Evolution of an Online Forum for Knowledge Management Professionals: A Language Game Analysis


  • Anne-Laure Fayard,

    Corresponding author
    1. Assistant Professor of Technology Management at INSEAD. Her research interests concern socio-technical practices, technology-mediated communication, distance collaboration and space.
    • Address: INSEAD, Bd de Constance, 77305 Fontainebleau, France

      Address: Fuqua School of Business, Duke University, Box 90120, Durham, NC 27708-0120 USA

    Search for more papers by this author
  • Gerardine DeSanctis

    Corresponding author
    1. Thomas F. Keller Professor of Business Administration at Duke University. She studies organization design and electronic communication in organizational contexts, including distributed teams and online communities. She serves as Associate Editor for Management Science, is a member of the editorial board of the Journal of Organizational Behavior, and is on the advisory board for Information Systems Research.
    • Address: INSEAD, Bd de Constance, 77305 Fontainebleau, France

      Address: Fuqua School of Business, Duke University, Box 90120, Durham, NC 27708-0120 USA

    Search for more papers by this author


Using Wittgenstein's (1953, 1969) language game framework, we analyze the evolution of an online discussion forum of information systems professionals in India. Three major dimensions of the forum's language game are documented: roles, social identity, and linguistic style. Our analysis shows how rules of linguistic interaction emerge around these three dimensions and help to shape social order in the forum over time. The forum's developmental pattern mirrors phases of small group development, including forming, norming, performing, and storming. Based on our findings, we posit general inferences regarding how professional groups with loose social ties can develop sustainable online discussion forums.


In today's fast-paced, global business environment online discussion forums present a potentially valuable venue for interaction among busy, working professionals. To this end, there is an emerging genre of online forums geared toward meeting the development needs of professionals with common interests and complementary knowledge needs (Gray & Tatar, 2004; Herring, 2004). Sometimes referred to as professional development forums, these are a modern incarnation of listserv discussion groups initiated by academic scholars in the early days of the Internet (e.g., Hert, 1997). Today these forums are vast, covering innumerable specialty topics for a wide array of professional groups. People participate in these forums voluntarily and intermittently for the purpose of acquiring information, skills, and other resources relevant to their work interests (Lakhani & von Hippel, 2000). The forums provide an alternative to educational courses or dues-paying associations that require face-to-face encounters, bounded times of interaction or other formalities and obligations. But the forums generally produce no tangible products; nor do they provide the participants with tangible rewards or outcomes. Online participation is engaged via a shared professional focus and an opportunity to learn from colleagues. As such, attracting contributors and sustaining the life of the forum is an ongoing challenge.

According to Herring (2004, p. 345), professional development forums can be considered “successful” if participation is “continuous and self-sustaining, unlike course-based CMC, which is task-focused and temporally bounded.” How does a forum realize this success? What developmental stages characterize a professional forum that emerges and thrives in the vast sea of Internet discussion groups? What mechanisms operate to shape the forum's life and keep it intact over time? Our paper explores these questions through an in-depth study of one professional development forum. The forum was convened by one person in August of 2000 and remains active today. There is no fee to join. The forum does no advertising and has no formal ties to any organization. There is no formal management structure. Anyone can join, and the contents are fully public. It is not a particularly large forum; people come and go, with the majority of visitors contributing one message. Still, the forum survives and has kept its focus and momentum for over four years. We examine this forum with an eye toward documenting its development, thereby learning from its success. We regard the forum as successful due to its self-sustaining participation over time (per Herring, 2004). Our goal is to derive practice guidelines that might aid others interested in developing such professional forums and, more importantly, to formulate theoretical propositions that can drive further study of loose, professional groups online.

Wittgenstein's (1953, 1969) language game framework frames our analysis. Language games are routines of discourse that constitute everyday life and give form, function and meaning to human interaction. Language games are “spheres of activity…specialized forms of discourse” engaged in by a community (Astley & Zammuto, 1992, p. 444). Games, as social routines, form as patterns of speech that are produced via the constant alignment among speakers as they generate discussion around ideas or objects of common interest. Wittgenstein's framework has been used by scholars to understand development and action within formal organizations (e.g., Astley & Zammuto, 1992; Boland & Tenkasi, 1995; Rindova, Becerra, & Contardo, 2004); however, its applicability to online groups has not yet been explored. Because discourse constitutes the essence of discussion forums, Wittgenstein's framework would seem to offer a particularly powerful approach for studying how people who otherwise do not know one another can enter, participate and exit from an online social setting with ready ease.

Researchers are increasingly interested in taking a developmental approach toward understanding the social forces that give rise to sustainable online forums, and our study aims to contribute to this growing literature. For example, Ahuja and Galvin (2003) have documented how newcomers engage in more information seeking than established contributors and gradually become more succinct in their message content over time. Butler (2001) has documented how increased size of an online forum tends both to attract members and increase turnover; and Galagher, Sproull, and Kiesler (1998) have described how contributors establish legitimacy and authority in online discussion groups. Hesse, Sproull, Kiesler, and Walsh (1993) have examined the structure of scientific networks, and Kozinets (1999) has described how consumer-oriented forums develop cultural cohesion over time. We complement these efforts by undertaking a holistic analysis of a forum's development through the lens of its evolving discourse. Whereas other scholars have studied forum development using a range of attributes, such as media space design and user actions within these spaces (e.g., Mynatt, O'Day, Adler, & Ito, 1998), our focus is exclusively on user discourse. The forum's discourse provides a basis for understanding the process by which roles, routines, and social identity develop over time (Robichaud, Giroux, & Taylor, 2004).

The setting for our study is an online discussion forum of information systems professionals in India. India's business environment is highly competitive for information technology skills and demands continuous learning on the part of IT professionals who hail from many locations and cultures within this vast, pluralistic society (Dhesi, 2002, Kamath, Kohli, Shenoy, Kumar, Nayak, & Kuppuswamy, 2003; Pineda & Whitehead, 1997); thus India provides an interesting context for our study. The website for the forum, which we refer to as KMforum,1 states that the forum “seeks to integrate the various emergent thoughts in the field of Knowledge Management in a uniquely Indian way.” Again, following Herring (2004), we regard the forum as successful in its growth (from three to 485 contributors since its founding in 2000) and sustainability (of four years, to date). The purpose of our study is to document the forum's early developmental process. We track the development of this forum in detail for the first fifteen months of its life with the goal of understanding how the group forms and grows, thereby setting the stage for its long-term survival.

Our article proceeds as follows. First, we review some of the known properties of online discussion forums and the challenges they confront. Second, we explain the language game framework as it applies to our study. We describe how we collected and analyzed the data and provide descriptive information about our target of study, KMforum. Third, we present a detailed case study of KMforum's development, charting three dimensions of its language game: roles, social identity, and linguistic style. We integrate the results of the case study to highlight the major phases of forum development. We conclude with implications of our findings for design and management of professional development forums. We note our study's limitations and offer possible directions for further research.

Online Forums and their Language Games

Challenges of Online Forums

Researchers have examined a range of phenomena related to online discussion groups, including member contribution patterns, churn, sustainability, and motivations to contribute (e.g., Constant, Sproull & Kiesler, 1995; Rheingold, 1993; Wasko & Faraj, 2005), empathy expression (Preece, 2000), deception, conflict, and social control (see Smith & Kollock, 1999); and how work-oriented groups differ in their contribution patterns from more entertainment-oriented groups (Finholt & Sproull, 1990). Collectively, the research to date shows that intimate relationships and development of community are possible online, and that online forums can be productive and sustainable. But the research also reveals that online groups confront developmental obstacles that can hamper information sharing and lead to a forum's quick demise. These include lack of familiarity among individuals, distinctive thought worlds, disparities in verbal skill, differing cultures, status differences, and challenges associated with physical distance (e.g., Bechky, 1999; Earley & Gibson, 2002; Gruenfeld, Mannix, Williams, & Neale, 1996). Meaningful knowledge sharing among professionals often requires situated understanding (Bechky, 2003), a process that is difficult in the asynchronous, low-context setting of online text exchange. Smith (1999) has tracked thousands of Usenet groups and observed, “In many cases a newsgroup is a barren or cacophonous space” (p. 201). He reports that a fifth of groups are entirely empty, and most attract fewer than 50 contributors and produce fewer than 20 messages per month.

