Birds of a Feather Protest Together

Syst Pract Action Res DOI 10.1007/s11213-010-9167-3 ORIGINAL PAPER

Birds of a Feather Protest Together: Theorizing Self-Organizing Political Protests with Flock Theory Devan Rosen • Jang Hyun Kim • Yoonjae Nam

! Springer Science+Business Media, LLC 2010

Abstract The current research theorizes decentralized and self-organized political protests via flock theory (Rosen 2002), a theory of emergent self-organization in communicative human interaction. Flock theory draws from a theoretical basis of emergence and self-organizing systems, and is presented as a theory of decentralized communication structures. Focusing on the optimization of decentralized networks and roadmap based coordination of organizationally homopholous foci; the current theory poses a model of human interaction that captures the potentially egalitarian effects of cooperative evolution and collective action. Case studies of two separate decentralized political protests against Korean government and FARC are offered as evidences of such phenomena.

Keywords Self-organizing systems ! Social networks ! Political protests !

Decentralized networks ! Flock theory Introduction Effective political protests may be more like jazz improvisation than a classical symphony.

It may not seem like this would impact the structure of recent political protests, but this analogy has a great deal of application. At the basic level, the comparison between the two forms of musical performance revolves around the structure, or pattern, of the interaction.

Improvisational jazz allows the individuals flexibility of expression (Berliner 1997), and a symphony has highly centralized control providing little latitude of interpretation to the individual performers. The jazz group still has a base level of structure, such as what key and tempo the performance will be in, but there is substantial room for interpretation and flexibility of group direction.

D. Rosen (&) ! J. H. Kim University of Hawaii at Manoa, Honolulu, HI, USA e-mail: rosend@hawaii.edu Y. Nam University at Buffalo, Buffalo, NY, USA

123

Syst Pract Action Res

The structure of interaction among individuals coordinating to protest can have as much variation as musical performance; some are decentralized and open, providing participants more freedom to make decisions and pay close attention to each other. On the other hand, some organized protests are highly centralized, with most decisions coming from higher levels of the hierarchy; individuals are not able to participate in decision-making or communicate as freely with others. There is a great deal of impact that the structural composition of the organized group can have on an organization as a whole (Baron and Hannan 2002; Cummings and Cross 2003). Crossley (2003) points out that administrative systems and bureaucratic structures prove largely unreceptive to public opinion and pressure, amplifying the need for alternative self-organizing routes to new social movements.

Although social network analysis has been widely used in the study of organizational structures, the majority of the extant research is limited to implications of methodological results regarding related variables (e.g. performance). What is lacking from the extant literature is a theoretical base indicating the means by which the groups and organizations studied, such as political protesters, can foster and maintain a decentralized structure, and how communication networks are central to the maintenance of these structures. Further, there is little research that extends existing decentralization literature to political protests.

In this vein, the current research discusses the advantages of decentralized structures for collaboration and knowledge sharing, and applies this model to political protesters.

Political rally examples suggested in this study are the anti-FARC rallied around the world and anti-US beef import movement in South Korea. People with similar information, attitudes, and beliefs on the social agenda organized these political protests (Fuchs 2006).

No specific group led the whole process of self-organizing. Rather, the organizing process was improvised like jazz and was managed with systemic adjustment of cohesion, information exchange, and peaceful means. Further, the organizing process in both cases was mediated by the social media website Facebook. Overall, the two rallies were non-violent and self-regulated by participants. Leadership in the rallies was flexible-relying on emergent leadership—according to information availability and changing situations. Further, as stated above, there are rare extant studies to interpret these voluntarism-based protests organized online with systemic approaches like the flock model.

This paper is structured as follows. In the first section, a review of the social network (structural) perspective and research on the benefits of organizational decentralization are provided. The next section covers flock theory (for a more rudimentary description of the model see Rosen 2008), providing contributing literature and underlying concepts, the proposed theoretical flock model, and homophily of foci as a contagion norm for protest groups. Following the discussion of the flock theory model, a discussion of contributing literature on roadmap-based organization is used to frame the bipartite small worlds part of the flock model and is applied to the core flock theory model. Case studies on decentralized political protests are elaborated, and portions of the theory are illustrated. Then, limitation of self-organization and contexts that the flock model would not apply are explained.

Finally, conclusions and suggestions for future research are offered.

Literature Review Social Networks and Decentralization The social network perspective focuses on the structure of social systems and how the elements of a social system come together. From the network perspective, the social

123

Syst Pract Action Res

environment can be expressed as patterns or regularities in relationships among interacting units, often called structure.

The form of network that will be utilized herein is a communication network, defined as the patterns of contact that are created by the flow of messages among communicators through time and space (Monge and Contractor 2003). Communication network analysis identifies the communication structure shaped by the flows of information or other material/nonmaterial resources. The ties and exchanges among the nodes can vary along several elements, including direction, reciprocity, and strength.

Decentralization and Performance One of the essential explanatory tenets of network theory is that structure in social networks has the ability to either constrain or augment contact with important resources (Shaw 1964; Ibarra 1993). Sparrowe et al. (2001) and Cummings and Cross (2003) extend this research by investigating group network structure and performance, and how informal relationships potentially hinder individual and group performance. Research on the social structure of groups and organizations indicates that decentralized structures are superior regarding group performance and coordination.

For instance, Cummings and Cross (2003) researched effects of the structural properties of work groups and their consequences in terms of performance. Analyzing 122 work groups working on complex and non-routine projects, they found hierarchical and coreperiphery structures to be inversely associated with group communication. Hierarchical structures were negatively related to group both manager- and member-rated group performance. Greater core-periphery structure, where there is a cohesive core and a sparse and unconnected periphery, was negatively related to the manager-rated performance of the group. Also, greater the ‘‘structural hole’’ role of the group member or leader, which forces individuals in the network to navigate through the person that bridge relationships, were negatively related to both manager- and leader-rated group performance.

Work-related resources (e.g. task advice and strategic information) are exchanged through informal organizational networks, yet these networks also transmit social support and identity (Sparrowe et al. 2001). Specifically, Sparrowe et al. focus on the exchange of task advice and information via advice networks (the means by which individuals can exchange informational resources essential to task completion), and on obstructive relations via hindrance networks (relationships between coworkers who thwart task behaviors). Centrality in the advice network was positively related to individual performance, while centrality in the hindrance network was negatively related to it. Hindrance network density was inversely associated with group performance, which points to the conclusion that uncooperative behaviors are just as closely related to performance as cooperative behaviors, further compounding the importance of satisfied group members.

Although these studies show decentralized structures’ efficiency, there is a lack of theory suggesting both the means by which groups can maintain decentralized structure while still interacting in a cohesive manner and functions that are crucial for self-organized political protests. The following section reviews flock theory (Rosen 2002, 2008) as a means by which a group or network can have a decentralized structure as well as a cohesive interaction.

123

Syst Pract Action Res

Flock Theory Literature which helps to establish the flock model should be examined here, focusing on emergence, autopoiesis and self-organizing systems, and boids. Then, the flock model is developed and presented using these concepts.

