccer fan,

the student is presented with a “talk about

soccer” action option when meeting with a

character for whom there is no prior

evidence indicating an interest in soccer.

Ev

ent

ma

nip

ula

tion

Tailoring injects “interruption” events

to challenge or engage. Events are

appropriate for a one-on-one meeting

and may be used to cue meeting

transitions.

To engage a student who is not attending to

the simulation, tailoring introduces an “over

the top” reaction, perhaps emphasizing the

busyness of the character and the need to

make better progress.

3. Exploring Tailored Character Utterances in CMT

As suggested above, in the “small talk” domain of the CMT, the majority of feedback

within the simulation experience itself comes from the on-going conversation with

other characters. The native character content provides up to twelve different possible

responses to every player action or utterance. This approach provides a mechanism for

the existing system to customize responses to the situation. For example, a player who

has gained little trust with the character and makes a mistake may be presented with

“harsh” feedback (“Do not bother me with your pointless apologies!”). However, when

trust is greater, the character may be more accepting of the mistake and offer a response

without a negative valence. These “polarities” in the responses could be rigorously

mapped to theories of “positive” and “negative” face [18] but, in the present

exploration, we have only annotated the responses in an informal way. Having many

potential responses also increases variation, which in turn better supports replay.

Replay is important for novices because, due to mistakes and missteps, multiple

discussions with the same character may be required before gaining trust sufficient

enough to conclude a business meeting successfully.

42

In addition to having different “polarities,” some character utterances also contain

pedagogical guidance. For example, after the action of showing a picture of one’s wife,

which is generally inappropriate in Iraqi culture, the character can say, “She is pretty,

but you should not show this picture to an Imam, I think.” This kind of statement is

supportive of the student in terms of affect (it is warm) but also points out the mistake

in a constructive way that brings to mind the learning objective that has been violated.

In the current system, the choice of utterance is made via a function that depends

on the level of trust, a set of intrinsic character traits (e.g. level of power, belief) and a

random element. We are currently investigating a tailoring approach for character

utterances that will allow the system to select utterances according to trust and the

experience and skill of the student. Figure 1 illustrates the process. Our hypothesis is

that biasing the choice of response in terms of a student’s skill and experience will

provide scaffolding for new learners and even experienced students encountering

situations that cover new learning objectives. Further, this approach could possibly

improve engagement in more experienced learners as more pedagogically-oriented

content is filtered out for mastered learning objectives.

We have annotated all the character utterances in the system with tags indicating

the polarity of the utterance (positive, neutral, negative) and the pedagogical guidance

in the utterance (none, low, high). In our marking scheme, “high” pedagogical

utterances must refer directly to a learning objective (“do not show pictures of your

wife”) while “low” pedagogical utterances may offer less specific feedback about the

action (“You let your wife dress in this way? Your ways are clearly not ours.”).

Whenever a student selects an action, a “monitor” process now tracks student

progress and skill against specific learning objectives through the use of an expert

model. The goal is to contextualize the student action in terms of the student’s

experience in the domain and their level of skill. Depending on these factors, a tailoring

process chooses the level of guidance needed in the utterance and the desired level of

polarity. The request can be for an utterance with a specific value or a range of values.

The dialogue manager provides the tailoring system with utterances meeting the

constraints. The tailoring process then chooses an utterance which is generated by the

system for the player to see. The tailoring process is also communicating its choices

about the level of guidance and polarity it would like to use to a Coach [19] that

provides explicit guidance and feedback. For example, if the tailoring process desired a

high guidance option and the dialogue manager was not able to provide one, then the

Figure 1: An approach to tailoring character utterances by annotation for polarity & pedagogical

content.

43

Coach might decide to intervene and offer explicit feedback or a hint in the absence of

the desired character utterance.

While the specific details of the tailoring process are still to be worked out, we are

anticipating that for a novice student, new to the domain and the learning objective, the

tailoring process might bias utterances toward positive polarity and high guidance,

simulating a sort of “patient” character who easily forgives mistakes. As experience is

gained in the domain, different characters would respond with more variability in

polarity, even when offering high guidance for new learning objectives. As the student

gains experience and skill, the tailoring process for pedagogical utterances plays less of

a role, but may bias actions toward strongly negative polarity when the experienced

student makes careless/thoughtless mistakes.

Simulation developers typically focus on realism in the dialogs, rather than

utterances that are explicitly pedagogical. Thus, not surprisingly, there is a relative

dearth of character utterances with “high” pedagogical content in the currently-

authored dialogs. We are currently considering both re-authoring some of the existing

dialogs, to include more pedagogical content and authoring “pedagogical elaborations”

that would be added to the original utterances to bring out pedagogical content

explicitly. Our initial review suggests that we will only need to author a few

pedagogical elaborations for each set of character responses. This framework extends

naturally to non-verbal actions (body language, facial expressions) associated with

character responses. A minimally pedagogically informative response might have

neutral language with the non-verbal action conveying the positive or negative reaction.

4. Conclusions

Simulation environments are useful tools for both the development of “soft skills” like

intercultural competence and highly-proceduralized skills (e.g., flight control).

However, practice is typically more effective when it is targeted to the learner’s ability,

needs, and motivations and accompanied by supportive guidance and feedback. In this

paper, we have defined a framework for organizing and comparing different

approaches to targeting practice via adaptive tailoring. We introduced tailoring both in

the broad sense of its use in practice environments and described its potential role and

application in a specific cultural training system.