For these and other reasons, a number of researchers have argued that many forums fail to function as “communities” and instead operate as very loose networks with (at best) weak social ties (e.g., Jones, 1997; Jones, Ravid, & Rafaeli, 2004). Although case studies of tightly-knit online groups are of interest, more studies are sorely needed of loose, fragile forums that make up the bulk of the online landscape. In these professional discussion forums, ongoing contributions both constitute and drive the process of survival. Most of these forums are not particularly large in size, and sustaining ongoing contributions is an enormous challenge (Bechky, 2003). Visitors to a forum must be able to interpret text comfortably, find meaningful content, and know how to formulate messages that will interest others. Understanding who is saying what and how the group conversation is unfolding within a vast array of messages can be difficult, especially for newcomers; and with many people coming and going, a forum can lack a sense of continuity and coherence. We know that online conversations can have significant amounts of socio-emotional content (Rice & Love, 1987; Walther, 1992) and that the process of reading and writing in online networks can lead to a social construction of place and sense making (Duin, 1991; Harasim, 1993). Facilitators and experts can play important roles in stimulating conversation (Gray & Tatar, 2004). But little is known about the evolutionary dynamics of conversations and how these interweave to produce an ongoing, self-sustaining professional development forum.

Language Game Framework

Language games are patterns of group discourse that surround a particular object, idea, class, or other meaningful entity (Wittgenstein, 1969). For example, there are language games of colors, emotions, law, professions, work groups, and business organizations (Aldridge, 1992; Astley & Zammuto, 1992; Barge, 1994; Myrsiades, 1998; Smith, 1997). The notion of ‘game’ highlights the idea that communication is a system of representation and action on the part of the players (speakers). A game forms as speakers generate discussion routines. The language game includes the rules of social engagement related to the entity; these set the stage for interpersonal relationships and mutual understanding. We can describe the language game of an online forum in terms of the players and their routines of speech that relate to the online forum.

Although it would be possible to examine the language game surrounding a forum's topical focus (e.g., knowledge management), our interest is in the language game of the forum as an organizing entity. Thus Wittgenstein (1969) analyzed the language game of builders (and not of the buildings), and Grover (1993) analyzed the language game of project management (and not of software or other products of project management). Our interest is in the language game of the forum itself, that is, of how speakers converse in the professional development arena-what is acceptable or unacceptable, how roles are enacted, whether a common sense of “us” emerges among the players, the ways of speaking and acting that provide structure and facilitate common understanding. In order to unveil this language game (which is enacted but not fully visible, even to the participants), we can examine discourse surrounding various dimensions of organizing. For the current study we focus on three such dimensions: roles, social identity, and a patterned linguistic style. We attend to these dimensions because they are indicators of “organization” (Scott, 1995). We look for patterns of speech related to each dimension with the aim of then integrating these into a comprehensive understanding of the forum's overall language game.

Games involve different players who take on roles which, in turn, help to define the types of interactions and behaviors that occur in a group or organization. Roles reflect task responsibility and authority (Aldridge, 1992); they are the orchestrators of conversation and integral to the creation of meaning and knowledge sharing (Kogut & Zander, 1996). Important roles in an online professional forum include facilitator, moderator, and guru (or expert) (Gray & Tatar, 2004). We examine expression of social identity within the discourse to assess coherence among the players and the distinctiveness of the online language game from other language games of which the forum may be a part, such as the geographic locale or the larger professional community. Identity occurs as speakers define themselves in relation to the group. Identity can be found in surface language features that convey intimacy with others, such as reference to “we,”“us,” or “our group” (Ashforth & Mael, 1989; Weiner & Mehrabian, 1968) and in references to a common, larger community (“our KM professional community”). Expressions of identity reflect rules for coordination and learning in organizations and online groups (Finholt & Sproull, 1990; Kogut & Zander, 1996). Finally, there are language games associated with managing relationships—providing introductions, closings, and feedback. We refer to this dimension as the linguistic style of the online forum. Linguistic style is evident in the use of stylized vocabulary and speech routines that emerge across all or a meaningful subset of a group's messages (see, e.g., Mynatt et al., 1998). Linguistic style refers to the governance of speech inside the forum, defining what is acceptable or expected of players as they participate in the game (Grover, 1993; Myrsiades, 1998; Topp, 2000).

We will separately document language games related to roles, social identity, and linguistic style in the forum for purposes of analysis. However, we emphasize that these are intertwined and overlapping and, in fact, cannot be fully separated one from the other. As Robichaud et al. (2004) recently explained, there is recursive embedding of speech patterns in organizational life, from words and symbols within messages, to streams of messages, to overall metatext of the forum discourse. We can separate dimensions conceptually, but, in fact, they dynamically evolve and influence one another. Further, we emphasize that our focus on three dimensions will not be sufficient to reveal the complete language game of the forum. Roles, social identity, and linguistic style are merely three of the many dimensions of organizing that may occur within a forum's discourse. We choose these in the interest of parsimony and because of their potential importance to online forum life. By examining the evolution of discourse related to roles, social identity, and linguistic style, we can get a sense of the whole, that is, the overall language game of the online forum.


KMforum Attributes

We selected KMforum for study for several reasons. First, it is a professional development forum that addresses a topic of interest to many IT professionals. Second, it was founded in a locale where interest in participating has the potential to be high (India). At the same time, there are many competing forums on the same topic available via the Internet (DeSanctis, Fayard, Roach, & Jiang, 2003), and, in the face of competition, the forum has survived with a steady pace of visitors since its founding. A preliminary view of the contents suggested that the forum is made up of a loose collection of professionals—not a tight community—and meets the description of professional development forum as defined early in this paper. Third, the Indian context of the forum provides a setting where contributors are highly distributed and heterogeneous in their backgrounds. At the same time, they share a common national identity. These are competing forces, both for and against, formation of a coherent language game within the forum. Finally, the complete contents of the forum are publicly available. The language of the forum is English, and the topical content is not highly technical, making analysis by the authors a realistic proposition.

KMforum was founded in August 2000 by a senior executive in a computer services company. He indicated at the start that he was founding the forum in conjunction with two other people. The goal of the forum as articulated in the founding message (see Figure 1) and the banner website description (“to integrate the various emergent thoughts in the field of Knowledge Management in a uniquely Indian way”) are broad and ambitious. Most participants in the forum are located in India (e.g., Bangalore, Hyderabad, Secunderabad, New Delhi, Calcuta, Jamshedpur, Lucknow), but contributors also are located outside of India (e.g., Bahrain, United States). Anyone with access to the Internet can read and post messages to KMforum. There is no fee to join. Members must go to the forum website to read and post messages, which include text and limited graphics capability. There is no stated limit to message length, content, or frequency of posting. The forum has no formally stated rules or requirements for participation.

Figure 1.

Founding message for KMforum

Contributors to KMforum come from a variety of organizations, including local companies, universities, government, consultancies, and large multinationals (e.g., HCL Perot Systems, Ernst & Young, Deutsche Bank AG). An analysis of the first 20 messages indicated the following types of contributors: managers/administers (45%), technical/systems specialists (25%), researchers (5%), consultants (5%), does not say (20%). Contributors discuss a range of matters related to knowledge management, including KM concepts and principles, technical standards, systems implementation, applications to specific work settings, and career issues.

Our analysis focused on KMforum's founding period and initial growth, until it reached a steady state of contributions. In all, we examined 15 months of message content. During this period there were 527 messages posted by a total of 123 contributors. Like other professional development forums, participation is characterized by a core of regular activity complemented by significant variation in contributor behavior over time (Gray & Tatar, 2004; Herring, 2004). We segmented the data into five three-month blocks so that we could identify and compare patterns within and between these blocks of time. Figure 2 shows the message posting pattern for the forum over time, and Table 1 summarizes the types of contributors and their posting patterns. Note that although the absolute number of messages per month declines over time, the slope of the contribution curve flattens, core contributors persist in their presence, and visitors and newcomers continue to join the forum as the forum evolves.

Figure 2.