Underlying Concepts Emergence With the increase of work in fields such as nonlinear dynamics and complexity theories (see Nicolis 1989; Prigogine and Stengers 1984), natural systems and processes that cannot be explained by an overly reductionistic perspective due to the mathematical complexity of such phenomena (Goldstein 1999) have been elaborated. For this kind of research, the estimation of initial conditions will not suffice for accuracy, undermining the prospect for simplified prediction and reductionist explanation. A political protest, for example, may start in a cooperative manner, where novel ideas emerge through the interaction. But this does not imply that these initial conditions will determine the course of the entire interaction. Emergent properties are, by definition, not explainable in conditions of basic elements, and to explain systems of complexity it is essential to rely on more macro levels. Thus, flock theory balances partial reduction concerning individual actions, and macro elements such as group norms and political organization.

A decentralized political protest forms and matures without the use of a bureaucratic structure. Individuals become interdependent in that the ideas that they create arise out of interaction, and thus would not necessarily arise without interaction. It is in this sense that emergence allows for the creation of new ideas, as Goldstein (1999, p. 49) elaborates, ‘‘Emergence … refers to the arising of novel and coherent structures, patterns, and properties during the process of self-organization in complex systems.’’ Emergence research has a rich and multidisciplinary history of investigation into the characteristics associated with emergent phenomena, often falling under the titles of emergence (see Darley 1994; Gilbert 1995; Hodgson 2000; Wheeler 1928), complexity theory (see Gleick 1988; Prigogine and Stengers 1984), or self-organizing systems (see Contractor 1994; Contractor and Grant 1996; Contractor and Seibold 1993; Maturana and Varela 1980; Monge and Contractor 2001; Monge and Eisenberg 1987).

Emergence describes the patterns (e.g. building off of one another’s ideas), structures (e.g. no formal leader), and properties (e.g. constantly changing informal leaders) that some systems embody on the macro level. Emergence focuses on systemic self-organization, not particular outcomes. A common criticism of emergence has been that the concept is an ‘‘epistemic recognition of the inadequacies of any current theory for deriving macro-level properties from micro-level determinants’’ (Goldstein 1999, p. 59). It is argued here that the provisional nature of emergence can actually be a supportive element because science should be able to deal with phenomena for which there is less than perfect knowledge.

Sperry (1986) pointed out that the mind emerges out of brain functions, yet the mind can have contributory power in affecting the brain—if emergents have causal power, then they cannot be simply recognizing the inadequacies of other theories. For example, in a decentralized protest group the very structure (as in decreased leadership) has causal power by fostering increased involvement from group members and fluid response to treats.

Emergence perspectives may be crucial for social sciences as they provide a means to explain higher-level relations, avoiding the problem of analytic reduction to lower-level units. Where traditional physics has had the ability to study complete order or utter randomness, emergence offers the ability to understand the middle ground where there are

123

Syst Pract Action Res

elements of randomness paired with elements of order (Goldstein 1999). As a result, the absence of adequate frameworks for emergent order acts as a hurdle to emergence being accepted as ontologically viable. The current research hopes to fill some of this gap.

It is important to establish the nature of the self-organization that is most applicable to emergent phenomena when a system has a set organization that is closed to environmental forces (political pressure or regimes), yet remains structurally open to these forces (protesting them), such as in the case of autopoiesis.

Autopoiesis and Self-Organizing Systems The process that individuals undergo to attempt to increase the level of coordination with each other can be seen as a function of autopoiesis, or the recursive self-reproduction of components in a system. An example of an autopoietic system is a flock of birds; the flock doesn’t have to change its means of organizing in order to respond to environmental obstacles, such as buildings and trees. The flock may have to change its structure, but not the way it organizes itself.

One of the main functions of an autopoietic system is to maintain its autonomy, and thus can be further defined as ‘‘a network of processes that produce all the components necessary to embody the very process that produces it’’ (Krippendorf 1991, p. 138). In this sense, autopoietic systems recursively produce all the components necessary to have a historically reproductive network, and are likewise self-reproducing. Maturana and Varela (1987) argue that within the reproduction it is important for organization, or the system (i.e., the ‘‘flock’’), to maintain its identity while its structure can change to adapt to the environment. Thus, autopoietic systems have the ability to maintain organization in relation to its structure while remaining operationally closed. The system is structurally coupled with the environment and organizationally closed to it at the same time, conditions that catalyze political protests.

Here, organizational closure refers to the ability of the system to use its own identity as the crucial trajectory point of reference when attempting to undergo change. An autopoietic organization is a process of self-reproduction based on internal rules, thus closed at the level of organization, yet still open to structural environmental change (Maturana and Varela 1987). In addition, structural coupling is the process of how self-organized systems influence each other (Luhmann 1995), which allows for operational autonomy of autopoietic systems with the environment. Likewise, the structural coupling of the system and the environment, or other systems, does not necessarily direct the internal rules of the system.

Instead, the environment only causes structural changes within the system, revealing recurrent interdependencies between the environment and system (Maturana and Varela 1987). Thus a system lacks the ability to undergo structural change without structural coupling, explaining the foundation of the evolution of a system to a self-organizing system.

In the case of structural coupling of a system with another system (e.g. distributed protest groups as a part of a larger coordinated protest), the result of ontogenic structural coupling is a consensual domain, defined as ‘‘a domain of interlocked (intercalated and mutually triggering) sequences of states, established and determined through ontogenic interactions’’ (Maturana 1975, p. 326). Consensual domains shape the participants as well as the structure by which they came to exist. Further, the conduct of each individual is constitutively independent of the conduct of others, but serves as a source of ‘‘compensable deformations that can be described as meaningful in the context of the coupled behavior’’ (Varela 1979, p. 49).

The convergence of communication via emergent systems (Kincaid 1988) is then a coupling of the individuals, in which the individual organizes the internal structure of the

123

Syst Pract Action Res

group to adapt to the environmental forces. It is important to maintain internal organization, enabling this structural coupling to evolve based on pattern recognition and accommodating replication. A set of rules of interaction can maintain the cooperative evolution of a group regardless of the shifting of group members (e.g. protester turnover) or the setting the group is in.

The following section introduces a specific context, bird flocks, that characterizes selforganization and is used as a framework and metaphor for self-organizing group interaction.

Boids In 1987, computer scientist Craig Reynolds undertook the task of creating a dynamic computer rendering of a bird flock, where each movement of each bird (which Reynolds, being from New Jersey, called ‘‘boids’’) is not pre-programmed, but arises organically from a set of predetermined rules. He comments on flocks, A flock exhibits many contrasts. It is made up of discrete birds yet overall motion seems fluid; it is simple in concept yet is so visually complex, it seems randomly arrayed and yet is magnificently synchronized. Perhaps most puzzling is the strong impression of intentional, centralized control. (Reynolds 1987, p. 2) As Reynolds was working with the representation of such group movement, he derived three simple rules that can incorporate the vast complexity of a flock, and render an accurate rule-based simulation of its behavior (Reynolds 1987).

Rule 1. Separation: avoid crowding local flockmates.

Rule 2. Alignment: match average heading with nearby flockmates.