Much remains to be accomplished in terms of exploring and evaluating the

framework and the motivating hypothesis, especially in terms of measuring the effect

of particular tailoring strategies on learning outcomes and user satisfaction. Also, in the

case of simulations for learning, the question of fidelity is particularly important. Our

long-term goal is to explore the hypothesis that pedagogical concerns should play a role

in determining simulation behaviors. If a tailoring action produces a feasible and

believable effect, and is better for the learner, then trumping the default behavior of

the simulation engine seems reasonable.

We recognize the current framework is also incomplete; for example, it is not clear

whether adaptations designed to prompt reflection on learning would be a distinct class

of tailoring function or if this would be a category outside of framework (such as

explicit tutoring). However, our intent is that the framework offers a way for

researchers to organize, present, and discuss their approaches to tailoring independently

of the specific domains in which they are working and the implementation details of

their tailoring strategies. The framework may also prove beneficial in the design and

44

implementation of game- and/or simulation-based learning systems to facilitate

organization and prioritization of learner tailoring in these systems.

Acknowledgements

The work described in this paper was supported by the Air Force Research Laboratory,

contract number FA8650-08-C-6837. We thank LT Sarah Fockel and Dr. Joseph

Weeks for their support of this research, as well as Daniel Auerbach and Kelly

Christoffersen for their support in the development of the CMT. BiLAT was sponsored

by the U.S. Army Research, Development, and Engineering Command (RDECOM).

References

1. National Academy of Engineering, Grand Challenges for Engineering, V. Chase, Editor. 2008, National

Academy of Sciences/National Academy of Engineering: Washington, DC. 2. Johnson, W.L., H. Vilhjalmsson, and S. Marsella. Serious games for language learning: How much game,

how much AI? . in 12th International Conference on Artificial Intelligence in Education. 2005.

Amsterdam: IOS Press. 3. Kim, J., et al., BiLAT: A game-based environment for practicing negotiation in a cultural context.

International Journal of Artificial Intelligence in Education, in press.

4. Lane, H.C. and W.L. Johnson, Intelligent Tutoring and Pedagogical Experience Manipulation in Virtual Learning Environments, in The PSI Handbook of Virtual Environments for Training and Education, J.

Cohn, D. Nicholson, and D. Schmorrow, Editors. 2008, Praeger Security International: Westport, CT.

5. Magerko, B., B. Stensrud, and L.S. Holt. Bringing the Schoolhouse Inside the Box - A Tool for Engaging, Individualized Training. in 25th Army Science Conference. 2006. Orlando, FL.

6. Mott, B.W. and J.C. Lester. Narrative-centered tutorial planning for inquiry-based learning

environments. in Proceedings of the 8th International Conference on Intelligent Tutoring Systems. 2006. 7. Riedl, M.O., et al., Dynamic experience management in virtual worlds for entertainment, education, and

training. International Transactions on Systems Science and Applications, Special Issue on Agent

Based Systems for Human Learning, 2008. 4(2): p. 23-42. 8. Bennett, M.J., Towards ethnorelativism: A developmental model of intercultural sensitivity, in Education

for the Intercultural Experience, R.M. Paige, Editor. 1993, Intercultural Press: Yarmouth, ME. p. 21-71.

9. Kramsch, C. and R. Anderson, Teaching Text and Context through Multimedia. Language Learning & Technology, 1999. 2(2): p. 31-42.

10. Hill, R.W., et al. Pedagogically structured game-based training: Development of the ELECT BiLAT

simulation. in Army Science Conference. 2006. Orlando. 11. Blanchard, E.G. and R. Mizoguchi, Designing Culturally-Aware Tutoring Systems: Towards an Upper

Ontology of Culture, in Workshop on Culturally-Aware Tutoring Systems at the 9th International

Conference on Intelligent Tutoring Systems, E.G. Blanchard and D. Allard, Editors. 2008: Montreal. p. 23-34.

12. Davis, E. and N. Miyake, Explorations of Scaffolding in Complex Classroom Systems. The Journal of the Learning Sciences, 2004. 13(3): p. 265-272.

13. Wood, D., J. Bruner, and G. Ross, The role of tutoring in problem solving. Journal of Child Psychology

and Psychiatry and Allied Disciplines, 1976. 17: p. 89-100. 14. Pea, R.D., The social and technological dimensions of scaffolding and related theoretical concepts for

learning, education, and human activity. The Journal of the Learning Sciences, 2004. 13(3): p. 423-451.

15. Vygotsky, L.S., Mind and society: The development of higher psychological processes. 1978, Cambridge, MA: Harvard University Press.

16. St. John, M., et al., Overview of the DARPA Augmented Cognition technical integration experiment.

International Journal of Human-Computer Interaction, 2004. 17: p. 131-149. 17. Coffield, F., et al., Should we be using learning styles? What research has to say to practice. 2004,

Learning and Skills Research Centre: London.

18. Brown, P. and S.C. Levinson, Politeness: Some universals in language usage. 1987, Cambridge: Cambridge University Press.

19. Lane, H.C., et al. Coaching Intercultural Communication in a Serious Game. in ICCE. 2008.

45

The Relevance of the Culture Based Model

in Designing Culturally-Aware Tutoring

Systems

Patricia A. Young

University of Maryland Baltimore County

Abstract. Designing for culture through intelligent tutoring systems (ITS) is on

the rise. The needs of military personnel to communicate and understand cultures

other than their own in deployments, missions, and work-related assignments has

strongly encouraged the creation of culturally-aware tutoring systems (CATS) that

teach about other cultures. This paper provides insight into 3 systems (i.e., ELECT

BiLAT, Tactical Iraqi and VECTOR) and the frameworks that guided the design

and development process. The examination reveals that there is a need for a

comprehensive framework that directs project management and project design.

The Culture Based Model is offered as a framework to guide the design of CATS.

