Interacting with Dynamic Social Knowledge

Revealing Challenges through an Analysis of Pragmatic Aspects of

Problem Solving

Rodrigo Bonacin

, Heiko Hornung

, Julio Cesar Dos Reis

, Roberto Pereira

and M. Cecília C. Baranauskas

FACCAMP and CTI Renato Archer, Rodovia Dom Pedro I, Km 143.6 - 13069-901, Campinas - SP, Brazil

Institute of Computing – UNICAMP, Av. Albert Einstein nº 1251, 13084-722, Campinas - SP, Brazil

CR SANTEC, Public Research Centre Henri Tudor, 2A rue Kalchesbrück, L-1852, Luxembourg, Luxembourg

Keywords: Pragmatic Web, Social Network Systems, Organizational Semiotics, Knowledge Evolution, Knowledge

Visualization.

Abstract: In the Social Web users interact with each other in multiple contexts expressing meanings and intentions.

Knowledge production in this context can be understood as a dynamic socio-cultural process. Mechanisms

that support users to explore this knowledge in an effective and efficient way may bring benefits from a

personal and social perspective. However, the construction of these interaction mechanisms is dependent on

new models and techniques to dynamically represent and visualize the shared knowledge. The interpretation

of the content by users is influenced by meanings and intentions, as well as by the understanding of the

evolution of these aspects over time. This paper analyses the evolution of meaning and intentions in

collaborative problem solving scenarios using Social Network Systems. The analysis method has its roots in

Semiotics and Speech Act Theory. Results indicate research challenges for new interaction possibilities by

representing the evolution of the pragmatic aspects and their relations with the semantic ones. To address

these open research problems we present a preliminary conceptual framework for multidisciplinary research

in three interconnected perspectives: interactive, conceptual and technical.

1 INTRODUCTION

The Social Web (SocWeb) (Gruber, 2008) has a

dynamic nature with respect to both content and

enabled interactions. This dynamic nature affects the

users’ interpretation and intentions, and

consequently their possibilities of interaction with

the system. At the same time, the Semantic Web

(SemWeb) (Berners-Lee et al., 2001) proposes to

model the information in Web ontologies, aiming at

enabling knowledge interpretation by artificial

agents and by people. However, some of the main

issues of the SemWeb include how it truly enables

the connections of the Web of people who will use

it, and how to turn “messy” human knowledge into a

shared information space that is useful to everyone

(Hendler and Berners-Lee, 2010). Therefore, an

alignment of the SemWeb with SocWeb visions may

bring benefits, but also open challenges in terms of

requiring novel interaction methods and techniques.

In the context of the SocWeb, knowledge

representation models applied to social software

should take into account the dynamic aspects of the

knowledge produced and exchanged by people in

these systems. These models could enable richer

users’ interactions by representing the evolution of

the pragmatic aspects and their relations with the

semantic ones. In this paper, we study the evolution

of semantics and pragmatics as an integrated

process.

In many cases, it is desirable to maintain the

history of how content, interactions, as well as the

meanings, intentions and interpretations evolve over

time. In collaborative problem solving, for example,

usually the interactions, as well as the rationale and

history of the actions taken are as important as the

solution itself. However, the interpretation of

content generated during the collaborative process is

dependent on the analysis of the evolution of the

context, from the time when the content was created.

Therefore, the associated knowledge representation

models should evolve, maintain and present the

Bonacin R., Hornung H., Dos Reis J., Pereira R. and Cecília C. Baranauskas M..

Interacting with Dynamic Social Knowledge - Revealing Challenges through an Analysis of Pragmatic Aspects of Problem Solving.

DOI: 10.5220/0004004800540063

In Proceedings of the 14th International Conference on Enterprise Information Systems (ICEIS-2012), pages 54-63

ISBN: 978-989-8565-12-9

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

evolutionary history in a proper way, associating

meanings and intentions.

The construction of such systems relies on open

research issues, requiring a multidisciplinary view to

deal with. In this paper, research requirements are

analyzed and situated in three interconnected

perspectives: (1) Interactive Perspective: How users

will actually visualize, make sense and interact with

this dynamic content; (2) Conceptual Perspective:

How to conceptually make sense and model the

dynamic aspects of the knowledge (including

meanings and intentions); and (3) Computational

Perspective: How to implement and automate as

much as possible the construction of the models.

In order to explore requirements for these views,

the paper presents a study of dynamic aspects of

pragmatics in messages exchanged during

collaborative problem solving processes, within the

special education domain. Two scenarios were

explored: one in the “Vila na Rede” Social Network

System (SNS) (www.vilanarede.org.br), which

adopts a forum structure for questions and

discussions, and the other within “Yahoo! Answers”

(http://answers.yahoo.com/), which adopts the

structure of multiple answers to a single question.