Rule 3. Flock Centering: move to the average position of nearby flockmates.

When birds following these rules they can avoid environmental objects as well as split of from, and rejoin, the flock.

Reynolds’ ability to capture coordinated evolution in a flock setting is extraordinary, yet to apply this phenomenon to human organization is quite a different task; there are similarities and differences. Similar to a flock, emergent human interaction (such as decentralized coordination) is comprised of a series of autonomous individuals who ‘‘move’’ (e.g., to protest a political outcome) so that no individual predicts precisely where the group is going at each iteration, nor does any individual decide the moves at all times. Yet, the group moves as a whole where all members participate in an egalitarian fashion, such that everyone has equal control over the course and outcome of the interaction. Unlike a flock, human groups often need to keep in mind the goal of the interaction (although certain interactions exist with the absence of an end goal, e.g. chaotic revolution).

To sum up, we discussed conceptions of emergence, cooperative evolution, and nonlinear dynamics, to establish a basis for the explication of the flock model. Emphasis is placed on the importance of a non-reductionist approach of emergence, the importance of rules, and interaction complexity.

The Flock Model Combining the central concepts of emergence (Hodgson 2000) and autopoiesis (Maturana and Varela 1980) as explanatory processes, flock theory models the self-organizing principles of cooperative evolution in human interaction. The structure of the theory, based on the rules ‘‘template’’ that Reynolds (1987) used to simulate a bird flock, is extended to include concepts based on social science research (e.g., leadership and centralization

123

Syst Pract Action Res

concerns). Reynolds’ three simple rules incorporated the vast complexity of a flock, and render an accurate rule-based simulation of its behavior. When the birds in the computer simulation (called boids) followed these rules they can avoid environmental objects as well as split of from, and rejoin, the flock. The coordinating ability of birds (Reynolds 1987) is viewed as an exemplar that is used to elucidate structure, while simultaneously establishing three mechanisms of interaction, presented below.

Structural Distance Structural distance, based on Reynolds’ (1987) first rule—separation, captures the concept explained by Eisenberg (1990) as the balance of autonomy and interdependence, the ‘‘close but not too close, far but not too far’’ element. Groups that foster excessive autonomy may dissolve and groups that foster too much interdependence may stifle creativity (Tjosvold et al. 1991). Organization is created by the shared repertoire of communicative behaviors; the members of a group need to maintain a level of cohesion that allows for individual input without sacrificing group acceptance of new information (be it an idea for a task group, or a new key in a musical interaction). A group can become too close and corrode the ability for the group to navigate and change direction when necessary, which is common in some political protests.

Collaboration If structural distance is to be maintained in evolutionary processes, then the direction of change and the rate of change need to be a cooperative function amongst the group, as in Reynolds’ second rule—alignment. To have a constantly changing, emergent, and shared understanding, the group members must attempt to match both the direction (norms) and velocity (tempo) of the other members. Central to this construct is the creation and maintenance of a group culture, or the norms and tempo that allow group members to reach a mutual understanding of both their interaction direction and velocity.

The purpose of direction matching is to allow the group to evolve in a collaborative manner, while maintaining the structural properties of the system via normative behavior.

Decentralization If a leadership role is present, it must shift in a manner such that no one actor maintains leadership for too long, and that the group is led in a purposeful direction, which hinges largely on the differences between individual and group emergent leadership.

It is extremely pertinent in the case of political protests, where each member of the group has embedded knowledge, such as the presence of threats, which may not be available to any other group members. It is in this sense that the leadership shifts and the group must self-organize with a purpose.

It is helpful to conceptualize a decentralized leadership as a ‘‘goose rule,’’ where a goose flock must shift leadership in an effort to maximize energy decay. Energy decay can be related to groups in that a leader can exhaust their energy within the group, and the individual that has not led for the longest time has build up the most potential energy, and should then lead in one of the successive moves. This also guarantees the efficient use of intellectual capital, as well. Findings presented earlier regarding the success of decentralized organizations support the use of decentralized structures.

Maturana and Varela (1980) note the importance of structural coupling in terms of autopoietic systems, and this concept is applied to group leadership as self-organizing.

Thus, the role of leader is indeed passed from one actor to another in such a manner that intersubjectivity exists in the transformation of leadership shift.

Flock theory, as presented above, expresses a model for decentralized group organization. However, the means by which the group communicates and maintains a cogent

123

Syst Pract Action Res

knowledge of group goals and status (i.e., global information) needs to be established. The following two sections cover homophily of foci as a norm that produces contagion in the group, and a roadmap based coordination system, where there can be local organization and global information.

Norms Groups of individuals often display coordinated behavior, frequently without a central authority to intervene, based on the presence of norms (Axelrod 1997). A norm exists in a social setting to the extent that individuals act in a certain way and can be reprimanded when not acting in this way. The current research subscribes to the notion that norms are evolutionary, and the use of them in initial interactions will propagate the establishment of the norms in a self-organizing way.

Axelrod (1997) proposed an evolutionary perspective of norms, what works well for an individual is more likely to be used again, and what doesn’t work well is likely to be discarded. From the evolutionary perspective of norms, individuals will internalize behavioral norms as more correct to the degree that they see others performing them, and that the actions of other individuals in a group are central in how every individual decides to act (Cialdini 1984). A cooperative, decentralized group structure will foster the maintenance of that structure, as cooperators are more likely to stay in a group than are defectors (Orbell et al. 1984). Axelrod also found that norms could become established surprisingly quickly, and that cooperative norms can be hurried along with relatively modest interventions.

One of specific group norms that relate to flock theory is homophily, which is discussed in the following.

Homophily Homophily, defined as the selection or presence of others who are similar (Monge and Contractor 2003) or contact between similar people occurs at a higher rate than among dissimilar people (McPherson et al. 2001), has been utilized in a variety of contexts to explain network and group dynamics. The notion that similarity breeds connection (i.e., ‘‘birds of a feather flock together’’) will result in people’s personal networks being homogeneous across many sociodemographic, behavioral, and interpersonal characteristics. Brass (1995) notes that similarity in groups can ease communication, increase predictability of behavior, and foster trust and reciprocity. Such homogeneous connections can act as a network constraint on the information people receive and the interactions they encounter (McPherson et al. 2001) Yet, in a review of the literature on diversity in groups, Mcleod et al. (1996, p. 251) found that ‘‘both laboratory and field studies have shown that heterogeneity among group members with respect to age, tenure, education, and functional area is related to group and organizational creativity, adaptability, and innovation.’’ These opposite research findings suggest the need to focus on areas of homophily that is not centered on personal status.1 1

Although much of the homophily research has been centered around status homophily (i.e. similarity based on informal, formal, or ascribed status), such as race and ethnicity (Ibarra 1995; Kalmijn 1998; and Marsden 1987), gender (Marsden 1987), age and religion (Fischer 1982), education (Marsden 1987), and age and gender in distributed teams (Yuan and Gay 2006), homophily research centered on network position and organizational foci is presented here as more closely related to the group structures discussed in flock theory.