Keywords. Culturally-Aware Tutoring Systems, framework, The Culture Based

Model

Introduction

The design of culturally-aware tutoring systems (CATS) has come of age. Recent

examples of intelligent tutoring systems (ITS) reveal a culture-specific platform

engendered by the latest technology. These systems include: Tactical Iraqi, SABRE,

VECTOR, ELECT-BILAT, and ATL [12]. These American born cultural learning

systems, seek to aid military personnel nationally and internationally with learning the

culture of their allies and enemies. In this sense, meeting the needs of a few has become

as important as meeting those of the many.

Traditionally, designing ITS have excluded culture-specific considerations [18].

The student or learner was considered a static being whose knowledge became part of

the “box.” However, the learner remained devoid of an identity or in this case a

cultural identity. If we are too truly design for humans, than design considerations

based on culture is the nexus of this journey.

This paper evalutes several CATS by providing a brief description of the system,

an examination of the framework used to develop the system, and a summary of the

systems components. In particular, the frameworks are examined to determine their

potential to provide guidance in project management and project design. Project

management entails problem solving, planning, educating, evaluating, learning, and

decision making. Project design focuses on monitoring and content development [26].

Lastly, the Culture Based Model (CBM) is presented as a comprehensive framework

for the design of CATS.

46

1. Deconstructing CATS

ELECT-BILAT, Tactical Iraqi and VECTOR provide authentic samples as CATS.

This deconstruction is based on literature reviews versus a hands-on examination of the

systems.

1.1. ELECT-BILAT

ELECT-BILAT (Enhanced Learning Environments with Creative Technologies for Bi-

lateral negotiations) is a serious game-based simulation that teaches military forces

how to conduct business negotiations and meetings in a cultural context. BiLAT stands

for bilateral engagements that describe the deliberate events required in the meeting of

two parties [11][17]. A more thorough discussion of the ELECT-BILAT system and its

functionality can be found in [11]

ELECT-BILAT used a 5 phase approach compiled by Hill et al. [11]; this

approach sought to provide a “pedagogical structure” to the training objectives. Hill et

al’s. [11] 5 phase approach is a compilation of best practices used in the training and

development community. In phase 1, a task analysis is performed on data such as:

interviews, field manuals, research reports and articles. These data became the primary

sources for developing the training objectives. In phase 2, a storyboard prototype is

developed based on the task analysis results. The storyboard serves as a preliminary

tool to develop the content. In phase 3, a computer version is developed from the paper

prototype. In phase 4, training objectives are connected to the content. These objectives

provide on-going assessments of the learner’s acquisition of content knowledge. By

phase 5, supplemental materials are developed to support learners, instructors and

trainers. These materials serve as a printed version of the training system and an

individualized training module.

The 5 phase approach is a set of best practices to be considered when integrating

technologies with pedagogy, learning and training. Cultural elements are not included

in the 5 phase approach but some cultural elements became part of the design process

in ELECT-BILAT. Overall, the 5 phase approach suggests a focus on project

management and project design.

1.2. Tactical Iraqi

Tactical Language and Culture Training Systems (TLCTS) is a serious game platform

that teaches a foreign language and culture through interactive 3D video game

simulations. Several language courses have been developed under TLCTS; they include

Tactical Iraqi (2006); Tactical Pashto (2007), Tactical French (2007) and Tactical

Dari (2008) [1]. These are ITS integrated with serious game technology. Tactical Iraqi

focuses on a dialect called Arabic Iraqi [15]. Military forces use this training to develop

cultural competency. A more thorough discussion of the Tactical Iraqi system and its

functionality can be found in [1] and [15].

TLCTS implemented a task-based approach to the design of learning systems like

Tactical Iraqi. The task-based approach is derived from Doughty & Long’s [8]

research that proposes 10 methodological principles of “Task-Based Language

Teaching” (TBLT). TBLT principles aid in the production of courses that tailor to

learner’s communicative competencies, functional proficiency in a foreign language

and pedagogical practices specific to second language acquisition. These principles are

47

specific to the development of distance learning foreign language and computer-

assisted language learning programs. The 10 TBLT principles include: (1) use tasks,

not text, as the unit of analysis; (2) promote learning by doing; (3) elaborate input; (4)

provide rich input; (5) encourage inductive learning; (6) focus on form; (7) provide

negative feedback; (8) respect developmental processes and learner syllabuses; (9)

promote cooperative/collaborative learning and (10) individualize instruction.

TBLT is a basic framework for designing language learning technologies. Cultural

elements are not included in the 10 TBLT principles but some cultural elements

became part of the design process in Tactical Iraqi. Overall, TBLT suggests a focus on

project design.

1.3. VECTOR

Virtual Environment Cultural Training for Operational Readiness (VECTOR) is a

training program that provides experience based cultural familiarity through a virtual

serious game environment. VECTOR was first piloted with a focus on the Kurdish

culture in Iraq [6]. Military personnel use this training to develop cultural sensitivity as

it relates to cultures other than their own. A more thorough discussion of the VECTOR

system and functionality can be found in [2], [3] and [6].

VECTOR follows a framework that aids in the construction of training scripts and

scenarios [6]. Deaton et al. [6] adapted Wise et al.’s [21] Cultural Communications

Skills Template (CCST) that provides a taxonomy to organize elements of a culture.

This template is detailed enough to provide most of what should be known about a

culture and specifically focuses on the identification of behaviors. The three major

areas and subsequent criteria of CCST are: (a) Psychological Profile (cultural values,

relationships, and motivations), (b) Culture’s Shared History (historical tradition, social

structures, manners/protocol, linguistic characteristics, the arts, science and technology,

culinary tradition, recreation and leisure, sports, vice, hygiene, and ethnic or minority

groups), and (c) National/Regional Characteristics (geography, polity, legal system,

military institutions, economic structure, finance, buying/leasing/selling goods and

services, housing, health, communications, transportation, and educational system).