The analysis was performed in three steps: the first

step is related to a quantitative analysis of the

interactions; the second step involves the

examination of messages using the pragmatic

communication analysis artifact (Liu, 2000); and the

third synthesises the results and the exploratory

analysis of interaction possibilities towards

requirements for research in related fields. From the

results of this study we extracted general

requirements and challenges for the interactive,

conceptual and computational perspectives. Based

on the challenges raised, we present and discuss a

preliminary framework for future research.

The paper is organized as follows: section 2

presents the background in Knowledge Visualization

(KV), Knowledge Evolution (KE) and Pragmatic

Web; section 3 describes the study of pragmatics

evolution in collaborative problem solving in SNS;

section 4 presents the challenges and discusses a

conceptual framework to deal with the issues;

section 5 concludes the paper.

2 BACKGROUND

In this section, we present the main areas related to

this work. Background works in KV (section 2.1),

KE (section 2.2) and Pragmatic Web (section 2.3)

are used in this paper for prospecting research

requirements.

2.1 Knowledge Visualization

Visualization techniques are strategies to deal with

the increasing quantity and complexity of subject

matters in many domains (Keller and Tergan, 2005).

According to Pampalk et al. (2003) visualizations

can make complex relationships easier to

understand, and stimulate visual thinking.

Knowledge Visualization (KV) differs from

Information Visualization (IV) techniques (Eppler

and Burkhard, 2004) in many aspects, including

goals, benefits, content, or recipients (Keller and

Tergan, 2005). In order to precisely define KV,

many proposals discuss the differences of data,

information, and knowledge, as well as try to reach a

precise definition of the concept of knowledge. The

concept of KV “in a strict sense is restricted to

externalizing aspects of knowledge by the individual

[…] in a ‘freestyle mapping mode’” (Keller and

Tergan, 2005, p. 7). Also, the term KV has a focus

on structured visualizations for the representation of

conceptual knowledge (Keller and Tergan, 2005).

In general, KV methods are required to make

knowledge explicit and better usable, as well as to

make sense of information structures (Keller and

Tergan, 2005). Burkhard (2004) proposes a KV

Framework consisting of three perspectives (a

Knowledge Type Perspective, a Recipient Type

Perspective and a Visualization Type Perspective),

which need to be considered when creating

visualizations that aim to “transfer” knowledge.

Different techniques have been proposed to address

KV, such as: Scientific Charts, Concept Maps,

Knowledge Maps, (Conceptual) Diagrams, Visual

Metaphors, (Heuristic) Sketches (Keller and Tergan,

2005); (Eppler and Burkhard, 2004). There are also

ontology-based (Kuß et al., 2009) as well as graph-

based (Maseri et al., 2007) approaches to KV. KV

studies have also been applied to the SocWeb

context (Hoetzlein, 2007).

In fact, an envisioned scenario for the Web

(Hendler and Berners-Lee, 2010) regarding the use

of SemWeb artefacts within SocWeb environments

poses new issues that have not yet been deeply

explored in the KV literature. Therefore

investigations to clearly identify the challenges and

new possibilities opened by this contemporary

scenario are still required. There are dynamic

aspects of the shared knowledge in social networks

that are not addressed by the current KV methods. It

is out of the scope of this paper to propose new KV

InteractingwithDynamicSocialKnowledge-RevealingChallengesthroughanAnalysisofPragmaticAspectsofProblem

Solving

techniques; however, we expect to contribute by

clarifying interaction possibilities for KV in a

dynamic SocWeb perspective.

2.2 Knowledge Evolution

The Ontology Evolution (OE) problem has mainly

emerged with the use of ontologies in the context of

the SemWeb. A well-accepted definition for OE is

given by Stojanovic (2004). The author defines OE

as: “the timely adaptation of an ontology to the

arisen changes and the consistent propagation of

these changes to dependent artifacts”. Over the last

years, distinct methods and approaches to organize

the evolution steps have been proposed to treat the

OE problem.

Stojanovic (2004) proposes a six-step method

that focuses on different aspects of the changes: (1)

detecting, (2) representing, (3) defining its

semantics, (4) implementing, (5) propagating and (6)

validating. For each of these steps, various different

approaches are proposed in the literature. Flouris et

al. (2007) present a survey on ontology change.

The OE problem has also been addressed and

considered under different perspectives. For

instance, approaches defined in ontology languages

for the OE (Avery and Yearwood, 2003), ontology

versioning (Klein and Fensel, 2001), and

community-based OE (Leenheer and Meersman,

2007). These approaches and methods for OE have

resulted in the development of tools for supporting

the OE process. Software applications such as

OntoStudio

(www.ontoprise.de/en/products/ontostudio) or

Protégé (protege.stanford.edu) are generally

augmented with additional functionalities through

the use of plug-ins in order to support specific OE

requirements. OE methods, techniques and tools are

important for supporting the proposed conceptual

and technical perspective.

In this work we present requirements for the

development of new KE techniques that consider the

relations between semantics and users’ intentions.