Number of messages posted for the five 3-month periods of the study

Table 1. Types of contributors to the forum and their posting patterns
Contributor typesDefinitionNumber of contributorsPercentAverage number of messages posted (and standard deviation)
Core groupContribute T1-T5 or T2-T565%45 (49)
Highly active membersContribute for more than two of the three-month periods1613%4.9 (2.3)
Less active membersContribute for one or two consecutive periods only3125%3.3 (2.65)
Visitors (drop-ins)Post only one message T1-T54940%1 (0)
NewcomersContribute in T52117%1.3 (0-3)

Analytic Approach

Wittgenstein (1953) argues that words are like tools, and just as tools serve different functions, so linguistic expressions serve many functions. “Linguistic flexibility” means that language games cannot be understood in terms of strict coding schemes. Instead, deciphering language games requires closely examining and describing language in use. With this in mind, we undertook an in-depth longitudinal language-focused discourse analysis (per Herring, 2004) segmented first by time (all messages in the intervals) and secondarily by dimension (roles, social identity, and linguistic style). This approach allowed us to examine each dimension in context yet at the same track the developmental patterns separately. Our analysis proceeded as follows.

We employed qualitative data analysis methods (Eisenhardt, 1989; Miles & Huberman, 1984), generating detailed field notes for each of the five three-month time blocks, attending to the three dimensions of roles, social identity, and linguistic style. Where possible, we noted specific messages and message sequences that corresponded to each dimension. This step required multiple readings of the complete text of the all messages that composed each time block. On each pass, the notes were clarified and further expounded. The following sets of questions guided our documentation of each dimension:


What roles are discussed, and how are they discussed? Who enacts roles? What does the founder do? Is there a moderator (administrator of the forum) or facilitators (persons who direct discussion content)? What other roles are played? What are the rules of each role, and how do the contributors respond to role behaviors? Are there references to task responsibilities or organizational structure (relative importance of roles)?

Social Identity

Is there discussion of building a group? Do contributors use “we” or “us” to refer to the forum participants rather than “you” or “I”? Do they use the forum name, or refer to the group (“Hi, all!”“Dear friends,”“our group”)? To what kind of identity does the collective language refer: organizational identity (the KMforum or its contributors), KM community identity (the larger professional collective that shares the same interest), or geographic identity (India, or a common region or other geographic locale that is shared by forum contributors)? Is there discussion of the goal and rules of the forum?

Linguistic Style

How do contributors address the other participants (e.g., “Hi, Meg” or “Dear Sir”)? Do participants sign their messages or use signature files? What kind of signatures are they using (e.g., none, “John” or “J. F.”)? What is the tone of speech? Are messages matter-of-fact, or do they include forms of politeness or expression of positive regard toward others? Do messages include reference to prior messages, forwarding, or excerpting of earlier message content? How formal or informal are the messages? To what extent do contributors use punctuation, emoticons, capital letters, parentheses, etc.? Do contributors refer to other messages and link their comments to others' comments, express agreement or disagreement, or seek others' agreement or disagreement with a viewpoint or prior message? Do they express appreciation or acknowledgment (e.g., “Thank you,”“cheers,”“looking forward to…”)?

Next, we summarized the case in synopsis form for each of the five time periods and developed detailed tables (per Miles & Huberman, 1984) documenting the occurrence of discourse related to each of the dimensions. The tables highlighted similarities and differences in language patterns across the five time periods.

Finally, we reviewed the notes, case, and tables to articulate the language games of roles, social identity, and linguistic style, and the overall phases corresponding to their emergence in the forum. The phasic analysis allowed us to identify games that persisted and those that disappeared, and identify language games that were enacted across all time periods but through different communicative practices.

In all of these analyses our approach was exploratory, intended to generate insights into the development of the forum and its evolutionary phases. We now present our case synopsis, comparative findings, and phasic analysis. To protect the anonymity of the contributors, names are substituted with fictitious names in all of these results.


We first present a synopsis of the KMforum case for each of the five time periods (referred to as T1, T2, etc. below). We then describe the language games associated with the three dimensions of roles, social identity, and style. Finally, we describe the developmental phases for the forum's overall language game.

Case Synopsis

T1 (Aug 23 - Nov 2000)

GV, the founder, constantly celebrates growth (e.g., (6)2“we are finally in double digits…”). He also plays a role in stirring up discussions, raising questions, challenging the group to respond, and, at times, summarizing a discussion thread. For the most part GV acts as both moderator and facilitator; however, at one point (68) GV shifts his style and forwards articles and references without commentary, and without greetings. Then at message (85) he reverts back to the facilitator role, commenting: “nice to see this e-group churning rich thoughts.” At (103), he announces that he is not moderating the forum anymore. Yet, this announcement does not change the frequency, tone, or content of his subsequent messages. Although GV is the most active contributor, there are three other highly active contributors and some less active contributors whose messages generate extensive discussion. Participants discuss a wide variety of KM concepts, definitions, technology, and so on. “Hi All, Could you share your practical experience on COPs as how its being implemented in your Company etc?” (117). Most of the messages are fairly long (20-30 lines). Information seekers give detail on their interests and the context of their questions. Those providing information give extensive explanation.

The building of the community is a key topic in T1, with GV's messages often hailing the latest number of participants. There are many references to well-known leading thinkers in KM, some joining the forum, accompanied by enthusiastic welcoming messages, such as, “it gives me a great pride and pleasure to welcome Dr. M.S.” In addition, there are many references to the forum as a group. Newcomers introduce themselves—often indicating that they have been lurking for a while. It is clear from the names, signatures and some explicit geographical references that this forum is based in India. It seems that some people know one other, at least by reputation; however, there are no references to off-line meetings and, hence, no embodiment.

The tone is informal and friendly. Contributors make heavy use of parentheses to embellish their text. This creates an oral style. They also use punctuation (exclamation and question marks) to express emotions. There are quite a few e-mail messages that are forwarded to the group. This pattern could be interpreted as an attempt to build community: “let's share all what we have; let's avoid side discussions.” Contributors also tend to make reference to other messages (“this is in response to S's message…”), legitimizing the speaker's contribution and giving a sense of connection among the discussions.

T2 (Dec - Feb 2001)

GV is still the most frequent contributor and acts as moderator and facilitator. Following an aggressive exchange of messages, a discussion arises on the role and necessity of having a moderator. GV indicates that he is not a moderator anymore, but he nonetheless acts the part by stating rules and advice: (167) “when this e-group was started it was moderated so that such occasions do not occur. When they did not I made the group unmoderated.” Until the end, he more or less keeps moderating the forum. In this segment, PP, a consultant, starts posting messages: “I am interested in developing a network of Indian thinkers” (202); “I would like to dedicate some additional time to this forum” (211). PP posts frequently and in a few instances acts as a moderator. For example, (240) he welcomes a newcomer and states that he would like to start building a KM group in India tied into an agenda that he has defined. PP has a distinctive discourse style; he invokes authority (cf. Galagher et al., 1998), referring constantly to his work and his collaboration with well-known KM names. He does not try to facilitate the discussions or have feedback but instead makes pronouncements.

Along with GV and PP, several contributors post three or four messages in T2. As in T1, participants tend to introduce themselves on first posting and cite their experiences. In T2, there are many references to a common identity as well as some discussions about behaviors that are acceptable or not in the forum. There is a lively discussion on the nature of a healthy discussion, on the role of the moderator, and appropriate behaviors in the forum. A newcomer posts a message indicating that he felt lost and disappointed by the exchanges he read on the forum (152): “My experience over the last 3 months has left me more confused and ignorant than before…” One participant replies to this message with a nasty message, which leads to a discussion about appropriate messages and the definition of the forum. Participants then agree on the importance of being open-minded and polite and the key role of the moderator. One contributor makes a point regarding the difference between the type of discussions to be made online and those that need face-to-face contact. He ends his message with these words (159): “the best use of this forum is to share practical experiences…a group like this serves as solace, sounding board (virtual friend, philosopher and guide).” People use “we” and “us” to refer to the group. At a few points when the discussion slows, several participants comment on the low activity (174, 197, 243). The style remains informal and friendly with many references to the forum and the Indian KM community, with the latter now identified as (202) “a network of Indian thinkers.” Participants still make heavy use of parentheses, message forwarding, and targeted references to other messages. They also continue to provide positive feedback to the other participants.