123

Syst Pract Action Res

Festinger (1950) discussed using structural position as a means of social comparison, and thus a means of deciphering similarity. Burt (1982), Lawrence (2006), and Friedkin (1993) confirmed the initial claim by Festinger, finding that people who are more structurally similar to one another are more likely to communicate about attitudes, that their attitudes would be similar, and as a result they would have more influence over each other.

Advice, friendship, and association respond to this type of homophily as well (McPherson et al. 2001).

Research suggests that a group or organization that has a decentralized structure will foster increased interpersonal homophily and interaction based on the similarity of structural position, and a hierarchical/bureaucratic structure would inhibit interaction and divide members along structural lines. In centralized setting, people tend to be less satisfied with their interaction than decentralized groups (see Cummings and Cross 2003; Sparrowe et al.

2001). Structural homophily is the first form of homophily that augments the flock model; the second is organizational foci.

Focused activities encourage people to contact with each other and form personal relationships, and interaction is more likely to occur among individuals who share similar organizational foci (Feld 1982). The relationships formed around organizational foci are capable of reinforcing non-homophilous ties and allow them to operate as homophilous ones would (McPherson et al. 2001). Caldeira and Patterson (1987) found that ties formed around organizational foci led to shared friendship, information, understanding, and behavioral homophily. Yuan and Gay (2006) point out that groups where knowledge sharing and creation are a key it is increasingly important for individuals to have contact with people that are dissimilar from them, and allow the goals of the group to organize the ties, as seen from political protests organized through Internet by diverse kinds of people.

Combining homophily research on structural position and organizational foci crystallizes a central point of flock theory; decentralized groups cooperating for a common purpose may be able to transcend traditional norms of homophily (age, race, gender) and allow their purpose to be the tie that binds. So, as the adage says, birds of a feather do indeed flock together, but they can flock together regardless of the color of their feathers.

As organizational foci can provide a common sense of purpose, decentralized networks need to coordinate to capitalize on this purpose, the following section covers roadmap based organization as a means of this coordination.

Roadmap Based Organization Since research has shown that centralized groups and organizations inhibit performance, the solution lies in the ability for a decentralized group working on complex problems to maintain a cohesive idea of group purpose and direction. Contrary to Reynolds’ (1987) flocking simulations, where global behavior can arise from local influence, the flocking behavior explained in this section exhibits both local and global influence via communication.

Bayazit et al. (2002), researching flocking behavior in complex environments, have found that goal oriented tasks are completed by the flock with more success when the group has global information along with local organization. Reynolds’ flocking methods do not perform very well if complex navigation is needed. A map, containing a network of representative feasible paths, can provide path-planning queries in the environment. An example of such roadmap-based navigation is when a group is organizing to protest a specific political situation. Whereas each unit may be only locally organized, the global coordination of a research roadmap will diminish the likelihood of unnecessary resource duplication and the vulnerability of information distribution.

123

Syst Pract Action Res

Bayazit et al. (2002) investigated whether flocking models that use planning methods provided by roadmaps support more sophisticated group behaviors than purely locally organized flocking can allow. Two behaviors investigated will be discussed; (1) homing, where the goal is to move the entire flock from a starting point to a goal position, and (2) two kinds of exploring, covering and goal searching.

Computational experiments compared traditional flocking behaviors (i.e. Reynolds’ boids) with roadmap-based techniques as well as an ideal behavior with complete knowledge of the search status at all times (i.e. a priori knowledge of the goal). Results revealed that the ideal behavior was able to cover almost 91% of the environment in the first 30 s; the roadmap-based behavior took 90 s reach a similar coverage point of 91.6%, and the traditional (boid) flocking behavior was only able to cover just over 80%. The traditional behavior reached the 80% coverage slightly faster than the roadmap behavior due to its tendency to bounce around and discover easily accessible areas quickly.

Goal searching experiments revealed that individuals in the traditional flocking behavior were not able to reach the goal, and none of the individuals were able to discover the narrow passage out of the small region in which they started. Roadmap behaviors were very close to ideal behaviors, with the group reaching the goal only 5 s later. However, the roadmap behavior was able to get two individuals to the goal before any of the individuals in the ideal behavior.

In summary, these findings indicate that augmenting the locally organized flock model with the global knowledge of a roadmap results in more efficient goal achievement and exploration of new territory. Yet extrapolating these findings to human social systems requires explication of the means by which such systems can maintain the decentralized social structure, while simultaneously knowing what each other knows and where the group is going.

Recent research into a network phenomenon called the small world theory provides a useful vocabulary and construct to model the balance of decentralized structure and a centralized repository of knowledge.

Bipartite Small Worlds A small world can be explicated by a network that exhibits two properties; tendencies of local clustering, and short distances between nodes (individuals) such that any node could be reached in an average of only a few steps (Watts 2003). The tight clusters are groups of people in social networks that are associated with each other through redundant ties that allow for local clustering, also called strong ties or close ties. Yet, there are also a substantial amount of ‘‘weak’’ ties that we have in our networks; these are the connections that our close ties have with people that are not necessarily connected to, or have much in common, with us. Weak ties have been shown to offer a greater possibility for effective social coordination called ‘‘the strength of weak ties’’ (Granovetter 1973). So it is paradoxically the weaker ties that link people to each other and to novel resources that reside in other tight clusters within a network. Likewise, bridges between different tight clusters, represented by weak ties are catalyzed by accessibility to a large number of nodes.

The combination of tight clusters (strong ties), and the diameter of the network (quantified by the average shortest distance of links between two nodes) increasing logarithmically with the number of nodes, gives the network its small world (Amaral et al.

2000). The diameter of the network remaining small in relation to the addition of nodes allows the nodes in the network to be connected with only a few links, even if the network gets quite large. Thus, any person on the planet could be only a few links away from any

123

Syst Pract Action Res

other person by navigating their strong ties (local clustering) and utilizing the large number of weak ties associated with each strong tie (long range reachability).

There is a fundamental difference between the notion that a short path can connect any two people, and their ability to find that path (Watts 2003). Underscoring this issue is the difference between broadcast and directed searches. A broadcast search involves an individual activating every link in their network, telling everyone they know, in turn telling everyone they know, and on. For instance, Facebook ‘‘updates’’ section typically shows a broadcast of information in a network. Thus, if a short path exists the message will eventually navigate this path.

Directed searches involve the passing of a message one link at a time. A directed search is far more efficient than a broadcast search in terms of the number of messages sent, but far less efficient in finding the shortest path for that message to take. Thus the problem remains, how does an individual in a network find what they need in only a few links? The scale-free model (Baraba´si and Albert 1999) would require the use of a hub or connector to navigate to the target, but this solution still suffers from the vulnerability issue relating to hubs, and the difficulty of application to social networks. A potential answer, as it turns out, is concerned with the social lives of people and the identities they develop; specifically, the affiliation of people with their interests (i.e. a social roadmap) such as political issues, represented by affiliation networks.

In affiliation networks, or bipartite networks, two nodes can be thought of as affiliated if they participate in the same group or event (Breiger 1974). Affiliation networks differ from single mode networks in several ways. Affiliation networks are two-mode, meaning networks where a set of actors and a set of events describe collections of actions instead of simple ties between pairs of actors. Connections among members of one of the modes are based upon the linkages formed via the second mode. If two nodes A and K are a member of the same political party, then they are linked by one degree (through the second mode).