2.3 Pragmatic Web

According to Morris (1938), pragmatics can be

understood as the relationship between signs and

humans. It concerns aspects such as intentions,

communications, conversations and negotiations.

While the areas of KV and KE focus on aspects

related to the visualization of knowledge and

formalisms that describe knowledge evolution, the

Pragmatic Web is also concerned with the question

of how knowledge is actually constructed and how it

evolves during the collaboration among people that

is mediated by Web artifacts. Originally proposed as

an extension or a complement of the SemWeb, the

Pragmatic Web addresses topics such as context and

meaning negotiation in the Web (Singh, 2002);

(Schoop et al., 2006).

The Pragmatic Web perspective has been applied

to a variety of research domains, e.g., multi-agent

systems (Paschke et al., 2007), interaction design

(Hornung and Baranauskas, 2011), self-organizing

communities of practice (de Moor and van den

Heuvel, 2004), or Web Services (Liu, 2009).

Pragmatic Web research is often rooted in different

Information Systems research frameworks and

theories, e.g., the Language/Action Perspective

(LAP) (Goldkuhl and Lyytinen, 1982); (Winograd

and Flores, 1986) or Organizational Semiotics (OS)

(Liu, 2000).

The basic unit of analysis of LAP is a speech act.

LAP subscribes to the notion that we perform

actions through language. Thus, collaboration is

coordinated by the performance of speech acts,

which underlie socially determined rules (Schoop,

2001). In OS, basic units of analysis are affordances

and agents. Initially introduced by Gibson (1968),

the concept of affordances was expanded by

Stamper (1996) to represent patterns of behavior that

are governed by systems of norms in the physical

and social world. Agents are entities (persons or

groups of people) that can be attributed with

responsibility. OS’s basic ideas have been

formulated as “there is no knowledge without a

knower, and there is no knowing without action”

(Liu, 2000; p. 26). OS subscribes to the notion that

knowledge about the world, and the underlying

systems of norms are constantly changing.

Considering LAP with OS (Cordeiro and Filipe,

2003) as theoretical frames of reference, and having

as object of study Web-mediated collaboration and

meaning negotiation, the Pragmatic Web, thus,

provides an important basis for this work.

3 ANALYSING DYNAMIC

KNOWLEDGE

The empirical data that has been analyzed in this

work has been gathered during activities conducted

in the context of a research project named TNR

(www.nied.unicamp.br/tnr) in the domain of Web-

mediated continuous learning of Brazilian special

education teachers. Regarding the required training

for special education teachers, the Ministry of

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

Education defined an eighteen-month distance

learning course for regular teachers. During this

course, teachers learn to discuss a so-called “case”

of a student with special needs.

In Brazil, special needs are classified into 7

categories (visual, auditory, motor, intellectual

impairment, intellectual giftedness, pervasive

developmental disorder, multiple impairments), and

in order to constitute a meaningful and

representative group of 28 participants, 4 specialists

of each category who expressed a familiarity with

the use of information and communication

technology (ICT) were chosen randomly.

In the context of this project, the initial activities

aimed at learning more about the way the 28

participants use Web systems, and how they engage

in different forms of Web-mediated conversations.

To this end, four consecutive activities were

planned. Due to space limitations, the analysis

presented in this work is based on two of the four

activities, namely the discussions conducted in “Vila

na Rede” and “Yahoo! Answers”. Vila na Rede is an

inclusive social network that permits to post

“announcements” and to comment on them.

Comments are displayed in a hierarchical structure

below the announcement, and may contain text,

pictures, audio or video. If the creator of the

announcement authorizes it, other users may

“collaborate”, i.e., change or augment an

announcement’s text or media files. Yahoo! Answers

is a Web system that permits a user to post a

question and other users to post answers to that

question or vote for the best answer.

Out of the 28 teachers, 16 participated in the two

activities (9 participated in both activities, 6 only in

Yahoo! Answers and 1 only in Vila na Rede). These

teachers had no previous experience with the

mentioned systems, but were already used to Web

applications such as blogs, email, and forums. A

“case” was posted in each system and teachers were

asked to discuss and solve it.

3.1 Methods

The objective of the present analysis is to investigate

requirements and possibilities for a prospective

computational mechanism that explores the dynamic

aspects of meanings and intentions in the contents of

problem solving in SNSs. From the analysis of

probable meanings and intentions of the written

communication acts, we expect to explore new

possibilities and extract requirements for research,

design and implementation of such mechanisms, as

well as to reveal research challenges to be overcome.

The first step of the three-step analysis regards

the analysis of interaction including quantitative

aspects. The interactions (e.g., comments, questions,

answers) among users during the problem solving

are enumerated and analyzed in a temporal order,

resulting in an interaction graph. The activity of the

network is also observed. One interaction may

contain more than one message addressed to a

receiver.