References to “gurus” (experts) continue. Messages that provide information (e.g., state opinions, point to resources, describe experiences, pass on suggestions) are more frequent than messages that request information (i.e., ask questions). Many messages include excerpts from printed or electronic materials related to KM, cut and paste of articles, URL addresses, etc. There are also general discussions about definitions and concepts, and discussions about practical issues in specific contexts. GV spurs this kind of discussion, and quells conflict, for example:

GV (155): “… if you and other members of this e-group have been thinking that it will be easy to take a few defined concepts and find out what KM is… then I regret to say that this need will not be fulfilled at this forum… I believe that sense making is the actual knowledge creation process for the community…”

T3 (Mar - May 2001)

GV remains the most active contributor, acting as both moderator and facilitator even while, at times, explicitly denying the role. He writes (344): “Dear all, I am not going to be accessing the net or mails over the next three weeks, hence I am making the group unmoderated. Please keep in mind the objectives of the group and if somebody violates them, please do not hesitate to remind the person of the same. Keep the spirit and ethos of the group and take it higher.” PP continues to post long, complex and theoretical messages in the form of pronouncements, without seeking feedback. He is the second most active contributor after GV. He seems to use the forum as a stage to express himself. He posts several aggressive messages arguing with another well-known consultant. Some other participants then start to facilitate, stirring up involvement of others. In addition to GV, there is a steady, core group of five contributors (including PP). Newcomers continue to follow the pattern established in T1 and T2, typically introducing themselves, stating their experience, and saying why they are interested in KM.

As in T1 and T2, contributors regularly refer to the group and its shared history, although use of “we” and “us” declines in this segment. At one point a contributor calls on others to learn more about human resource policies in specific companies, and he generates many replies. Frequent references to “the Indian KM community” continue but, as before, there is no mention of offline interactions. Participants persist in the prior style of using parentheses, forward patterns, and giving positive feedback. The tone remains informal, although some tensions arise as reflected in several aggressive messages and pointed arguments.

T4 (Jun - Aug 2001)

GV is still a key contributor, although he becomes the second most active contributor, after PP, who becomes the most active contributor. GV still takes on the moderator role as well as the facilitator role, stirring up conversations. “Dear all (…) we have touched on this aspect a couple of times superficially in the past (…) what do you think?” He continues his pattern of welcoming newcomers (455). In T1, he praised growth, but in T2 and T3 he stopped referring to the number of contributors. He resumes praising of growth in T4 (455): “In this e-group we behave in spurts and starts. Well, as a small update we are around 300 KM enthusiasts on the group as of today.” The core group members are still active. In addition, other participants take up the facilitator role and try to stir up discussion (e.g., 397, 402, 425, 430), at times with success. There are many references to gurus, either from the Indian KM community (e.g., (455) “welcoming a familiar face in the Indian landscape of KM, Mr. B. Mani”) or the broader KM field (399, 402).

Use of greetings and references to experience persist, but there is more emphasis on the latter than the former in this period. There are fewer newcomers, at least who introduce themselves as such.

In T4, the group's identity seems to be “assumed.” There are still references to the forum, but these are fewer than in T1 and T2. There are more references to the broader Indian identity (446): “but the reality in Indian situation;” (449) “a GKEC ANSI/ISO standards working group is forming in India;” (457) “the real scenario of Indian companies.” There also are explicit references to geography, either locations of participants or locations of some events organized (e.g., (381) “All Mumbai/Pune based members;” announcement of a conference in Bangalore (382 and 385)). Some members refer to face-to-face encounters, such as conferences, meetings of local KM groups (e.g., (381) a group in Mumbai), and individual encounters.

The tone remains friendly, and participants are more intimate and informal in their messages. There is now heavy use of punctuation and emoticons to express emotions, creating an oral style. Participants still forward emails, but they do not explicitly refer to others' contributions as before (only one occurrence in T4). Positive feedback to other members continues, and grows in frequency.

T5 (Sep - Nov 2001)

GV is still active but much less than before. He is now the third most frequent contributor. He still plays facilitator and moderator roles but much less enthusiastically than in earlier periods, especially after an argument segment which occurs in this period; thereafter, GV mainly forwards messages and posts articles.

Early in this period PP is very active and uses the forum as a stage to display an argument he is having with JF, a participant in a different KM forum (491): “I suggest that this discussion between John and his supporters and I and my supporters move to the KM India forum.” The argument segment consists of six postings by PP, retracing his argument with John Frost (JF) (487, 488, 489, 492, 500, 501). The argument with JF was occurring in another online forum (which we will call KM-2forum), but the moderator of that forum stopped posting PP's messages (according to PP). Consequently, PP decided to use KMforum as a platform for his argument. He pastes messages of JF's and other messages from the KM-2forum into the KMforum. At several points PP asks JF to post directly to KMforum, but JF refuses. PP claims that his messages are not posted on KM-2forum while they are on KMforum (491, 500). PP abruptly stops posting after posting a last message on this argument (message 501).

In the argument, there are many references to the KMforum and to a “private” forum, called EI forum. PP posts messages on what constitutes an ideal community, the ideal form of governance, and on the philosophy of KM (487): “the society that adopts…becomes a rule following society that will punish those who have divergent viewpoints simply because these viewpoints do not fit in the social norm (…) Is this type of society that the leadership of KMforum wants?” There are many references to right and wrong behaviors (487): “this is a disservice to the KM community… Now I am aggressive about this, yes. It is because I feel that the KM community must better understand that argumentation of the type championed by John Frost is fundamentally wrong as a KM foundation.” In his last posting, PP asks the other participants to remove John Frost from his official position in the KM-2forum.

PP is nearly the only contributor to post messages during this argument segment, and when others post they don't refer to the argument. Members ignore the argument and become notably silent. It is as if the argument brought up by PP has frozen all the lively discussions in the forum. GV stops stirring up discussions and only provides announcements, often posting articles with no greetings. The period right after the argument corresponds to some sort of recovery phase. There is minimum relationship management after the argument among the regular contributors, that is, little friendly discourse and informality. Life seems to reappear only in the last few messages posted during the period. For example, a student from a Canadian university posts a series of questions near the end of T5. He introduces himself and asks members to answer questions on KM leadership (534). He is very polite, within the traditional style of the forum. This message generates several replies with personal addresses. In a similar vein, a new member posts a message (525) in which she just briefly introduces herself and states her desire for “an interactive and educative association with the KMforum.”

There are several references to KM gurus in T5: (460) “first proposed by takeuchi and nonaka” (sic); (462) “So as John Frost and other leading KM thinkers,” as well as to Indian thinkers: (472) “I asked my guru…;” (476) ”Repeating GURU DR. W.S.'s dictum…” There are no references to a governance structure for KMforum but there is much discussion about leadership and the broader KM community. PP refers to leaders and membership: (489) “In a series of discussion between I and several KM-2forum leaders;” (ibid) “Paying member of the KM-2forum (…) and has been voted or appointed to be an official of the KM-2forum.” PP forwards an email from JF in which JF says: (489) “I'm a representative of KM-2forum. I'm its Executive Vice President of Education, Research and Membership…” The argument between JF and PP reportedly started because of power issues. PP thinks that JF does not play his role properly in KM-2forum and did not treat another member well. It seems that PP invited JF to join the forum he is moderating (EI forum) and then cancelled his membership.

Identity is still discussed in T5, although it is upstaged for a while by the argument segment. There are fewer references to the KMforum group than in the previous periods. There are some references to the KM community (most of them made by PP in his argument with JF and discussing the ideal KM community). There are many references to the Indian context and, as in T4, frequent references to face-to-face encounters, e.g., a conference in Delhi (470); a Mumbai chapter (532, 533, and 534).

The style of discourse is somewhat different than in the prior periods. It changes after the argument and starts to resume as previously only during the last messages of the period. People use fewer greetings, especially after the argument. There are still several personal forms of greeting which, as in T4, create a sense of intimacy. However, the style, which had become very informal during T4, is less informal in this period. There is less of an oral style. The forward pattern is still enacted. The referring pattern resumes as it was in T2 and T3. People still give positive feedback; however, this has declined since T4. Among the six occurrences in T5, five occur before the argument.