The basic template is applied to a distinct culture; this application furthers the

development of a culture-specific template. The culture-specific template is then

compared to mission templates, and these data are used to design training courses.

CCST is a basic framework for defining, designing and analyzing cultures, and

more specifically identifying cultural communications skills within a culture. CCST’s

template consists of 27 cultural elements and some elements became part of the design

process in VECTOR. Overall, CCST suggests a focus on project design.

1.4. Understanding the 3 CATS

The conscientious effort to consider culture makes these 3 systems culturally-

aware [4]. Table 1 provides an overview of ELECT-BILAT, Tactical Iraqi and

VECTOR. The table specifically highlights some important components to the design

of CATS. In these cases, the types or genres are serious games. These systems sought

to teach skills and content directly related to the Middle Eastern culture. There is a

focus on the learner or user of the system; this focus is carried throughout the design

process. There is also a focus on the particular society, culture or target audience (i.e.,

Middle Eastern). When designing CATS, specific technologies are needed to enhance

48

the overall design. In each case, the most available and applicable technologies were

integrated into the systems. A teaching method became important to the design process

as these systems integrated skills-based, task-based or case-based learning. The

learning category specifies how content is delivered; in each case there is a focus on

individualized instruction and, to integrate the research on cognition and learning,

VECTOR incorporated cognitive apprenticeship. Consistent across ELECT-BILAT,

Tactical Iraqi and VECTOR, the development of these systems is guided by an

approach or framework. The investigative work that needs to be done in creating CATS

maybe referred to as ethnographic research, because there must be a collection of data

through participant observations, interviews, and in-depth field work[9]. Broadly

defined, cultural elements refer to those criteria used to identify a particular society,

culture, or target audience. The data collected and cultural elements integrated in

ELECT-BILAT, Tactical Iraqi and VECTOR is consistent with ethnographic research

methods.

Table 1. Culturally-Aware Tutoring Systems. Data gathered from [1], [2] [6], [11], [15] and [17].

Product ELECT-BILAT

(Middle Eastern)

TACTICAL IRAQI VECTOR

(Kurdish)

Genre/Type serious game serious game serious game

Goal teach skills for conducting

business negotiations and

meetings in a cultural

context

teach Arabic language &

culture

teach skills needed to

interact in a variety of

cultural contexts

Learner/User military forces US military, Australian

Defense Force

military personnel

Culture/Society Middle Eastern Iraqi Kurdish

Culture-Target

Audience

Middle Eastern people

Iraqi people

Kurdish people

Technologies -Integrating Architecture

-SmartBody

-PsychSim

-Automated Speech

Recognition

-Enabling Learning

Objectives

-Lithtech Jupiter Game

Engine

-iGEN Intelligent Agents

Teaching

Method

skills-based task-based learning case-based learning

Learning individualized instruction individualized instruction -individualized instruction

-cognitive apprenticeship

Framework 5 phase approach [11] Task Based Language

Teaching [8]

Cultural Communications

Skills Template [21]

Data Collection -bilateral materials

-interviews of subject

matter experts

-analyzed transcripts of

Iraqi-American role-

playing scenarios

-live simulation training

exercises

-Iraqi people role play in

mock-up Iraqi towns

-Sources: field manuals,

research reports and

articles from military

personnel

-Arabic language

-Kurdish culture

-Sources: books, articles,

pamphlets, military

reports, cultural

information about Iraq and

other cultures

Cultural

Elements

values, social interactions,

behaviors, norms

cultural knowledge,

language, communication

(verbal & non-verbal),

social interactions

behaviors, cultural

knowledge, gender,

language, religion, status,

perceptions,

predispositions, attitudes,

space, social interaction,

emotions, personality

Supplemental

Materials

printed version of training

system

online & hardcopies of

reference materials

do’s and don’t booklet

49

2. The Culture Based Model & CATS

Frameworks provide solutions or guidelines to design [19]. The 5 phase approach

provides a limited number of management and design features and the cultural

elements are not explicit in the approach. TBLT does not focus on project management

and project design is limited with this approach. Also, TBLT does not contain any

cultural elements to guide the design process. CCST does not focus on project

management; it is more focused on project design. Alternatively, CBM focuses on

project management and heavily on project design, Given this, CBM is offered as a

comprehensive framework to design CATS.

The Culture Based Model is a basic framework; it qualifies as a model of culture.

In this sense, CBM works with other frameworks, models, approaches, templates, etc.

It could be seen as a bridge that connects one discipline to another or as a layer in

building systems, concepts, or products. CBM defines cultures. The framework does

not label cultures with terms such as individualistic, because there is too much

differentiation within societies, cultures and people. CBM can be applied to more than

one culture. Use the basic framework to create culture-specific guides.

CBM is an intercultural instructional design framework for the design of culture-

based information and communication technologies (ICT). Within CBM’s framework,

the term ICT is used to define the various forms of technologies. In this context,

culture-based means the ICT can be generic, culture neutral and considered

internationalized. Or the ICT can be specialized, culture-specific and considered

localized. The model provides guidance in building generic and specialized products

and services. CBM has 8 areas aligned with the acronym ID-TABLET (Inquiry,

Development, Team, Assessments, Brainstorming, Learners, Elements, Training).

Inquiry and Elements focus on project design – all other areas focus on project

management. There are a total of 70 design factors with additional subfactors not

included in this total.