The second step involves the communication

examination based on the pragmatic analysis

presented in Liu (2000). We propose the use of this

technique because it provides a structured way to

analyze pragmatic aspects in messages. According to

the Speech Act Theory (SAT) (Austin 1962, Searle

1976) the acts are classified as locutionary (i.e.,

actual utterance and its ostensible meaning),

illocutionary (i.e., propositional contents carrying

intentions) and perlocutionary acts (i.e., effect on the

addressee). As presented by Liu (2000), a message

can be divided into two parts: the content part of a

communication act that manifests the meaning of the

message as it is expressed in the proposition; and the

function part of a communication act specifies the

illocution that reflects the intention of the speaker.

An interaction between users identified in the first

step can be broken down into one or more messages.

In this sense, for example, a long answer is

considered one interaction unit, but it may contain

more than one message unit from the

communication act point of view.

In Liu (2000) the illocutions are grouped into

three dimensions: time (i.e., whether the effect is on

the future or the present/past), invention (i.e., if the

illocution used in a communication act is inventive

or instructive, it is called prescriptive, otherwise

descriptive), and mode (i.e., if it is related to

expressing the personal modal state mood, such as

feeling and judgement, then it is called affective,

otherwise denotative). By using these dimensions,

the illocutions are classified as: 1. Proposal (future,

prescription and denotative), 2. Inducement (future,

prescription and affective), 3. Forecast (future,

description and denotative), 4. Wish (future,

description and affective), 5. Palinode (present/past,

prescription and denotative), 6. Contrition

(present/past, prescription and affective), 7.

Assertion (present/past, description and denotative),

and 8. Valuation (present/past, description and

affective).

Figure 1 shows the proposed evaluation form

used in the second step. For each message, two

analysts attributed continuous values from 0.0 to 1.0

for each dimension, for example, 0.0 for a message

InteractingwithDynamicSocialKnowledge-RevealingChallengesthroughanAnalysisofPragmaticAspectsofProblem

Solving

that the analyst judges to be totally denotative, and

1.0 for a totally affective one. An analyst might, for

instance, attribute a confidence level smaller than 1

to a phrase (s)he could not classify confidently.

Figure 1: An example of communication message

analyzed using the proposed form.

Based on the values attributed before, a

predominant classification is attributed for each

message, and optionally the analyst can also indicate

a secondary (or alternative) classification. The

analysts also identify aspects associated with the

content part: the role of who performed the message

(“speaker”), and the main object, affordance or

proposition that the message refers to. This step was

manually performed by the two analysts.

The third step is the synthesis of the results of

step one and two. Graphs showing the evolution of

each of the three dimensions of the illocutionary acts

were produced to facilitate the analysis. A free

exploratory analysis, in the format of a

brainstorming, was performed aiming to investigate

how the dynamic aspects could potentially

contribute to interaction design for continuous

learning, regarding mainly topics such as content

recovery or search. Finally, a synthesis of the

challenges and future research needs was performed.

3.2 Results

Figure 2 presents an interaction graph produced in

the first step of the analysis of Vila na Rede case.

The circles represent the 12 users (10 participating

teachers and 2 facilitators from the research team)

that performed at least one interaction, and the arcs

represent their interactions. Circle sizes are

proportional to the total number of interactions the

respective users were involved in. The arc thickness

is relative to the number of interactions performed

by each pair of users.

By observing Figure 2, it is possible to identify

that the users V1 and V10 were the most active ones.

In average, each of the 12 users performed 9.25

interactions, and approximately 77% of all

interactions where performed as an answer to a

previous one. This can be interpreted as evidence

that the problem solving and meanings were

constructed in an interactive process, in which

messages were constructed dynamically over the

interpretation of previous messages.

Figure 2: Interaction Graph of Vila na Rede Case.

In the second step, the content of the messages

was analyzed using the results of step one as a

starting point. Figures 3 to 5 present the evolution of

the time (Figure 3), invention (Figure 4), and mode

(Figure 5) dimensions during the problem solving

process in the Vila na Rede activity. Some

interactions analyzed in the first step consisted of

two or more illocutionary acts. A total of 170

illocutionary acts (contained in 110 interactions)

were identified and plotted on the horizontal axis of

Figures 3 to 5. The dashed lines in Figures 3 to 5

represent the polynomial trend lines of each

dimension. Looking at the trend line we can

visualize the evolution of the dimensions in the

problem solving process. For example, in Figure 4,

in the interval 61 to 101, there is a predominance of

descriptive messages, while after message 151, the

prescriptive messages predominate.

Figures 6 to 8 present the evolution of the time

(Figure 6), invention (Figure 7), and mode (Figure 8)

dimensions during the problem solving process in

the Yahoo! Answers activity. Each interaction

analyzed in the first step consisted in average of 8

illocutionary acts. A total of 318 illocutionary acts

(contained in 39 interactions) were identified and

plotted on the horizontal axis of Figures 6 to 8. The

vertical axis represents the respective dimension

values attributed to each communication message

using the form of Figure 1. The dashed lines in

Figures 6 to 8 represent the polynomial trend lines of

each dimension.