Language Games

Summary documentation of discourse related to roles, social identity, and linguistic style by time period is shown in Table 2. Through description of the discourse we are able to identify the emergence of language game patterns, including those that persist and those that change or disappear. Some patterns, such as those associated with social identity, were enacted during all five periods but through different communicative practices as the forum evolved.

Table 2. Documentation of discourse related to roles, social identity, and linguistic style by time period
ROLES Moderator - the administrator of the forum who regulates the technology or its use by contributors. Facilitator - a person who directs discussion content. GV: Founder, moderator and facilitator - The most frequent poster. Celebrates growth. Stirs up discussion (about 13 messages). - Other participants (48, 109, 120) refer to him as a key actor. - Evolution: around message 68, he starts mainly forwarding articles or references. Often with no greetings. - (85) He is “back.”“nice to see this egroup churning rich thoughts.” - He reaffirms his role in message 103; he apologizes for not having been active. Paradoxically, he also announces in this message that he won't be the moderator anymore “this egroup is unmoderated now.” GV: Founder, moderator and facilitator - Still the most active poster. Tends to post long messages that are just articles or references. In several messages, encourages discussion. - Others' perception: KMforum needs a moderator, and GV is the one. After a quite aggressive email sent by a member, HS writes that they needed a moderator (158); another participant suggests GV “our most prolific contributor” as the best candidate (159). GV replies saying that the group is now unmoderated and signs “a reluctant moderator” (167). GV: Founder, moderator and facilitator He continues these roles but is less active than in T1 and T2. No reference by others to his facilitator role. GV: Founder, moderator and facilitator - still active, but the third most active poster (GV posts 10 messages in T4). - GV posts some facilitating messages (4): “Now we are really talking! Any ideas to share about working with patterns” (430). - One reference to the moderator: (398) “we thank the moderator.” GV still active, but less of a moderator/ facilitator role. - Extensive forwarding and “passive” knowledge providing in the second part of the period (after the “end” of the argument).
Other facilitators: Is there only one facilitator, or several? Only one. HA takes on a facilitator role in (69). Only one. BH (143 and 147). Several. MH (3 messages) VS (1) PK (1) Several. VS (397); PP (402 and 408); Roshan (425); Derzko (438). Several. PP (462, 471); GN (472, 476); VS (507); SV (527).
  PP: Did not post. PP: Newcomer. Posts quite a bit (second most active poster). Seems to be willing to play an active role”“I am interested in developing a network of India thinkers” (202). PP: Posts frequently. A lot of authority discourse. Long postings. Quite theoretical. More a push model than an interactive one. No reference to the KMforum; some references to the Indian community. - A short argument with another participant (299, 301, 304, 306). PP: an important actor. He is the most active poster. Many posts refer to his work at the EI forum and to some work on Topic. Map (401) - he wants to start a new forum. PP is very active, especially in the first part. Initiated an argument with a member of another forum and uses KMforum as a stage to display the argument. He stops posting after message 501 (last reference to the argument). 
References to an organizational structure No reference.No reference.No reference.No reference to an organization structure for KMforum but references to other organizations: to the Global Knowledge Economics Council India and the ISO standards (402, 428, 449 - emails posted by ES).Reference to some organizations but not to the forum structure; to external organization as the GKEC and the organization of KM-2forum.
Gurus - valued experts; people who are referred to by others as knowledgeable References to Indian thinkers (7 messages refer to 4 different “gurus”).References to Indian thinkers (e.g., 4 to Dr RO), scholars (2) and to KMforum contributors (MA and DE; (237)).No references to gurus other than PP referring to people that he knows.Reference to gurus (6 messages): either KM thinkers (399, 402), or to Indian gurus (454, 455) “welcoming a familiar face in the Indian landscape of KM, Mr. B. M.”) and to DE (437, 454).References to important people in KM (460); to people in the KM-2forum (462) and to Indian thinkers (466, 472, 47, 517). In 472, NG writes, “I asked my guru, Dr. S. SS…” Indirect reference to Professor MS: GV posts two of his articles (504 and 516).
SOCIAL IDENTITY Building.Reflection.Stabilization.Identity is “assumed”- more informal; more intimate.Argument (487, 488, 489, 492, 500, 501) - frozen - starts again.
Praise- acknowledgment of number of members: Several times. “we are 58 KM enthusiasts currently in the group” (20).No praise of numbers.No praise of numbers.One reference to the number (like in T1): “This egroup we have here behaves in spurts and starts. Well, as a small update we are around 300 KM enthusiasts on the group as of today“ (455).No praise of numbers.
Collective language (“we” / “us”)Few “we”: 2Several “we” / “us:” 10Only one “we”Only one “we”No “we” / “us”
Reference to the KM-forum Use of the name, or reference to the group (“Hi, all!”“Dear friends,”“our group”).Many references to the group and KMforum (15+ messages).Many references to the KMforum group (17+ messages).Still several references to the group (12 - mostly by newcomers); Many newcomers introducing themselves.11 references to the group.Few references to the group (6 postings): - One before the argument; one by PP during the argument; four after the argument which corresponds to some sort of revival - three from necomers “Hello KMforum group… waiting for member support” and one by a member: “welcome to KMforum. I hope that we will have good and mindblowing discussion in this group” (527).
Reference to the KM community - - reference to the larger professional collective who share the same interest. Few references to the KM community: 4Some references to the KM community: 7Many references to the KM community: 12Some references to the KM community: 7, e.g., addresses to “Dear KM professionals” or “Dear Fellow KMers.”Few references to KM community: 5. Most of the references are made by PP in his argument when he discusses at length what the ideal KM community is.
Geographic identity: Reference to India - locale and shared culture. A few references to the Indian identity (in the early messages 2 and 5), then mostly implicit: names and signatures; “happy diwali” (116).More explicit references to the Indian identity (11+ messages).Quite a lot of references to the Indian identity: 12+ and calls on people (occurred in 4 messages): e.g., (274) subject: KM in India- replies: “but I guess the employees of these firms in India would be in the best position to answer as to how they are utilizing their KM systems;” (279) “Perhaps an HLL employee should be able to comment in detail.”Many references to the Indian identity: 16; e.g., messages with explicit references: (443) “but the reality in Indian situation;” (455) “MD and others in India (…) the last several posts by Indian colleagues here (KMforum) have reminded us that knowledge is something experienced by an individual…;”“Dear Kmers in India and beyond… welcoming a familiar face in the Indian landscape of KM…” First references to embodiment - e.g., (381) “All Mumbai / Pune based members” (AB asks them to contact him to create a Mumbai chapter - make the link with the discussion on the Mumbai chapter in T5: 532, 533, 535); (382) a conference in Bangalore.Many references to the Indian community: 20+ postings. (488) PP in the argument: “I include back the KMforum (KM India) because it is important to study what is going on here and to communicate back into the American community the feelings from India about this.” Most of them involve some embodiment e.g., (470) a conference in Delhi; (507) “I was at an e-HRD conference that was held by HRD network, India;” (532) “HI all, I was wondering if there is a Mumbai chapter of this community/ Any plans of an offline meeting among the Mumbai community?”
Discussion on the goal and rules of the forum. No discussion on the forum goals and rules.Some reflection on the rules and the aim of the community (9 messages) ± 4 messages referring to the activity of the forum. “It was good to see active discussion after a little lul” (162); “I found this group less active, so friends let us now have a round of vigorous discussion on few topics that you may all suggest” (243).No discussion on the forum goals and rules.No discussion on the forum goals and rules.References to rules and netiquette (as in T2). NB: all these references are made by PP. In (471), he says that it is “harmful” to repeat text when you reply (cut and paste patterns). This is not a pattern in the forum (there might be few cases, but far from being a pattern). He is referring to this rule in one of his emails during the argument.
LINGUISTIC STYLE GreetingsGreetings: most of the time.Greetings nearly always.Greetings nearly always.Greetings nearly always.Greetings before the argument… a pause after and then starts again.
Degree of formality or informality in speech (e.g., punctuation, emoticons, etc.) Some use of punctuations and informal style (but not very frequent about 5 mess). (26) “Phew! I guess I'll leave it like that.”“Etc. etc.!!… quite a lengthy topic eh??…”Some use of punctuation and informal style. (150) “I am not going to define it anyway myself, but am thinking aloud” - challenging some of the discussion. (162) “I can't resist quoting GR here…”Some use of punctuation and informal style. (295), “Hi! rambling account, but the truth lies somewhere in between U can skip the intro, but I suggest you read it, just to appreciate why I think I know what I am talking about” (295)“Pompousness apart;-) I write below my humble opinion about the Truth in KM. Your comments are invited.” (275) “Hi all!!!!” Smiles: (295), three in (353)More use of punctuation and emoticons and of informal style. e.g., “We should not forget it bcos IT. I otherwise could not have even interacted with you people!!!” (388) “… SOS! Any suggestions would be handy” (424) (441)“So bottom line:–think not to capture tacit knowledge but share it (…) Throw in a ritual dance or two as well…;-) evolve man, why remain textual only when the technology ain't.” 4 smiles.Less use of punctuation and emoticons. Some informal style but mostly before and during the argument and then one in (527). Smiles and exclamation marks in 2 messages before the argument (462, 476); (489) argument: use of capitals to emphasize. After argument, no use until message (523): emoticon in the title.
Parenthesis: Contributors make heavy use of parentheses to embellish their text; this creates an oral style. 7 occurrences9 occurrences13 occurrences5 occurrences Many 50 in 10 postings (most of them in the argument period).
Forward pattern 7 occurrences8 occurrences13 occurrences8 occurrences11 occurrences
Referring to: Contributors also tend to make reference to other messages (“this is in response to S's message…”). 1210414
Feedback - expressing agreement or disagreement, or seeking others' agreement or disagreement 354106 (5 before the argument; 1 after)
Personal addresses -to one specific individual, e.g., “Hi Meg” or “Dear Vibha” FewFewFew7+: creates a sense of intimacy.Several personal addresses. 12 (Including 3 from PP to JF): Creates a sense of intimacy. Intimacy / references to off-line communications: 11 (Including 5 from PP referring to his private discussions to JF and other KM-2forum leaders).