CBM can aid CATS developers in creating content for domain models; capturing

attitudes, emotions and behavior [22]; building cultural profiles of learners [4] and

collecting the necessary data to build systems. Building CATS through CBM’s ID-

TABLET requires making cultural considerations during the management and design

process or from beginning to end. It is a conscientious effort on the part of the designer,

team, client and all stakeholders to develop an authentic product.

CBM begins with getting stakeholders to identify design goals. These goals can be

classified in one of the following categories: C: Custom development; A: Add-on; R:

Re-engineer; or D: Diagnostic evaluation. Custom development is building a project

from scratch. An add-on means adding on to an existing product. Re-engineering

requires revamping the existing product in terms of making changes to the software

codes, restructuring, or rebuilding. Diagnostic evaluations serve as an ethnographic

measurement tool.

Given the direction of the project, corresponding areas of CBM’s ID-TABLET are

selected and a database or knowledge management system is built. This database serves

as a “culture guide” for building training programs and other products. Extensive

details on CBM can be found in [26].

The next sections briefly outline how CBM works, presents the areas of the ID-

TABLET and offers connections to CATS. Where applicable, connections are made to

ELECT BiLAT, Tactical Iraqi and VECTOR. Inquiry

50

Inquiry monitors development, automates the internal flow of the design process and

functions as internal sensors. This area offers a series of questions to be asked and

answered before, during and after the products production. These questions are meant

to guide the designers or teams conceptualizing and keep everyone focused on the

society, culture or target audience. There are 6 Inquiry design factors: I1. Genre, I2.

Framing, I3. Omission, I4. Backgrounding, I5. Foregrounding, and I6. Visual

representations. Each design factor serves a different function in the design process.

Genre aids in the selection of the ICT, considers affordability, and determines

other influences on the product’s design. The first question asks: What ICT are being

used and why [24][25][26]? In selecting technologies for CATS, there should be a

focus on issues such as access and equity. These questions get at whether the learner

can access the technology and whether there have been considerations to provide an

accessible technology. ELECT BiLAT, Tactical Iraqi and VECTOR access is granted

and restricted to military and government personnel.

Framing assists in maintaining the target audience’s perspective. These questions

align the project to the target audience with a focus on content. For example one

question asks: How is the content presented to the target audience [24][25][26]? In

designing for CATS, there should be considerations for whose perspective is being

presented and how is the content delivered. The presentation of content can be different

based on the technology platform (web based versus software) or where objects are

placed (text on left or right of screen).

Omission helps in assessing the design; these questions determine what has been

omitted intentionally or unintentionally. An omission question asks: What has been

intentionally omitted and why [24][25][26]? In designing for CATS considerations as

to why content, ideas, and people were omitted needs to be evaluated. The personal

ideologies of the team should be intentionally omitted from the design. There should be

more focus on the target audience or culture rather than one’s own beliefs and values;

thereby, bias is minimized.

Backgrounding helps to provide a balanced design. A balanced design is assessed

by determining what content has been backgrounded intentionally or unintentionally

and whether this backgrounding is detrimental to the project. A backgrounding

question asks: Is the backgrounding intentional or unintentional and why

[24][25][26]? In designing for CATS, considerations for whether backgrounded items

are implicit (implied) or explicit (clear) need to be discussed and discovered. A

community representative or cultural expert may readily notice a backgrounded item as

being appropriate or offensive.

Foregrounding assists in providing an objective design. This objectivity is

determined through an analysis of what and why something is emphasized. A

foregrounding question asks: What is emphasized and why [24][25][26]? In designing

for CATS, considerations for foregrounded items should be consistent with the

expectations of all stakeholders. Foregrounded items could include: text, color, icons,

objects, etc.

Visual representations assist in conveying meaning to the design. Meaning

maybe conveyed related to purpose, product, process, selection and placement of

images. A visual representations question asks: Who is portrayed in these visual

representations [24][25][26]? In designing for CATS, determine whether the images

portrayed are those stakeholders wanted to communicate.

51

2.1. Development

Development provides the management structure for problem solving. It considers

those features that are important in the overall development of the product. There are

10 Development design factors: D1. Consider technical, aesthetic, content, culture-

based, and target audience (TACCT) design specifications, D2. Mass distribution

formats, D3. Effective technology to produce the product, D4. Diversify ICT

format, D5. Understand target audience, D6. Explore environmental &

individual/group cultures, D7. Quality design, D8. Authenticate product, D9.

Control for interference, D10. Model the product or process [24][25][26].

In terms of CATS, providing a management structure to the design process is

needed. D1 offers TACCT as a preliminary assessment tool to evaluate pre-existing

technologies. D2 helps the designer consider issues of access to the technology and

equity in the number of equal chances learners have to engage the technology. D3 aids

the designer in considering whether the technology is effective in delivering content;

that is, the human to machine interaction is considered. ELECT BiLAT, Tactical Iraqi

and VECTOR, as indicated in Table 1, also provided supplemental materials that were

online and print based. This is consistent with D4 that suggests diversifying the ICT

format to better meet learners learning modality. In the research, there is much concern

about understanding the learner [22], D5 gets the team thinking about these issues

before any prototype or storyboards begin. Consistent with “Acquisition–Oriented

CATS (systems trying to teach intercultural skills to learners)” [4, p. 24], D6 begins the

discussion to consider the multiple perspectives of societies, cultures and target

audiences. D7 looks at the quality of the product and the possibility that others might

emulate its design. D8 begins the ongoing discussion of accurately representing the

society, culture or target audience; this is balanced by the collection of authentic data in

CBM Elements. D9 focuses on minimizing bias, attitudes and prejudice throughout and

minimizing interference related to humans, machines and the environment. Phase 2 of

the 5 phase approach [11] (see also 1.1) requires the development of a storyboard

prototype. This is consistent with D10 to model a product or process through some

form of visual language [5].