3.3 Possibilities and Needs for Novel

Interactive Mechanisms

When cross-referencing the function analysis with

the content analysis, aspects regarding the problem

solving processes can be observed.

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

Figure 3

Rede.

Figure 4:

Rede.

Figure 5

Rede.

Figure 6

Answers.

Figure 7:

Answers.

Time dimen

Invention di

Mode dime

: Time dime

Invention di

ion evolution

ension evoluti

sion evolutio

ension evoluti

graph of Vil

n graph of Vi

graph of Vil

graph of Y

n graph of Y

a na

hoo!

Fig

dis

fiel

cas

dis

top

the

fun



dis

ass

the

esp

dis



aff

cas



spe

(fu

sol

the



sol

(

sol

for



re 8: Mode

wers.

The content

ussed issue

instance, a

d of Figure 1

under dis

ussion, there

cs. After de

were cross

ction analys

ing this third

In Vila na

ussion, the

rtions, some

case disc

ecially intere

recover por

ussion and p

In Vila na R

ominated by

ctive) and

otative) rega

solution. T

port the ide

alternatives

In Vila na R

cialists focus

ure, prescrip

tion is impo

solution itsel

In Yahoo A

tion proces

itional data

ted the q

criptive, de

sages with

sent, descri

tion process

someone w

n some ca

ary questio

In Yahoo

cess focuse

imension evo

analysis li

(from the

marked in

, the speaker

ussion. At

is a single o

ermining th

referenced

s. Some m

tep of the an

Rede, at t

specialists

valuations a

ssion itself.

ting, for exa

ions of di

anning.

de, the middl

valuations (p

ssertions (pr

ding facts a

is part of the

ification an

hat support t

de, at the en

d on a synt

ion, denotat

tant to visua

swers!, at

, messages

bout the cas

estion (i.e.

otative), th

assertions

tive, denotat

s is very i

o wants to

e details ar

nswers!, the

rese

lution graph

ks message

bserved aff

the object/a

was talking

each mome

at most few

se “dominan

ith the grap

jor aspects

alysis were:

e beginnin

in educati

d few requ

These as

ple, when

logs that

part of the

resent, descri

sent, descri

d alternativ

dialog can p

visualizatio

e solution.

of the discu

esis of a sol

ve). This

ize the form

the beginnin

ocused on r

from the pe

, proposals:

is was foll

regarding

ive). This p

portant, for

understand

not prese

middle

ing soluti

f Yahoo!

to the

rdances).

ffordance

about the

t of the

dominant

topics”,

s of the

observed

of the

n made

sts about

ects are

sers want

eal with

iscussion

tion and

s for the

otentially

of facts

ssion, the

tion plan

of the

lation of

of the

questing

son who

future,

wed by

he case

rt of the

example,

the case,

t in the

of the

n plans

InteractingwithDynamicSocialKnowledge-RevealingChallengesthroughanAnalysisofPragmaticAspectsofProblem

Solving

(Special Education Pedagogic plans). Some plan

proposals (future, descriptive, denotative) contained

many items, and followed a structured format. This

part of the processes is important for someone who

wants to visualize the concurrent proposals, for

example.

 In Yahoo Answers!, the end of the process

focused on assertions (present, descriptive,

denotative) and valuations (present, descriptive,

affective) concerning the case and solutions. This

was followed (at the very end) by the prescription

(future, prescriptive, denotative) of a final plan. This

part of the solution is important to visualize the final

plan, as well as the valuations about the case and

alternatives (e.g., some answers valuated the role of

the family and the school in the described case, and

proposed alternatives to work with this aspect).

 It was possible to visualize blocks of messages

with the same (or close) values, e.g., messages 11-17

and 71-77 in Figure 3. Many blocks occurred

simultaneously in more than one dimension and

were related to the same affordance and/or were

performed as answers to the same message. More

data is required to understand the significance of

these blocks for a problem solving process.

However, these blocks are likely to represent

correlated messages and once identified, they could

be explored, for instance, by recovery and

visualization techniques.

 By including the pragmatic analysis, a more

refined classification of the messages is possible.

For instance, it is possible to visualize and recover

sequences of messages that valuate or judge one

specific alternative.

 A more refined analysis of users’ participation is

possible. One possibility is to differentiate when a

user requested information about one specific

technique to solve a problem, or valuated

alternatives, or did a lot of assertions about the

techniques. This is especially important in social

networks when we are looking for someone that has

experience with a specific problem. Thus, it is

possible to identify someone who commented about

one topic, and also what the declared intention was

(proposal, assertion, etc.).