Few roles enter the discourse of the forum and none refer to an official organizational structure, but the roles that are enacted are very important to structuring the discussion and in allowing the larger language game of KMforum to develop and survive. Two roles are noteworthy: facilitating/moderating and being a guru. Although various participants play the role of facilitator at different times, GV is the primary player of this role and is sometimes acknowledged as such by the participants. The language game of guru involves formal recognition as such; gurus are identified, usually with titles (e.g., “Dr.,”“guru”) and treated with respect by others in the forum. Whereas many newcomers are formally introduced (by themselves or others) and explicitly welcomed to the forum (usually by the founder), gurus are given special introductions and more explicit respect (e.g., praise) by participants. This may be related to the high value placed on titles, and formalities in use of titles, in Indian culture (India Review, 2003).

The role of facilitating is blended with moderating and is largely taken on by the founder, GV. He is the dominant player of the facilitating/moderating role throughout T1-T5, although he recedes in T4 as PP becomes the most active poster. In T5, GV resumes as most active contributor, but he diminishes in his facilitating role; he mostly forwards information and does not stir up discussions. In T1, GV is explicitly moderating and facilitating the forum with a clear aim to build the community. He celebrates growth and stirs up discussion, and other participants refer to him as a key actor. In T2, he is still very active, but at one point he states that the forum is un-moderated (167). However, he still seems to play a facilitating/moderating role. As GV recedes as facilitator in T3 and T4, other members of the core contributing group become more active in stirring up discussions.

The role of PP is interesting in that he tries to create a role for himself in the forum, but others are not willing to play the game with him and, as a result, no new language game develops despite his domination of the discussion. He enters the forum in T2 and is the second most active contributor in T2 and T3. He becomes the most active contributor in T4. Most of his posts are affirmation of his convictions without request for feedback or discussions. PP brings argument into the language game of the forum; but since others do not reply to any of his postings regarding the argument, an argumentative discourse fails to take off. He posts a last message on the argument in the middle of T5 and then stops posting.

There is a small but active core group (see Table 1) that shares the facilitation role with GV, especially as his level of participation declines in T4 and T5. Apart from the core group, there are three members who are highly active in T1 (i.e., post more than 5 messages) who then stop making any contribution after T1 or T2. This small group is interesting in that they are early members who are involved in creating a lively atmosphere and in “setting the stage.” Once this is done, these early members disappear and others enter and join the core group of players. The majority of forum contributors do not participate consistently. They post one or two messages once in a while. It seems that the core group has enacted a discourse pattern that is sufficiently defined and yet simple enough to allow the participants to enter and quickly know how to behave and play the language game of the forum.

Social Identity

Language games related to social identity are enacted during all five periods, although with some variations and through different communicative practices (see Table 2). Some practices, such as explicit reference to the social groups of KMforum, the KM professional community, and the Indian community, are enacted through all five periods. Other practices, such as the use of collective language, the reference to rules and netiquette, and the praising of the number of members, are enacted only during some periods.

References to the KMforum as group (“Dear KM-ers,”“interesting for this e-group,” etc.) are present during the five periods, but these decrease over time, especially in T5. References to the larger KM professional community, while fewer overall, are consistent in frequency over the whole period. References to the national/cultural identity of India are present during all five periods and increase over time; further, the nature of these references changes. In T4 and T5, many of the references to India are concrete references to geography and potential face-to-face meetings, such as conferences and the creation of regional or city-specific chapters (e.g., (470) reference to a conference in Delhi; (502) announcement of a conference “sharing of KM implementation in the Indian corporate context”; and discussion of the creation of a Mumbai chapter (381, 532, 533 and 534)). This increasing prominence of locale is interesting in that face-to-face contact among the participants follows their formation of a group (forum) identity rather than preceding it. We might expect common locale to aid social identity in an otherwise distributed group such that KM professionals from the same geographic region would more readily identify with one another than with people distributed around the world. Commonality of culture and experience are more likely to promote social solidarity (Markovsky & Chaffe, 1995), identity synergy (Pratt & Foreman, 2000), and thus commonality in a group's language game. References to face-to-face meetings, which are nonexistent at the start, enter the language game related to the Indian locale only after the forum develops. We refer to this phenomenon as embodiment of the language game-the link between the forum and sharing of physical space.

Linguistic Style

Language games of discursive expression are the most complex and in KMforum appear to include practices that enable management of relationships and giving and receiving feedback. Members of KMforum use an informal and friendly style of interacting, most of the time using greetings (either general “Hello,”“Hi friends” or personal addresses), closings (“cheers,”“best regards,”“ciao”), and expressions of politeness and gratitude (“Please forward this announcement,”“I would be grateful…,”“Thanks”). Beyond these basic forms of relationship management, the forum contributors develop several idiosyncratic communicative practices. Some of these are maintained during the five time periods, whereas others are patterned early on but then are replaced with new practices in later periods.

Two patterns develop early on and persist throughout the forum's life. First, participants make heavy use of parentheses to express thoughts and feelings. By doing this, the discourse takes on an informal, oral style. Second, participants tend to forward others' messages-either messages that were posted in another forum that the participant wants to share in KMforum, or personal email replies received to a message that were sent only to the poster.

A stylistic pattern that emerges and then fades is linking of messages to one another. In T1 and T2 participants frequently make explicit references to others' postings (“As × said in his previous email…”), but this diminishes substantially in the later periods. (There are only four occurrence of this referring pattern in T3, one in T4, and four in T5.) On the other hand, the pattern of feedback—expressing agreement or disagreement, and seeking others' reactions to comments—is rarely used in T1-T3 and increases markedly in T4, which corresponds to the period where the forum is well-established. Similarly, personal addressing—“speaking” one-on-one to others in the forum—grows markedly in later periods. One could explain the emergence of these practices as a more intimate discourse that emerges after the forum has established a strong enough feeling of social identity. Perhaps not surprisingly, in T5 the feedback pattern nearly disappears after the argument (it is used only once).

The language game of linguistic style interplays with those associated with roles and social identity. These conversational patterns are introduced and reinforced by the core group members, especially (but not only) the founder. The language game of style appears to create a context that maintains the “link” across the five periods and gives some sense of stability and cohesion to the forum. The well-defined and friendly patterns of discourse allow newcomers to enter and participate, even as the forum's game is ongoing with others. The forum's social identity is reinforced even as newcomers enter and older players smoothly exit.