Affiliation networks is important in regard to the individuals’ membership in collectives, or social circles (Simmel 1950), which provide conditions for development of interpersonal connections. Individuals are brought together through joint participation in social events, such as protests, and participation in more than one event establishes a linkage between the two events; thus overlapping group membership allowing for the transmission of information between groups. Therefore, this can be interpreted as either actors being linked by events, or as events linked by actors (Breiger 1974).

Watts (2003) discussed the ability of multidimensional networks, where people have several dimensions to their identity—thus belonging to multiple groups, as to catalyzing an individual’s navigation of their network. Yet, this proposition still requires the use of existing social networks for the acquisition of information on the reachability of a specific person. What is being proposed here is that searching for information contained by someone that is unknown is better done by searching via the content of what is needed allowing for a more directed navigation, much like in the roadmap based flocking experiments covered above.

Applications to Flock Theory As covered in the previous section of this paper, the fundamental notions of flock theory allow for a network of individuals to: (a) manage their social structure, (b) collaborate (i.e. propose ideas and concepts that are supported by the network), and (c) maintain decentralized, cooperative evolution.

123

Syst Pract Action Res

Structure, Mode One The first element of flock theory proposes that for cooperation to occur, individuals in a social network should maintain a structure that is close-but-not-too-close (extreme cohesion) and far-but-not-too-far (breakdown of cohesive group). Notions of distance are central to the small-world construct, where local clusters are close, but not so close as to prohibit long range connections (as seen in clusters that are only strongly tied to themselves); and the random links to other clusters allow them to be far, but not too far (as is the case when short path lengths don’t exist). The challenge then becomes the management of this distance.

The balance between randomness and order is crucial in small-world networks (Watts 2003). It may be possible to apply this to the flock model: The group of individuals needs to maintain a balance between autonomy and interdependence. Notions of distance based on a balance of randomness (autonomy), and order (interdependence), are discussed here as relating to the social structure mode of a bipartite (two-mode) network.

Individuals in a group (network) maintain a structure that allows for the close reachability of other members, while simultaneously allowing a certain amount of distance, or freedom, from the group structure. Like the bipartite network discussed in the previous section, this is done in relation to the second mode of the network—the roadmap.

Roadmap, Mode Two The second element of flock theory posits that the collaboration requires the maintenance of distance, and is achieved through matching the ‘‘motion’’ of the other individuals. If distance is to be maintained in the evolutionary processes, then the direction of change (either topically or task oriented), and the rate of change needs to be a cooperative function amongst the group.

However, without the roadmap, it is difficult for individuals to understand how far they are from each other, given that distance can be both physical as well as topical.

Outcomes of group interaction can often come from the peripheries of the network instead of from the center of a network (its leaders). Seemingly small events can result from group members’ knowledge that seemingly random encounters can lead to a plethora of individual decisions, each made without any knowledge of a larger plan; yet, by some means aggregating into a momentous and unanticipated event (Watts 2003). In such cases the centralities of the individuals in the network would not reveal much about the outcome, as the center emerged as a result of the event itself. In the absence of a central actor, the attention then turns to the roadmap reveal where the hubs truly are.

If hubs form in the content mode of a bipartite network, then these content hubs allow the individuals to affiliate themselves with the topics and connect with each other in the physical mode via these affiliations. Further, the evolution of the system can be better understood as a shift of the hubs, such as a new protest forming. Individuals remain clustered in their networks and navigate their distance based on their affiliation with the changing topic, or their position on the roadmap. Thus, supporting an idea in an online political forum allows the other individuals to have a deeper understanding of how far they are from each other. It is important to note that if a topical direction change is to occur, the structure and connectivity of the network can have as much of an influence on the likelihood of adoption as the attraction of the idea itself (Watts 2003), so the social distance mode is a precursor to the development of the content mode.

A group’s reception of a directional change, or shifting of a topical hub, can be conceptualized as an information cascade, or the widespread reception of an idea. Information

123

Syst Pract Action Res

cascades occur when there is a specific network structure that allows for the breadth of the individuals in the network to be receptive to some specific information. It is in this sense that the combination of both modes of a bipartite network must be considered to understand how decentralization can exist amongst a set of individuals. The positing of an idea allows people to attach to it, either supporting it or not. If there is to be a cooperative evolution amongst the group, directional changes need to be initially supported (creating a new hub), as the individuals will be structurally linked to each other via new ideas and, at least initially, thus the distance is also maintained. Individuals in the network can maintain a close but not too close distance by connecting through the hubs, as well as supporting directional changes by allowing the new hubs to form and develop.

Even a fairly small group of individuals interacting can maintain a small-world structure, where everyone is only a few degrees of separation from each other, embodied by central hubs as well as local clustering. Yet there is still an important element to consider in this model—the avoidance of central hubs in the distance (social) mode of the network, as this can lead to network vulnerability and the end of a protest. Hubs in the social mode can lead to dominant leaders as well, breaking down the cooperative element. Both of these issues can be resolved by shifting any presence of a leader (or at least having a decentralized decision process) in the social mode of the group, often as a result of individuals posing new ideas or topics.

Decentralized Evolution The final element of flock theory posits that decentralization is crucial, and leadership amongst the group must shift in order to maintain an egalitarian, decentralized evolution.

Instead of a network where a central leader directs network evolution, such as in the hub dominated scale-free small-world, important innovations can originate from the less central individuals (Cummings and Cross 2003; Shaw 1964; Sparrowe et al. 2001). The first two elements of flock theory, discussed above, unpack the means by which a network can have the content of the interaction and predict the connections that form amongst people, yet there is still a possibility that all of the ideas presented will come from a small set of individuals (i.e. core/periphery), even in the case of an information cascade. To resolve this issue the group should attempt to shift the sources of input through the entire group so each individual can update the roadmap, as is the case with many of the communication technologies.

In small-world terms, the presence of a strong leader will have a similar effect as a hub, dominating the network and making the least connected nodes practically invisible. More importantly, the less connected nodes would have to go through the hub to access each other, which creates a situation that is far from egalitarian. Likewise, the group would be very susceptible to breakdown if the central hub were to possess an inability to solve a specific problem or distribute information properly. Much like a hierarchical organization, information brokering by individuals in upper levels of the hierarchy can often lead to information overload and a limited response to environmental ambiguity (Lewin and Stephens 1995).

Ambiguity requires communication between people who have a variety of expertise and are thus mutually dependent, thus requiring distributed communication in an informationprocessing network (Stark 1999). A successful information-processing network will then distribute information to the roadmap with as much equality as possible, where a hub based scale-free model would require the central actors to unrealistically process an exponential amount of information. A robust information-processing network needs to distribute the workload of leadership as well as the redistribution of information; in line with the flock

123

Syst Pract Action Res

model, where the leadership shifts and information stems from the breadth of individuals.

The research on decentralization presented earlier underscores the need for distributed leadership and information distribution.