 The identification of the illocutions can also be

used as a parameter in the syntactic and semantic

disambiguation process. For instance, the word

“who” could be an interrogative pronoun in a

request (e.g., who did that), but also a relative

pronoun in an assertion (e.g., that’s the guy who did

this).

 Palinodes and contritions were the least frequent

in the processes of the analyzed cases. Nevertheless,

they are extremely important in some situations. In

the Vila na Rede case, one specialist proposed a

synthesis of future actions, however in the

subsequent messages she apologized for a mistake in

one of the proposed actions.

 The analysis also revealed that there were

assertions, valuations and proposals that were not

related to the problem in focus, such as greetings

and messages about the use of the computational

systems. The identification of such messages might

enable to filter, or even highlight the messages

associated with the “core” problem (substantial

messages). Anther possibility is to filter

“transversal” issues, such as the use of the

functionalities of the systems (communication and

control messages).

 Although comparisons between Vila na Rede and

Yahoo Answers! are out of the scope of this paper,

the pragmatic analysis also revealed important

aspects regarding what forms of conversation or

dialogue are supported in each system. For example,

in Vila na Rede, questions and comments are

displayed in a hierarchical structure. This resulted in

less illocutions per interaction/post and more diverse

illocutions types, while the Yahoo Answers!

discussion was predominated by proposals to solve

the posted question in long post messages that

contained a high number of illocutions.

Starting from these possibilities we extracted

requirements for the construction of features in

SNSs that deal with and take advantage of the

dynamic pragmatic aspects. A major requirement is

the possibility for the users to visualize, explore and

make sense of the dynamism of the pragmatic

aspects of problem solving.

Users should be able to recover or filter parts of

the solution processes using adequate parameters

and be able to interact with the results in a useful

way. For illustration purposes, consider the

following hypothetic scenario: Someone wants to

know what are the previous solutions for a specific

problem discussed in the SNS, including, for

example, the rationale and the evolution of the

solution, who gave opinions and valuations, when,

what were the alternatives and why they were not

adopted (valuations about discarded alternatives),

etc. How do users specify what they need; their

information requirement? Should they directly

specify the illocution classification, e.g., valuations

about extra-scholar activities to support deaf

students in the last 2 months? How to visualize the

results? How to explore the results (e.g., select a

term to show who was referred and when in the

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

discussions)?

The questions presented are a small subset of

possible issues that need to be explored. Most of

these issues demand a multidisciplinary approach,

which may involve techniques, methods and theories

of various research fields.

4 CONCEPTUALIZING

DYNAMIC KNOWLEDGE

In this section, we present a preliminary research

framework for conceptualizing pragmatic aspects of

dynamic knowledge. This framework is the result of

an exploratory investigation of areas, methods and

technologies related to dynamic aspects of

knowledge in the SocWeb. The scope of this section

is not to provide a well-defined process or an

exhaustive list of methods that can be directly

translated into a solution. Rather, the framework

presented here is intended to point out areas,

technologies, and needs for deeper investigation.

Figure 9 presents the three perspectives of the

proposed framework. Examples of candidate areas

and technologies are shown for each perspective.

The Figure shows that there is a significant overlap

of the perspectives. Each item is classified according

to the perspective where it is expected to contribute

most. For example, in our proposed framework,

SemWeb technologies are most relevant for the

computational perspective, although they also might

inform the conceptual or interactive perspective.

In the interactive perspective, two major

challenges to be addressed are: 1. how to provide

alternatives of interaction design to enable the user

to recover, explore and manipulate data and models

regarding pragmatic aspects, and 2. how to explore

pragmatic aspects to provide better interfaces when

exploring knowledge considering the temporal

aspects. In this sense, new interaction design

techniques have to be investigated.

IV and KV techniques are particularly important

for providing means to visualize huge amounts of

complex information (interactive perspective).

However, it is necessary to articulate these

techniques with theories and novel frameworks for

understanding knowledge (conceptual perspective).

The approaches so far explored in the literature are

not able to accurately answer the question of

whether the knowledge structures used really make

sense for the target users.

Another issue to be addressed in the interactive

perspective is how to capture the users’ declared

intentions. In some systems, users may declare their

intentions through interactive interfaces, thus one of

the main difficulties here is how to design

appropriate interaction alternatives according to the

context.

Participatory Design practices (and user-centered

design) may provide alternatives of interaction with

users. By working with users at design time, it is

possible to determine how users make sense of the

representations, and what the appropriate design

options are, for example, regarding searching,

filtering, visualizing and exploring information

using pragmatic dimensions similar to the ones

presented in section 3.

In the conceptual perspective, one of the major

questions is how to employ adequate theories that

support modeling the conceptual aspects of

pragmatic knowledge. These models have to

guarantee conceptual consistence and provide

theoretical grounding to what is implemented in the

computational models (computational perspective),

and presented to users (interactive perspective).