Developmental Phases

Table 3 highlights the overall evolution of KMforum's language game over time. The rules of the game—who plays (roles), how they speak (linguistic style), and the manner in which they identify themselves as playing together (social identity)—are established early on and, for the most part, persist long into the future. But we also observed subtle shifts suggestive of developmental phases in the evolution of the forum's language dynamics. The five time periods we analyzed correspond to developmental phases described by Tuckman and Jensen (1977). T1 is the building or forming period (the community is attracting members, praising growth and starting to form routines of conversation among participants). T2 is a reflective or norming period. During this time there are many discussions about appropriate behaviors, the aim of the forum, and its activity. Roles and expression of identity become routine, and the stylistic attributes of the discourse, established in T1, are repeated. T3 is a stabilization period in which the patterns of the past are performed in a steady stride of conversation. The foundations for forum discourse appear firm, as patterns enacted in T1 and T2 are maintained and further reinforced. T4 is also a stable period, a continuing period of performance in which the forum seems well-established and interactions are lively. Members are more informal and more intimate. There are more and more references to the Indian locale as well as to face-to-face encounters.

The last period, T5, is a critical period because in the middle of it an argument is brought into the forum by an active poster, PP. A period of routine performance of the language game is disrupted by a “storm.” PP presents his argument to the KMforum and, at times, tries to draw participants into the fray. The forum members do not react, however, and instead ignore the fight. Once PP stops posting, a change in style can be noticed. The forum enters a brief frozen period in which the language game appears to stop, or at least pause in its rhythm. There are few discussions during this period; instead, passive knowledge is provided (mostly by the founder, GV). Greetings and other forms of relationship management and feedback tend to disappear. However, after this frozen period, the last postings of T5 are informal and friendly again—referring to the community and to potential exchange and discussion. The group re-norms as it recovers and resumes its language game.

Tuckman's stage theory argues that group phases reflect members' concerns with resolving interpersonal relationships and task activities as the group moves through its life. Although he postulated a defined sequence, with all groups proceeding sequentially through phases of forming, storming, norming, performing, and (in some cases) adjourning (i.e., dissolving), other group research suggests that phases of group development vary across groups and contexts and do not necessarily occur in defined sequences (e.g., McGrath, 1991; Poole, 1983). Instead, groups can cycle through developmental phases, repeat them, or in some cases skip phases. Our analysis suggests that the development of KMforum's language game can be meaningfully described in terms of the stages Tuckman hypothesized. The phasic sequence differs from Tuckman's postulation; moreover, one can anticipate that the exact sequencing of the phases will vary across forums, depending on the players of the language games, the precise rules that they enact, and the particular conversational experiences that take place as the forum evolves. Nonetheless, the phases of forming, norming, performing, storming, along with the cycle of repeated norming and, later, re-norming, are noteworthy. They provide useful insight into how this conversation of KM professionals established itself and survived as an online forum.

Of particular note in the life of KMforum is its storming period, the disruption this segment caused in the forum's ability to “perform” its language game, and its eventual recovery. The argument involved one of the active members of the forum, PP, but the target of his argument (JF) was not a member of KMforum and never replied to PP's postings in the forum. Beyond the personal and power issues, issues of rules and appropriate behaviors in the KM community are discussed. Hence, PP writes, (487) “the rules of the norm can become larger… Is this the type of society that the leadership of KMforum wants? (…) the imposition of this dogma will lead to violence,” and he claims that (490) “people leave this forum because of this behavior by one or more of the dominant participants…” PP presents his argument with JF as a way to advance a perspective on what the KM community should be. Hence he writes: (489) “This copied message below was the fourth or fifth in a series of discussions between John and I and several KM-2forum leaders. The nature of the discussion can be reproduced here and we as a community can look at the issues that are raised.” Although this “storming” segment disrupts the language game, it fails to yield new rules of play or to cause mass exiting of contributors such that KMforum dissolves. The forum risks reaching an adjourning phase (Tuckman's label for the end of a group's life), but instead it resumes its life. Why? We see two reasons for this.

First, PP is the only “speaker” during the argument. He summarizes what was said by JF and others, or he includes quotes or cuts and pastes their messages. But no one replies to PP's postings; hence, no one joins him in the language game that he is initiating. Though PP disrupts the game, he fails to transform it or substantively change it, because others do not play with him. Instead, they are silent.

Second, the rules of the KMforum game have been performed for a long period (the six months of T3 and T4). PP is not playing by these rules and, further, is advocating new norms and ways of behaving in KM forums. The well-established routines of discourse constitute a sort of organizational capability that is not readily replaced. Major change is difficult, perhaps impossible, at this point in the forum's life. In other words, long-term performing seems to prevent any movement to adjourn as a forum.

It is only towards the end of T5 that the forum seems to come back to life and the interactions take on the pattern performed in T3, T4, and before the argument. Hence, KMforum survives the argument and, following a short period of low activity and more formal interactions, goes back to the form of life that had been enacted in the previous periods. The forum experienced steady growth in November 2001 (at about the same pace as in T4 and T5) and remains active today.


This analysis has shown how the language games of roles, social identity, and discourse style can collectively provide insight into the evolutionary dynamics of a professional development forum. We learn from KMforum that a loose collection of professionals who share interest in a common topic but otherwise do not know one another or share task goals or incentives can develop a rhythm of conversation that allows them to develop sustainable interaction. The group is porous, not insular; people come and go, presenting ideas and sharing insights within a congenial and lively setting of knowledge exchange. They are held together by an interaction context that contains very few roles, an emergent social identity, and a limited set of rules for discourse exchange, nearly all of which are implicit rather than explicitly stated. The loose nature of the forum makes its robustness and long-term survival impressive, indeed.

We observed several important features of KMforum's dynamics that appear to underlie its successful development. Some of these features have been noted in prior research whereas others are perhaps new. First, the moderator/facilitator orchestrated the activity of the forum. He set the pace and rules of the game early on. As the rules became established and well performed on the part of the group, the moderator/facilitator became less active; nonetheless, he remained present throughout the 15 months of the forum's life. Other contributors to the forum occasionally took on the facilitative role and could do so in a smooth, effective fashion, mimicking the moderator/facilitator's behavior. Thus, sharing of the moderator/facilitator role was done by other players in the game, although it was dominantly played by the same individual. Second, a small core group of six contributors acted to launch the language game of the forum and keep it in play over time. The people in this core group changed over time, but the presence of this small, active group of contributors worked to keep the language game of the overall KMforum in steady play for the first 15 months of its life. This steady play of activity established the foundation of the group's life and set the stage for its future.

The third important aspect of KMforum's development was discourse related to the group's social identity. The language of social identity was not limited to the forum itself (organizational identity) but instead incorporated the larger social groupings of KM professionals (professional identity) and the Indian locale (geographic identity). The language of organizational identity was loose, emphasizing growth of the group and openness rather than tightness; intimate language developed slowly and with an openness to include new members. In this way, the group did not become tight or insular but instead formed social identity rooted in commonalities of profession and locale. Their language game emphasized a confluence of commonalities in social background that created a feeling of “we” but with open opportunity for people to come and go within the group. The group became embodied, in the sense of meeting face-to-face as an outgrowth of their interaction, after their language game was well formed.

The final important aspect of KMforum's development that we observed in our study was the evolution of its discourse style. Friendly greeting of new members and expression of politeness and gratitude were evident in the language game from the start and persisted long into its life. Parenthetical and other informalities of expression created an oral style of conversation that, along with linking of messages one to the other led to a meaningful, lively conversation despite turnover in the players as the language game progressed. Over time, the game's discourse became more intimate as players addressed one another by name and issued invitations and follow-on notes about face-to-face events. Rapid formation of the forum's discourse style, followed by norming and repeated performance of the style over time, operated as an organizational structure of sorts. This structure, along with forces of organizational identity and roles, provided sufficient strength for the forum to survive in the face of moves by PP to change the nature of the game. The game paused (became silent) when disrupted, but it eventually restarted and resumed play.

With these findings in mind, we offer the following propositions for future research on online professional development forums:

P1. Language games form as a sequence of observable stages in online forums that can be described in terms of discourse related to roles, social identity, and discourse style.

P2. Developmental stages of language games may not occur in the same sequence for all forums, but they include: forming, storming, norming, and performing.