A particularly fitting example of distributed leadership and information distribution is the trade off of specialization and cross-fertilization in coordinated political protests. The division of labor in protest groups requires a large amount of information retrieval and sharing, yet the leadership is often widely distributed. Difficulties of distributed leadership are compounded by the frequency of the protest group members often having an equal organizational rank (e.g. all common citizens). Protest members are often spread across different specializations, creating communication problems. Combining distributed leadership of equal stature with cross-fertilization creates many potential problems for protest groups, problems that may be explained by the model proposed in this paper.

Case Studies of Flocking Political Protests Political rallies organized using communication technologies have garnered considerable interest by social scientists and cyber-culture critics. Rheingold (2003) named the organizing of people through the mediation of new network technology as a ‘‘smart mob.’’ Such organizing minimizes costs to set up and develop relationship because cyberspace offers a high level of freedom of time and space for the participants. Like thousands of Atlantic herring migrate thousands of kilometers in a form of self-organizing systems, people utilize Internet as a medium of ‘‘the speed of thought’’ (Gates 1999) to plan space and time for their meeting online and offline. Actually the ‘‘planning’’ does not involve any coercive leader, and rather anyone can play a role in constructing the gathering. For instance, the two cases in the following section has a commonality that they were suggested by a single individual or a small number of people, but people’s decision on ‘‘when and where’’ to meet in the offline space was voluntarily made (e.g., Seoul City Hall Square) and shared by other web users.

The drastic downfall of communication cost through the web invites an increased political self-efficacy among web users. For instance, political bloggers tend to share antiestablishment mentality and belief in the web as an alternative and ‘‘real’’ democratic space to existing political system (Matheson 2004). This type of mentality may become a basis a courageous call for an offline rallies through their favorite channel, the web. This trend is also called ‘‘self-representative politics’’ (Coleman 2004) or ‘‘advocacy democracy’’ (Dalton et al. 2003).

Case 1: Anti-FARC Rallies Ignited on Facebook On February 4, 2008, millions of Colombians marched against The Armed Forces of the Colombian Revolution (FARC). The protest occurred simultaneously in 27 cities throughout Colombia and 104 major cities around the world, mainly in Latin America, but some in Europe, Asia, and the United States. ´ scar Morales, This international event was triggered by a young Colombian engineer, O using the social networking website Facebook.com. This event was not the first grassroots political movement organized through Facebook (‘‘Facebook becomes a force,’’ Pe´rez 2008). In October 2007, thousands of people around the world took to the streets to demonstrate against Myanmar’s repression of pro-democracy marches organized by that country’s monks. This event was planned and organized through Facebook by a student group to support the monks.

123

Syst Pract Action Res

The FARC is one of best-known major Marxist guerrilla groups in Colombia. As the largest group, with approximately 17,000 members, FARC claims to represent the rural poor in opposition to Colombia’s wealthy classes and opposes US influence in the region and neo-liberal policies. The FARC’s financial base is built on kidnapping, extortion of licit and illicit businesses in regions under the group’s control and the theft of municipal funds. Through these methods, the FARC has become one of the world’s most successful self-financed groups of insurgents (Shifter 1999; Sa´nchez and Palau 2006).

The FARC is reportedly now holding approximately 770 captives, who are an important source of revenue (Colombia’s conflict 2007). These captives include, for example, politicians, members of the Columbian security forces, and US counternarcotics contractors.

Some captives have been held for almost a decade.

Even though some of the protests against the FARC had been organized by Colombian politicians and nongovernmental organizations (NGOs), their efforts failed to prevent the FARC from engaging in acts of kidnapping and violence. For example, two of the most notable early anti-FARC protests occurred in October 1999 and July 2002. The October protest took place in Colombia, attended by over a million demonstrators. It was organized by an association called No Ma´s, led by the country’s then vice president, Francisco Santos Caldero´n. The July event in Bogota’s central square was organized by Colombian women and was attended by 20,000. Even though both of these protests were fairly large, they were structurally centralized in nature and thus could not garner widespread support and exponential numbers.

Two more recent incidents finally led Morales to use Facebook to organize a virtual protest. Facebook is an example of a social-networking site, a form of communication technology within which users can form social links to other users or groups. These social links allow for the management of a large number of network ties, and can be shared and added very easily. Social-networking sites such as Facebook act as roadmaps for users’ personal networks, as well as interest groups, and provide the means by which users can organize and collaborate without geographic and temporal limitations. It is in this sense that Facebook as an information roadmap allowing for the protesters to flock in their interests.

In June 2007, 11 politicians among the high-profile FARC hostages were killed by close-range small-arms fire in a jungle camp. After this violence, Venezuelan President Hugo Chavez suggested that the FARC should be recognized as a legitimate army instead of as a terrorist group. According to him, this recognition would likely to win the release of dozens of other hostages. But President Chavez’s suggestion sparked large disputes in Columbia as well as in other countries. At this time, Morales and his friends started the group ‘‘A Million Voices Against the FARC,’’ in Facebook. The group grew in just a few weeks into a virtual protest. Bolstered by the enormous response from other Facebook users, the support of the Colombian government, many NGOs, and some political parties, Morales decided to call for an international event against the FARC. By using the communicative roadmap that Facebook offers, protesters were able to join the cause by decentralized means and via informational, knowledge networks instead of traditional structural social networks.

Case 2: South Korean Protests Against Unrestricted US Beef Import A similar case of the Internet as a tool for decentralized political protests occurred in Korea. Health-related risks like mad cow disease, which people face without geographic and time limits in post-modern society, create strong emotions because they involve a

123

Syst Pract Action Res

potential but fatal disease (Beck 1992; Rozin et al. 1993; Washer 2006). Since people infer risk directly from their emotions, not from evaluating consequences cognitively, and also perceive risks on the basis of anxiety and lack of control over potential hazards, managing and analyzing emotions, reasons, and risks are essential to policy-makers (Slovic 1987; Loewenstein et al. 2001; Sinaceur et al. 2005).

The Korean government’s efforts, however, seemed to be ineffective in alleviating people’s feelings about perceived risks. On April 18, 2008, the Korean government agreed to re-open its market to US beef after a ban of more than 3 years. The government agreed to remove restrictions on certain cuts and, conditionally, ages of US beef. On April 29, people’s sentiment had been elevated by an in-depth report by MBC, a major Korean broadcasting network, on the potential dangers of human derivative variant CreutzfeldJacob Disease (vCJD), the so-called human mad cow disease. MBC also reported the fact that downed cattle were slaughtered and sold in the US in early 2008, leading to a massive recall of US beef, as it has a higher risk of having mad cow disease (Korean Herald 2008).

The television program’s most sensational excerpts—with the support of scientists’ serial studies that the infectious agent of Bovine Spongiform Encephalopathy (BSE) is distinctively endurable in high temperature and is not destroyed during normal cooking and can remain dormant for decades without any particular symptoms—were spread online through blogs, video-sharing websites, and discussion boards.