Semiotics can provide us with theories and a

conceptual basis for future works (conceptual

perspective). For example, the Norm Analysis

Method (NAM) (Liu, 2000) provides a systematic

way to understand and model the human behavior in

society. However, some research issues have to be

addressed, such as practical applications that provide

better interactive and computational models, and the

question of how to deal and model norms changes

and evolution.

In the proposed framework, SAT and LAP are

part of the conceptual background regarding the

analysis of pragmatics. Methods associated with

SAT and LAP might be explored, and they may

contribute to the development of methods and

techniques towards a (semi-)automatic content

analysis. The propositional attitudes (i.e., the effect

of the illocutionary acts), which were out of the

scope of this paper, should be investigated through

the analysis of the perlocutionary acts. Moreover,

theories developed in sociology and anthropology,

especially those applied to Social Network Analysis

(SNA) may complement the conceptual basis.

In the computational perspective, two of the

major issues to be addressed are: how to produce

computer interpretable representations of the

pragmatics, and how to automate the construction of

these models to provide scalable solutions over time.

SemWeb technologies can be a starting point to

the implementation of such solutions. For instance,

concepts of Web Ontology Language (OWL) may

have to be adapted or extended to deal with the

pragmatic aspects. Moreover, in this context, the KE

InteractingwithDynamicSocialKnowledge-RevealingChallengesthroughanAnalysisofPragmaticAspectsofProblem

Solving

field has produced a set of methods and techniques

to computationally deal with domain evolution

through the use of SemWeb ontologies. However,

research in this field is required to cope with the

evolution of concepts and models related to

pragmatics.

Figure 9: Areas and Techniques in the three dimensions.

Text mining techniques and tools are also

particularly important for providing scalable

solutions that include models or ontologies modeled

from large volumes of content. It is not feasible to

analyze thousands or millions of documents to

produce ontologies. Text mining technologies can

utilize statistics, graphs, ranks and other

representations that support the knowledge engineer

in the construction of complex models, like

ontologies, conceptual graphs and mind maps.

However, compared to modeling exclusively based

on semantics, pragmatic aspects introduce additional

complexity that need to be faced.

5 CONCLUSIONS

Collaborative problem solving in SNS can

potentially produce a huge amount of complex

information. This information is of great value, for

example to users in the context of continuous

learning who want to recover information about past

problems. However, the correct recovery and

understanding of this information is influenced by

pragmatic aspects that evolve over time. These

aspects have a dynamic nature, and involve the

examination of the intentions in the discussions,

which is a complex issue.

In this paper, we presented a study of problem

solving processes carried out in SNSs. The results

pointed out new interaction possibilities, but also

requirements for future research in related areas. The

research needs were translated into a preliminary

research framework structured in three perspectives:

interactive, conceptual and computational. This

framework indicates that novel approaches are

necessary for dealing with the dynamic aspects of

knowledge in these systems. Moreover, we stressed

the importance of providing solutions that are able to

conceptually and computationally articulate

semantics and pragmatics as an integrated process

over time. These solutions may demand, for

example, studies in Organizational Semiotics and

new KE and KV techniques.

We hope that this preliminary research

framework contributes with ideas towards

facilitating meaningful Web-mediated interactions.

Further work involves gathering more empirical data

to test and refine the analysis method, coping with

the challenges described in the previous sections,

and investigating novel KV and KE methods based

on the background disciplines.

ACKNOWLEDGEMENTS

This work is partially funded by CNPq

(#560044/2010-0, #141058/2010-2), by Proesp/

CAPES (#23038.01457/2009-11) and CAPES (#01-

P-08503/2008). The authors also thank colleagues

from InterHAD for insightful discussions.

REFERENCES

Austin, J. L., (1962). How to Do Things With Words.

Oxford University Press: Oxford, England.

Avery, J., and Yearwood, J., (2003). Dowl: A dynamic

ontology language. Proceedings of the IADIS

International Conference WWW/Internet 2003,

Algarve, Portugal.

Berners-Lee, T., Hendler, J., and Lassila, O., (2001). The

semantic web. Scientific American, 284(5), 34-43.

Burkhard, R., (2004). Learning from Architects: The

Difference between Knowledge Visualization and

Information Visualization. In Eight International

Conference on Information Visualization (IV04),

London, July

Hornung, H. and Baranauskas, M. C. C., (2011) Towards a

Conceptual Framework for Interaction Design for the

Pragmatic Web. In 14th International Conference on

Human-Computer Interaction, 2011. Heidelberg:

Springer, 72-81.

Cordeiro, J. and Filipe, J., (2003). Language Action

Perspective, Organizational Semiotics and the Theory

of Organized Activity – A Comparison. In

Proceedings of the workshop DEMO. Tilburg, the

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

Netherlands.

de Moor, A. and van den Heuvel, W.-J., (2004). Web

Service Selection in Virtual Communities.

Proceedings of the 37th Annual Hawaii International

Conference on System Sciences, Washington, DC,

USA. IEEE Computer Society. 10p.