P3. The basic role of moderator/facilitator can be sufficient to orchestrate the activity of the forum; further role differentiation is not required, so long as the moderator/facilitator remains active in setting the pace and rules of the language game and contributors are willing to take on the moderator/facilitative role when the game experiences pauses or the moderator/facilitator becomes less active for some reason.

P4. A small core group of active participants is critical to sustaining a forum's language game. This group can have high turnover; its most important component is frequent contribution in which the rules of the language game (in terms of role, social identity, and linguistic style) are performed.

P5. Strong organizational identity is not required for forum survival. Social identity that incorporates loose organizational, professional, and geographic-based referents can create sufficient cohesion for forum survival.

P6. Rapid formation of a forum's discourse style, followed by norming and repeated performance of the style over time, facilitates long-term survival.

P7. Forums with well established language games can sustain major disruptions (e.g., arguments) and avoid adjourning.

Implications for the Design and Management of Professional Discussion Forums

Our study was limited to analysis of one professional forum, and we examined the discourse for only the first 15 months of its (now) four-year old history. With these limitations in mind, we offer the following tentative guidelines for design and management of professional development forums, especially those of similar size and scope as the one we studied. Some of these guidelines reinforce those suggested in prior studies (e.g., Mynatt et al., 1998; Preece, 2000; Wenger, McDermott & Snyder, 2002). Others provide new insight or enhance existing guidelines.

Build a Conversational Rhythm Early On

As noted by Wenger et al. (2002) and Mynatt et al. (1998), building a conversational rhythm is key to the development of online forums. The first periods of a forum's development (the forming and norming phases) are critical to establishing patterns of group conversation. Efforts have to be made right at the beginning not only to get the discussion going but also to set the context and behaviors that will become the basis for the forum's routine of discourse long into its life. Two roles can help in this process: (1) the facilitator/moderator who welcomes newcomers, stirs up the discussion, sets the pattern for greetings and closings, and in general is instrumental in starting up the forum and building the stage for lively interactions; (2) a small core group of very active members who engage the play of a lively pattern of friendly discourse in the forum. Jarvenpaa and Leidner (1998) noted the importance of forming trust early on in small distributed work teams. To set the stage in an online forum we suggest something similar, a lively rhythm of discourse with a lot of interactions, attention to relationship management, a good rate of response and frequent positive feedback.

Create A Context To Facilitate And Support Interactions

The context—the communicative practices and tone enacted in the forum—is crucial (see Powazek, 2001; Preece, 2000). The forum's context provides a linguistic model of behavior for newcomers, a sort of behavioral script for participating in the language game. An established context assures that the success of the forum does not rely solely on few key individuals. The forum's context provides a sense of continuity and coherence, as well as a sense of social identity, in a setting where the majority of the contributors do not participate consistently. For example, in KMforum the references to social identity created a sense of shared social practice that was open to participation by newcomers. Whoever was joining the forum could have a sense of social belonging by reading the messages, often starting with “dear group,”“dear KM Forum members,” etc. The enactment of some communication routines such as this in the first periods of a forum's life provides a sense of continuity and coherence, and gives a model to the newcomers who then replicate the same routines when they start posting. Taking the time to articulate some of these routines, as it was done in KMforum during T2, can also help to build the context and reinforce the sense of common identity as it provides clearer definitions of what it means to interact in the forum. What are the aims and what are we expected to do as we participate?

Develop Strong Sociability Criteria

As part of creating a context for developing interactions and sharing knowledge, the discourse style and the level of sociability it conveys are crucial. Members of KMforum developed a language game for relationship management, most of the time using greetings (either general “Hello,”“Hi friends,” or personal addresses), closings (“cheers,”“best regards,”“ciao”), and expressions of politeness and gratitude. Relationship management, acknowledging others' contributions, and providing positive feedback need to be emphasized and nurtured (Preece, 2000). It is important to develop a friendly, informal style and a positive tone right from the beginning. This is a key part of creating the context. It is also important to be vigilant and avoid letting nasty emails escalate, while at the same time making it clear that people can express themselves freely and have differing opinions. The facilitator/moderator can set up the rules and make sure that they are followed, as suggested by Preece (2000). However, more than a formal netiquette, it is the way the messages are written—how informal, friendly and positive they are—that will create a sense of sociability and belonging, allowing newcomers to have a sense of the forum's style as soon as they start reading the postings.

Know How To Avoid Conflict Escalation

Our analysis provides an alternative to the general advice to groups that they acknowledge conflict or, in the case of online communities, “take the conflict outside” (Powazek, 2001, p. 113) or resolve conflict by using email or other means external to the public forum (Elliott & Scacchi, 2003). In the case described in this article, ignoring conflict was a smart and wise choice by forum participants as it allowed the participants to maintain their language game. Any replies to PP might have created an uncontrollable escalation of conflict. Ignoring conflict helped the forum to survive. In other situations, it might be more efficient to acknowledge conflict and explicitly manage it. In fact, in T2, one participant replied to a newcomer's nasty email by expressing his disappointment. This message was not ignored; instead it led to a very lively discussion regarding the netiquette and aims of the forum, and a subsequent agreement on the importance of being open-minded and polite, as well as on the key role of the moderator. This quick reaction from the moderator followed by lively discussion once again avoided flaming as well as frustration. The lesson is that conflicts have to be managed, and the moderator can play a role here in addressing or ignoring the issues depending on the context (topic, history, individuals involved, etc.).

Support Embodiment

Our case study shows the importance of locale as a source of social identity discourse over time. Hence, the references to physical places in India, which were nearly nonexistent at the start, increased over time. In the last periods, postings increasingly referred to conferences and some members suggested the creation of local chapters. The forum, therefore, became a facilitator for face-to-face interactions, and technology was bootstrapping the socialization process. This suggests that there may be benefit to creating some “physical” structure to “ground” the online community and support interactions. One could also add some functionalities such as pictures or videos in order to embody the different actors and their postings (per Mynatt et al., 1998). Alternatively, one might start with a “physical” forum of professionals and use the technology to extend and augment it; but our study suggests that the reverse is also possible: the online forum can spur face-to-face forums for people to converse.

Wenger et al. (2002) have highlighted the importance of embodiment for distributed communities of practice (e.g., arranging teleconferences, face-to-face meetings, etc.). However, they described cases of local communities that are extended and distributed within global organizations, whereas the online forum we studied is a forum where participants belong to different organizations, and where the online interactions spurred embodiment-local organizations and face-to-face meetings. Our results suggest that Wenger et al.'s advice may apply to online professional forums that form outside of corporate boundaries.


The language game framework provides a useful and relevant lens for researchers interested in taking a developmental approach to the study of online discussion forums. By examining a forum's discourse as the interweaving of language games, such as those related to roles, social identity, and style, the researcher can detect the continuity of general patterns at a macro level (the whole of a forum's development) without ignoring the subtle variations at the micro-level (development of specific types of discourse over time). The language game approach needs more elaboration through further empirical studies of online discussion forums. One important future direction is to study language games as a function of professional domain. For example, studies could compare online forums discussing medicine, supply chain management, or graphic design. Other studies might take a comparative approach to investigate linguistic development at the level of topical domains, such as comparing the evolution of forums devoted to the KM theme or comparing forums on particular medical topics, to surmise whether there is a specific language game enacted for each of these topics.

Regardless of the forum being examined, researchers using the language game lens should take care to attend not only to the stylistic aspects of online discourse but also to the discourse surrounding the roles taken up by the participants and the social activities performed (e.g., identity formation, relationship management, feedback). Examination of language games beyond those considered in our study would be particularly worthwhile. For example, the activities of information sharing, sense making, negotiation, argument, and agreement within professional development forums could be examined within the language game framework. Such an approach would help to highlight the dynamics of learning within professional development forums.

In sum, a language game approach offers researchers who are interested in studying online forums a powerful framework to investigate such forums' evolution and dynamics. We hope the initial study reported here leads to further inquiry into the language games of online forums and their developmental processes.


This research is supported by the National Science Foundation under Grant No. SES-0135602. The assistance of Mike Roach and Siyuan Huang is gratefully acknowledged.


  • 1

    The forum name, participant names, and other personal identifiers have been substituted with fictitious names.

  • 2

    Numbers in parentheses preceding quotes refer to message number in the forum.