These worries about US beef stirred up teenagers who use the Internet daily and caused them to respond first by attending weekend protests at the center of Seoul. These small public incidents triggered huge on-street protests by common citizens, including workers, students, housewives, and members of religious groups. Most of protestors held candles during the night to demonstrate against beef imports from the US. However, the Korean government’s lukewarm attitude about this issue caused massive public anger against President Lee, Myung-bak. Eventually, the protest developed into an anti-government movement online and on the streets that continued for 5 months on a nationwide scale.

In this case, the Internet was used as a method of public participation in politics and for individual journalism. People who participated in the on-street beef protest equipped themselves with various digital devices, such as cameras, webcams, mobile phones, and laptop PCs with wireless access to deliver vivid live streaming of the public anti-beef demonstrations to others everywhere. Furthermore, the Internet incapacitated the power of traditional media. From the beginning, some of major newspapers, which sometimes have been accused to have a pro-government bias, tried to appeal to reason and rationality with scientific and statistical evidence in order to mollify the public’s ire and fears. However, these attempts were overwhelmed by sensational online videos and photos, for example, the scary scenes of the mad cows’ slaughter, interviews with relatives of people who might have vCJD, and police arresting crying middle-school girls who were taking part in the candlelight vigil protests. One of the largest web portals in Korea, Daum ( http://www.daum.net), and its subpage ‘‘Agora’’ (http://agora.daum.net) became a major channel of sharing information about organizing of the rally. Web users enthusiastically updated the situations on the street, government actions, and discussed the direction of protests. Due to a rapid hike of readership and posting, Daum became an ‘‘agora’’ for Korean Internet users. Also, rally supporters who could not go to streets volunteered to become a local hub which distributes information about rally updates, street crowdedness, a possible outlet from police arrest, and even the best way to go back home from rally sites.

With Internet and video-enabled cell phones, Korean protesters maintained the proper level of information throughout their rallies.

123

Syst Pract Action Res

A specific video clip was especially prominent which captured the scene where a college female student who was hiding under a bus from helmeted policemen’s violent quelling of rallies was kicked by police troops ignited a mass outrage and resulted in mass resistance and protest. Because the Agora page was acting as the affiliation mode of the protesters networks, their roadmap, the video clip was opened by an Internet media and shared by numerous people at the speed of ‘‘thought’’ (Gates 1999).

The chorus of efforts by the distributed and decentralized protesters was successful, and the Korean government met the demands of the protest. The protesters were able to act as a coordinated flock, where local information could lead to global coordination, and information could be shared and retrieved without the need for a central actor.

Interpreting Two Cases from the Flock Model Table 1 summarizes how the flock model can be applied to interpret two cases portrayed above. First of all, throughout the two cases, people’s emotional arousal and cognitive understanding on the issues including human rights violation by FARC and perceived risk of vCJD by importing US beef were coordinated by participants for peaceful offline demonstrations and online community formation. In this process, ‘‘weak ties’’ based on shared intent and concerns were utilized by participants to maintain their social cohesion just enough to initiate and maintain the collaborative actions. This regulation was performed both in cyberspace and ‘‘real’’ space.

Second, group norms and culture were developed among participants. However, these culture and norms were not directed, fixed, or coerced by any institution. Rather, the selfestablished norms were fluid enough to adapt to changing environment. In the Korean case, throughout the on-street demonstrations lasting around 100 days, the frequency of violence was maintained at the low level because protesters sometimes restricted other participants’ attempt for physical collision with police. This nonviolent culture contributed to the attraction of hard-to-join social classes such as moms with babies, junior high school students, and females in the ‘‘peaceful’’ rallies.

Also, it should be noted that there was the coordination (intended or not) between homophily and diversity in the anti-FARC and Korean cases. Participants were composed of different kinds of people in terms of gender, age, vocations, ethnics, and social classes in the two protests. This diversity was possible because they flocked not based on their perceived similarity of social traits, but of shared interests and concerns.

New communication technologies (e.g. Facebook and other social media) were used to share the information, mobilize people with similar thoughts and positions, and set up and adjust the time and location of offline rallies. Namely, social media such as Facebook, blogs, and mobile technology were used as a ‘‘roadmap’’ for people to create and maintain group norm and culture such as non-violence, voluntary sharing of knowledge, schedule coordination, and express humanity by helping other protesters in need.

Third, decentralized protests promoted increased homophily among participants by a reinforced sense of belonging and heightened commonality of sentiments. Under the atmosphere of voluntarism among protesters, leadership emerged and shifted flexibly according to changing situation and need. One of the causal factors of these flexible leadership changes was availability of information. For instance, communication technologies were sometimes used as a real ‘‘roadmap.’’ Protesters who lost their way because of police blockade got help from those who stay home but watch roadside cameras offered by police website and other protesters’ real time update of knowledge useful for fleeing away from police arrest (in the Korean case). Those who received useful information

123

Syst Pract Action Res Table 1 The flock model: an application to political protest context Flock model elements

Application to political protests

FARC and Korean cases

Separation: Maintaining structural distance to avoid crowding local flockmates

Managing cohesion level that allows Participants maintained their optimum level of cohesion for individual input without emotionally (in cyberspace) and sacrificing group’s acceptance of physically (on the street). For new information instance, protesters in the Korean Not too high or low cohesion level case shared information of helped maximize the benefit from crowdedness in major streets of weak ties as well as sense of Seoul to avoid physical belonging overcrowdedness

Alignment: Collaboration to change or adjust the direction and velocity

Creation and maintenance of group 1. Those who wanted to use physical violence against police were culture and norm among discouraged by many other protestants participants requesting For instance, group culture of ‘‘nonviolence.’’ (in Korean case; a voluntarism, nonviolence, and humanity to help other participants formation and maintenance of nonviolent culture) in need were formulated and maintained among participants of 2. The harmony of homophily (involvement in humanity and the two cases public policy) and diversity (participants are composed of different kinds of people in terms of gender, age, vocations, and social classes)—a two mode network 3. Internet as a roadmap (using blogs, social network sites, and mobile Internet technology in both cases)

Flockcentering: Decentralizing. Movement to the average position of nearby flockmates

1. Decentralized protests promoted Leadership shifts according to increased homophily (sense of availability of information among belonging; shared sentiments) protestants 2. Those who have the information of crowdedness and risk of police arrest voluntarily shared it with participants and those who received the information led crowds to safer path (Korean case)

through social media diffused their information voluntarily with others. Concurrently, leadership changed to adapt changing situations.

Limitations of Self-Organizing and Flock Theory The limitation of the flock model is in a situation where there is a social dilemma, such as civil unrest or non-peaceful protests, or when order needs to be restored to a highly entropic situation. Solutions to such situations cannot rely on the group to self-organize, as the group may not have all of the components necessary to produce organization. Thus, organization must arise from an outside force, e.g. a police officer gaining control over a rapidly decaying situation of unrest. This is not to say that there will not be a continued decrease in organization and order with the presence of an organizer (i.e. police officer), yet it may be the only means of order.

123

Syst Pract Action Res

Another situation where a self-organizing system is undesirable is where the individuals within the system desire organization and the decrease of entropy, yet cannot figure out the means by which cooperate with each other to organize. An example of this would be a large population of people that wish to protest a specific governmental decision, yet do not have the means by which to organize. Such situations represent systems that benefit from an external organizer; such people are often fittingly called organizers. However, if the system lacks an organizer, coordinated roadmap communication can be used as a means for self-organization, as covered earlier in this paper.