2.2. Team

Team focuses on the recruitment of a culturally sensitive design team; this is where

much of the decision making happens. The 3 Team design factors include: T1.

Cultural expert(s), T2. Enlist educators, and T3. Culturally informed team

[24][25][26]. Creating any type of artificial intelligence is a team effort; however in

creating CATS there must be cultural experts as indicated in T1 [21]; [4]. Like most

educational products, there is a need for subject matter experts as indicated by T2. In

Table 1 under data collection, ELECT BiLAT used subject matter experts as a resource

for content. Last, T3 suggests that team members are educated about the society or

culture and accept the various racial, ethnic, linguistic, and other diversities of the

target audience. In the case of ELECT BiLAT, Tactical Iraqi and VECTOR, team

members would receive training about the Middle Eastern culture and its people. Team

members must positively accept and respect the Middle Eastern culture to be part of the

team.

52

2.3. Assessments

The area of Assessments covers multiple evaluation options. These evaluations provide

evidence of the products effectiveness and evaluate the goals set for the target audience.

There are 4 Assessments design factors: A1. Multiple evaluation options, A2. Assess

the assessment, A3. External review, and A4. Culture-specific assessments

[24][25][26]. ELECT BiLAT, Tactical Iraqi and VECTOR each have some type of

assessment tool. A1 suggests that the assessment be internal and external. In the case of

the 3 systems, this would mean that ITS have internal assessments and that the

assessments match the content delivered via the ITS. External evaluations might be

pilot studies, field tests, prepackaged exams, instructor made exams or standardized

tests. A2 provides an internal monitoring of the system to account for whether the

learning outcomes are being properly measured and whether the testing methods (e.g.,

fill-in or multiple choice questions) are effective. A3 suggests hiring external reviewers

to evaluate the product; the goal is to provide an objective evaluation of the system. A4

recommends the development of culture-specific assessments that culturally align with

the needs of the target audience and content.

2.4. Brainstorming

Brainstorming determines the direction to proceed and serves as an initial review

of the design process. It is the first step in planning the project. There are 10

Brainstorming design factors: B1. Financial support, B2. Pilot studies/field tests of

product, B3. Assess community’s response, B4. Community representative on

team, B5. Investigate target audience to authenticate product, B6. Reflect and

assess learning goals, B7. Affordable design, B8. Meet needs of target audience, B9.

Discuss and consider cultural context, B10. Present and consider outcomes

[24][25][26].

ELECT BiLAT, Tactical Iraqi and VECTOR received financial support from the

United States government. B1 advocates that financial funding be secured for the

duration of the project because culture-specific projects require more initial upfront

monies. TLCTS engages in ongoing pilot studies of its upcoming ITS [13]. B2 suggests

that pilot studies or field tests be conducted throughout the development process. These

ongoing evaluations reduce victimization of learners/users from failed products. B3

promotes an assessment of public opinion to the products development; this might be

carried out through opinion polls or focus groups. Similar to the cultural expert, B4

suggests a community representative who is versed in the racial, ethnic, social class and

economic conditions of the target audience; this person may have lived or still lives in

the target community. Discussions about the level of ethnographic data collected begin

in B5; the collection of data might include: interviews from cultural informants,

observations, artifact collection, commissioned reports, and articles. In Table 1,

preparation for ELECT BiLAT, Tactical Iraqi and VECTOR included data collection

consistent with that suggested in B5. B6 suggests that there be ongoing reflections and

assessments of learning goals throughout the products development. In meeting the

needs of the many and the few, B7 advocates for an affordable design that the target

audience can afford or obtain free through some other means of accessibility. B8 begins

discussions on how the product meets the instructional or cultural needs of the target

audience. The design of nonbiased products requires much discussion about the target

audience in terms of their history, social interactions, politics, economic status,

53

religious affiliations and other factors in CBM Elements; discussions about this topic

begin in B9. B10 suggests formalizing expectations for achieving learner outcomes or

meeting the user goals.

2.5. Learners

Learners centers on the needs of learners and learning. These design factors assist in

building a learning environment that supports the learner’s cultural frame of reference.

There are 10 Learners design factors: L1. Extend learning, L2. Differentiate

opportunities to learn, L3. Empower and engage learners, L4. Teach proactive

learning, L5. Identify educational objectives, L6. Culture-specific instructional

strategies, L7. Enrich instructional content, L8. Adapt instruction to learner, L9.

Plan for instruction, L10. Enculturate the learner [24][25][26]. L1 suggests the

need to extend learning opportunities by building adaptive designs that meet learners

where they are in terms of knowing and understanding; these opportunities should

extend or advance instruction as needed. Much research in artificial intelligence

promotes the building of adaptive learning environments [22]. L2 promotes

differentiated instruction that provides learners with options in instruction [7][10], and

applying these ideas more broadly to the design of CATS means that there can be

differentiation in hardware, software, teaching and instruction, and learning. L3

suggests that learners be empowered by and engaged in the content. By example,

ELECT BiLAT, Tactical Iraqi and VECTOR provide an engaging learning

environment through the use of game technology. L4 suggests proactive learning

requires building mechanisms into CATS that provide incentives to learners to improve

their learning; this can be achieved by creating multiple pathways for the learner to

improve, build, or develop skills, abilities, and experiences. When building any e-

learning system or product the educational or learning objectives should be identified

first—as suggested in L5; this means that the learning outcomes are known to the

designer and the learner. L6 advocates for consideration of culture-specific

instructional strategies that are specialized to a particular individual or group. These