Eppler, M., and Burkhard R., (2004). Knowledge

Visualization. Towards a New Discipline and its

Fields of Application. In Working Paper of Net

Academy on Knowledge Media, St. Gallen.

Flouris, G., Manakanatas, D., Kondylakis, H.,

Plexousakis, D., and Antoniou, G., (2007). Ontology

change: Classification and survey. The Knowledge

Engineering Review, 23(2), 117-152.

Gibson, J. J., (1968) The Ecological Approach to Visual

Perception. Houghton Miffin Company, Boston,

Massachusetts.

Goldkuhl, G., and Lyytinen, K., (1982). A Language

Action View of Information Systems, in Proceedings

of the 3rd International Conference on Information

Systems, Ann Arbor, MI, pp. 13-29.

Gruber, T., 2008. Collective knowledge systems: Where

the Social Web meets the Semantic Web. Journal of

Web Semantics, 6(1), pp.4-13.

Hendler, J. and Berners-Lee, T., (2010). From the

Semantic Web to Social Machines: A research

challenge for AI on the World Wide Web. Artificial

Intelligence. Elsevier. 174, 156-161

Hoetzlein, R., (2007). The Organization of Human

Knowedge: Systems for Interdisciplinary Research.

Master's Thesis, University of California Santa

Barbara, USA.

Keller, T. and Tergan, S., (2005). Visualizing Knowledge

and Information: An Introduction. In S.-O. Tergan and

T. Keller (Eds.): Knowledge and Information

Visualization, LNCS 3426, pp. 1 – 23.

Klein, M., and Fensel, D., (2001). Ontology versioning for

the semantic web. In Proceedings of the International

Semantic Web Working Symposium (SWWS), Stanford

University, California, USA.

Kuß A., Prohaska S., and Rybak J., (2009). Using

ontologies for the visualization of hierarchical

neuroanatomical structures. In 2nd INCF Congress of

Neuroinformatics.

Leenheer, P. and Meersman, R., (2007). Towards

Community-Based Evolution of Knowledge-Intensive

Systems. In R. Meersman and Z. Tari et al. (Eds.):

OTM2007, Part I, LNCS 4803, 989–1006.

Liu, K., (2000). Semiotics in Information Systems

Engineering. Cambridge University Press, New York,

NY, USA.

Liu, K., (2009). Pragmatic Computing – A Semiotic

Perspective to Web Services. In Filipe, J. and Obaidat,

M. S., editors, E-Business and Telecommunications.

ICETE 2007, Revised Selected Papers,

Communications in Computer and Information

Science, Springer, Berlin, Germany, 23,3–15.

Maseri, W., Mohamad, W., Embong, A., and Mohd Zain,

J., (2007). Improve Knowledge Visualization through

an Interactive Graph-based Dashboard System with

Key Performance Indicator: A Case Study of

University Dashboard for Higher Education. In

National Conference on Software Engineering and

Computer Systems 2007 (NaCES 2007), Legend

Resort Kuantan.

Morris ,C. W., (1938). Foundations of the theory of signs.

International Encyclopedia of Unified Science, (vol. 1

pp. 2). University of Chicago Press, Chicago.

Pampalk, E., Goebl, W., and Widmer, G. (2003).

Visualizing Changes in the Structure of Data for

Exploratory Feature Selection. In SIGKDD ’03,

Washington, DC, USA.

Paschke, A., Boley, H., Kozlenkov, A., and Craig, B.

(2007). Rule Responder: RuleML-Based Agents for

Distributed Collaboration on the Pragmatic Web. In

ICPW ’07: Proceedings of the 2nd International

Conference on Pragmatic Web, 17–28, New York,

NY, USA. ACM.

Schoop, M., de Moor, A., and Dietz, J. L. G., (2006). The

Pragmatic Web: A Manifesto. Communications of the

ACM, 49(5):75–76.

Searle, J. R., (1976). A Classification of Illocutionary

Acts, In Language in Society, 5 (1), 1-23

Singh, M. P. (2002). The Pragmatic Web. IEEE Internet

Computing, 6(3):4–5.

Stamper, R., (1996), Signs, Information and Systems, in

B. Holmqnist, et. Al. (Eds) Signs of Work Semiotics

Information Processing in Organisations, Walter de

Gruyter, N. Y.

Stojanovic, L., (2004). Methods and tools for ontology

evolution. PhD. Thesis. University of Karlsruhe,

Universitat Karlsruhe (TH), Institut AIFB, D-76128

Karlsruhe.

Winograd, T. and Flores, F., (1987). Understanding

Computers and Cognition: A New Foundation for

Design. Addison-Wesley Longman Publishing Co.,

Inc., Boston, MA, USA.

InteractingwithDynamicSocialKnowledge-RevealingChallengesthroughanAnalysisofPragmaticAspectsofProblem

Solving