instructional strategies take into consideration how the target audience learns and the

best way of approaching this learning. Instead of applying the learning theory or

instructional strategy to the system, assessments are made to determine the best

learning strategies that work for the learner. For example, cognitive apprenticeship

would not be integrated into the ITS as with VECTOR (see Table 1); the ITS would

determine the best instructional strategy based on the needs of the learner. Then the

most appropriate learning theory or instructional strategy is applied by the system. L7

proposes that instructional content can be enriched by expanding the content to culture-

specific elements of learners such as: politics, morality, ethics, beliefs, language,

identity, and social actions. L8 recommends an adaptive instruction that is not too

grade level or age level specific but focuses more on cognitive ability; connections to

cognitive ability seem more in line with the goals of building artificial intelligence. L9

suggests that the long and short term instructional needs of the learner be considered;

this means how knowledge is acquired over a span of time. Finally, L10 proposes that

the product can enculturate the learner into a culture. Enculturation into the Middle

Eastern culture seems to be the goal of ELECT BiLAT, Tactical Iraqi and VECTOR.

54

2.6. Elements

Elements facilitate content development. These Elements seek to be

comprehensive in providing the fundamental total of which all culture is composed.

There are 25 Elements divided into 3 sections.

Section I: The Anthropology of Culture covers design factors E1–E13. This section

draws from key concepts in the fields of anthropology, communications, demography,

economics, language, history, and the physical and environmental sciences to explore

the depth of culture. The design factors include: E1. Cultural aesthetics, E2. Cultural

artifacts, E3. Cultural capital, E4. Cultural classification, E5. Cultural

communications, E6. Cultural demographics, E7. Cultural environment, E8.

Cultural history, E9. Cultural knowledge, E10. Cultural language, E11. Cultural

physiology, E12. Cultural relations and E13. Cultural resources.

Section II: The Psychology of Culture covers design factors E14–E20. This section

draws from the fields of cognitive anthropology and cultural psychology that focus on

cognitive, psychological, and social realms. The design factors include: E14. Cultural

beliefs & values, E15. Cultural experiences, E16. Cultural ideas, E17. Cultural

identity, E18. Cultural interests, E19. Cultural misconceptions, and E20. Cultural

ways.

Section III: The Science of Culture covers design factors E21–E25. This section

draws from key concepts in the fields of biological science, earth science, ecology,

physical science, futures research and cross-cultural studies to explore the scientific

nature of humanity and the possibilities of cultural futures. The design factors include:

E21. Cultural anomalies, E22. Cultural cultures, E23. Cultural futures, E24.

Cultural infinities and E25. Cultural nature [24][25][26].

The database of information about the society, culture or target audience is built

from the Elements. Data is collected in the applicable areas of the Elements. These data

become a culture guide to the particular society, culture or target audience. Once the

database is built, multiple ITS could be developed based on these data. By example, a

comprehensive database of information on the Middle Eastern culture could have aided

in building the 3 systems: ELECT BiLAT, Tactical Iraqi and VECTOR.

Through Elements the human aspect is brought to design. The design factors

under Section I: The Anthropology of Culture could benefit the development of CATS

in the following ways:

• builds a comprehensive taxonomy of a society, culture or target audience

• aids in the development of content for domain models [22]

• aids in understanding the different forms of cultural knowledge

• offers options to consider multiple parameters about a society, culture or

target audience

• aids in building an authentic cultural profile of learners [4]

• allows for the development of new cultural models to generate training

modules or interaction scripts [20]

• allows for intercultural analysis

• assists in determining cross-cultural clashes

The design factors under Section 2: The Psychology of Culture could benefit the

development of CATS in the following ways:

• builds a comprehensive taxonomy of the human being

• aids in capturing attitudes and emotions [22]

55

• assists in understanding human behavior

• aids in understanding communication and language issues

• allows for an understanding of reasoning specific to culture [4]

The design factors under Section 3: The Science of Culture could benefit the

development of CATS in the following ways:

• builds a scientific taxonomy of what is and what could be

• allows for considerations based on cultural diffusion (borrowing,

acculturation & appropriation)

• considers how the environment impacts humans and other species

• considers the future of humanity

• makes considerations for the science of humanity

2.7. Training

Training is the education of individuals. This is another phase in providing full

management of a project. There are two design factors for training: Tr1. Product

training and Tr2. Culture-based training. Most of the information for this training is

gleamed from the data gathered through CBM Elements [24][25][26]. Tr1 suggests

training instructors, practitioners, and others about the product. Tr2 recommends

culture-based training of team members; this provides comprehensive coverage of the

society, culture, or target audiences.

3. Conclusion

Frameworks provide solutions or guidelines to design [19]. The needs of designers and

developers to implement the 5 phase approach TBLT, and CCST suggests that: (1)

there is a need for a framework, (2) there is a need to explore interdisciplinary research

and colleagues for some answers, and (3) it takes a village to build CATS.

Empirical studies suggest that CATS can improve learning, motivate learners and

teach information to learners. An empirical study of BiLAT reports that the system aids

learners in acquiring content material; this is due to the interactivity of the system and

interactions with the virtual human characters [17]. Preliminary studies of Tactical

Iraqi indicate that students gained knowledge, increased their understanding of the

Arabic language, improved job performance and enhanced “organizational outcomes”

[14]. The games, in general, motivate learners and motivated learners leads to a

successful learning experience [16]. Continued empirical research is needed to provide

more conclusive measures of learning gains, strengths and weaknesses of CATS, the

impact of cultural content on teaching and learning, and future areas of improvement.

Designing CATS is a conscientious effort that may take more upfront time but

provide many benefits in the end. This review demonstrates that there is a process in

the design of CATS. This process should be guided by a framework that provides

assistance in project management and project design. The development of culture-

specific products will eventually move beyond those focused only on language or a few

localization features. Therefore, a culture-based model like CBM provides the

comprehensive framework needed in building CATS.

56

References

[1] Alelo, Inc. (2009). Tactical Language and Culture Training Systems. Retrieved May 7, 2009, from

http://www.tacticallanguage.com/

[2] Barba, C., Deaton, J. E., Santarelli, T., Kneer, B., Singer, M., & Belanich, H. (November 2006a). Virtual

environment composable training for operational readiness (VECTOR). Paper presented at the 25th Army

Science Conference, Orlando, FL.

[3] Barba, C., Santarelli, T., Glenn, F., Bogert, D., & Belanich, J. (2006b). An approach to scenario authoring for

virtual environment training. Paper presented at the BRIMS Conference, Baltimore, MD.

[4] Blanchard, E. G., & Mizoguchi, R. (2008). Designing culturally-aware tutoring systems: Towards an upper

ontology of culture. Paper presented at the Intelligent tutoring systems: Past and future -- Culturally-aware

tutoring systems conference, Montreal, Canada.

[5] Botturi, L., & Stubbs, S. T. (2008). Handbook of visual languages for instructional design: Theories and

practices. Hershey, PA: Information Science Reference.

[6] Deaton, J. E., Barba, C., Santarelli, T., Rosenzwieg, L., Souders, V., McCollum, C., et al. (2005). Virtual

environment cultural training for operational readiness (VECTOR). Virtual Reality, 8, 156-167.

[7] Dewey, J. (1897). My pedagogic creed. New York, NY E. L. Kellogg & Co.

[8] Doughty, C. J., & Long, M. H. (2003). Optimal psycholinguistic environments for distance foreign language learning. Language, Learning & Technology, 7(3), 50-75.

[9] Ferraro, G. (2004). Cultural anthropology: An applied perspective (5th ed.). Belmont, CA: Thomson Learning,

Inc.

[10] Gagne, R. M. (1962). The acquisition of knowledge. Psychological Review, 69(4), 355-365.

[11] Hill, R. W., Belanich, J., Lane, H. C., Core, M., Dixon, M., Forbell, E., et al. (2006). Pedagogically structured game-based training: Development of the ELECT BiLAT simulation. Paper presented at the 25th Army

Science Conference.

[12] Hussain, T., & Feurzeig, W. (2008, March 24-28). Methods and tools for the development of effective training

games. Paper presented at the American Educational Research Association: Research on schools,

neighborhoods, and communities: Toward civic responsibility, New York, NY.

[13] Johnson, W. L., & Valente, A. (March 2008). Tactical language and cultural training systems: Using artificial intelligence to teach foreign languages and cultures. Paper presented at the Innovative Applications of

Artificial Intelligence Conference 2008, Chicago, IL.

[14] Johnson, W. L., Valente, A., & Heuts, R. (2008). Multi-platform delivery of game-based learning content.

from http://www.alelo.com/publications.html

[15] Johnson, W. L., Wang, N., & Wu, S. (2007). Experience with serious games for learning foreign languages

and cultures. Paper presented at the SimTecT 2007 Simulation: Improving capability and competitiveness, Brisbane, Australia.

[16] Johnson, W. L., & Wu, S. (2008). Assessing aptitude for learning with a serious game. Paper presented at the

Intelligent Tutoring Systems 9th International Conference, Montreal, Canada.

[17] Lane, H. C., & Hays, M. J. (2008). Getting down to business: Teaching cross-cultural social interaction skills

in a serious game. Paper presented at the Intelligent tutoring systems: Past and future -- Culturally-aware

tutoring systems conference, Montreal, Canada.

[18] Sleeman, D., & Brown, J. S. (1982). Intelligent tutoring systems. London: Academic Press, Inc. Ltd.

[19] van den Akker, J. (1999). Principles and methods of development research. In J. vandenAkker, R. M. Branch,

K. Gustafson, N. Nieveen & T. Plomp (Eds.), Design approaches and tools in education and training (pp. 1-

14). Boston, MA: Kluwer Academic Publishers.

[20] Weiland, W. J., Deaton, J., Barba, C. A., & Santarelli, T. P. (April, 2003). Virtual environment cultural training for operational readiness: VECTOR (No. ARI Research Note 2003-10). Lower Gwynedd, PA: CHI

Systems, Inc.

[21] Wise, J. C., Hannaman, D. L., Kozumplik, P., Franke, E., & Leaver, B. L. (1998). Methods to improve cultural

communication skills in special operations forces (No. ARI Contract Report 98-06). Alexandria, VA: United

States Army Research Institute for the Behavioral and Social Sciences.

[22] Woolf, B. P. (2008). Building intelligent interactive tutors: Student centered strategies for revolutionizing e-learning. Burlington, MA: Morgan Kaufmann/Elsevier.

[23] Young, P. A. (2008a). Integrating culture in the design of ICTs. British Journal of Educational Technology,

39(1), 6-17.

[24] Young, P. A. (2008b). The Culture Based Model: A framework for designers and visual ID languages. In L.

Botturi & T. Stubbs (Eds.), Handbook of visual languages for instructional design: Theories and practices (pp.

52-75). Hershey, PA: IGI Global.

[25] Young, P. A. (2008c). The Culture Based Model: Constructing a model of culture. Journal of Educational,

Technology & Society, 11(2), 107-118.

[26] Young, P. A. (2009). Instructional design frameworks and intercultural models. Hershey, PA: IGI

Global/Information Science Publishing.