Perhaps the most undesirable effect of self-organization is in the case of herd behavior.

Commonly discussed in relation to flock behavior, herd behavior is characterized by local organization, similar to a flock, but is far more vulnerable to dangerous headings that can lead to system destruction. Much like a herd of cattle leading each other off of a cliff, herd behavior can typify situations where the lack of a central point of organization allows even a small direction change to cause widespread entropy.

One of the most common contexts that herd behavior is used both as metaphor and measure for social movement is in economics and investment trends, but is certainly possible in a protest situation. A wide range of research has used the herd metaphor as a basis for economic research and the explanation of cascading trends in investment (see Abhijit 1992; and Bikhchandani and Sharma 2001).

Findings across the financial disciplines suggest that investors and analysts will make investment moves similar to that of a herd, even when they believe the investment is a bad decision. A principle reason for this misguided action is that the investors believe that others may know something about the return on investment and their actions reveal this.

Central to this kind of herding is the lack of knowledge about what the other individuals know, i.e. allowing local organization to be predicted by only localized action, disregarding the need for global information and coordination. Herding situations can specifically be avoided when using a roadmap-based coordination in a flock model, as proposed in the current research.

For example, consider fifty politically motivated activists each with their own assessment, possibly different, of the benefits of protesting a specific political situation. Ten of these activists believe that the protest is sensible and the remaining 40 do not. Each activist knows their own estimate of outcome, yet they do not know the estimate of others’ or which way a majority of the activists are leaning. If they shared their information, they would concretely decide that the opportunity is not worthwhile. However, the individuals do not make their information public, and they are not making their decisions at the same time. Thus, if the first several protestors that make their move are amongst the ten that believe positively, others that were initially more pessimistic may revise their conclusions assuming that others knew something that they didn’t, and protest.

Herding examples such as these are not rare in emerging market investments, and the example provided above is very typical if the activists were replaced by investors in the example, which clearly points out a potential problem with this form of self-organizing system; without a means of real-time information sharing and retrieval a decisionmaking group can easily become a blind herd headed for a cliff. In the example presented above, if the activists pooled their knowledge they would have come to a collective, self-organized decision not to protest. The current research stream presented a model by which a group can be both self-organized and able to react to potentially dangerous outcomes.

123

Syst Pract Action Res

Conclusion The cases presented above have brought about various political and social controversies over issues such as the role of the Internet in social movements and people’s political participation around the world. The events have typified the momentum of political protestors recognizing the potential power of the Internet as a means to foster and maintain a decentralized, coordinated group focused on a particular problem.

In both protest cases presented, the Internet effectively served as a road map for searching, collecting, and exchanging information, and catalyzing the decentralized coordination of a political movement. The rapid development of communication technologies is increasingly enabling anyone to take part in an online community or online discussion related to issues that matter to them most, and organize about these central foci.

Additionally, there is minimal time and resource cost surrounding the organization, further enabling these groups to self-organize. Various factors, such as new policy proposals from the government and potential risks of these new policies to the governed, are instantly evaluated by people and collectively evaluated via the online communication. The issue or policy spreads among people interacting by using communicative roadmaps, such as social networking websites, and protests can form as needed.

Future research should include technologies that are emerging as important catalysts of coordinated protests. For example, Twitter has become a valuable tool and should be included as an important tool in the toolbox of decentralized coordination. Likewise, context aware mobile technologies are also becoming more and more powerful in that users can be aware of the location of other users without having to explicitly communicate this information. Such capabilities allow for a much more accurate roadmap and allow for more individual attention to be placed on important content of interaction instead of location.

In addition to including newer communication technologies, future research should also seek to analyze the actual social networks of protests groups in a dynamic manner to uncover the detailed means of self-organization. Analyzing longitudinal data will likely give a more detailed account of the mechanisms and messages that are most important to a successful decentralized protest. Dynamic network analysis has traditionally been difficult, but recent advances in software development are beginning to open those doors to social science researchers.

The flock model presented above serves as an elaboration of the mechanisms that allow decentralized political protests to occur, and can elucidate the structural advantages of Internet technologies for these purposes. However, the use of communication technology as a roadmap is not limited to political protests; many other contexts such as social support groups, resource exchange sites, online consumer activism, and environmental movements all benefit from the decentralized nature of the Internet. It is indeed a small world, and instrumentally navigating that small world is integral for bottom up self-organization and egalitarian evolution. People power can overcome, but organizing that power is paramount.

References Abhijit B (1992) A simple model of herd behavior. Quart J Econ 107:797–818 Amaral L, Scala A, Barthelemy M, Stanley HE (2000) Classes of behavior of small-world networks.

Proc Natl Acad Sci 97(21):11149–11152

123

Syst Pract Action Res Axelrod RM (1997) The complexity of cooperation: agent-based models of competition and collaboration.

Princeton University Press, Princeton Baraba´si AL, Albert R (1999) Emergence of scaling in random networks. Science 286:509–511 Baron JN, Hannan MT (2002) Organizational blueprints for success in high-tech startups: lessons from the Stanford project on emerging companies, California Management Review Bayazit O, Lien J, Amato N (2002) Roadmap-based flocking for complex environments. In: Proceedings of the 10th Pacific conference on computer graphics and applications (PG’02), IEEE Beck U (1992) The risk society. Sage Publications Ltd, London Berliner P (1997) Give and take: the collective conversation of jazz performance. In: Sawyer RK (ed) Creativity in performance. Sage, Greenwich, pp 9–42 Bikhchandani S, Sharma S (2001) Herd behavior in financial markets. IMF Staff Pap 47:279–310 Brass DJ (1995) A Social Network Perspective on Human Resource Management. Res Person Hum Resour Manag 13:39–79 Breiger R (1974) The duality of persons and groups. Soc Forces 53:181–190 Burt RS (1982) Toward a structural theory of action. Academic, New York Caldeira GA, Patterson SC (1987) Political friendship in the legislature. J Polit 49:953–975 Cialdini RH (1984) Influence—how and why people agree to things. Morrow, New York Coleman S (2004) Blogs as listening posts rather than soapboxes. In: Ferguson R, Howell M (eds) Political blogs: craze or convention?. Hansard Society, London Colombia’s conflict (2007) Strategic comments 13 (10) Contractor N (1994) Self-organizing systems perspective in the study of organizational communication. In: Kovacic B (ed) New approaches to organizational communication. SUNY Press, Albany, pp 39–65 Contractor NS, Grant S (1996) The emergence of shared interpretations in organizations: a self-organizing systems perspective. In: Watt J, VanLear A (eds) Dynamic patterns in communication processes. Sage, Newbury Park, pp 216–230 Contractor NS, Seibold DR (1993) Comparison of structurational and self-organizing perspectives to the study of GDSS. Hum Commun Res 19:528–563 Crossley N (2003) Even newer social movements? Anti-corporate protests, capitalist crises and the remoralization of society. Organization 10(2):287–305 Cummings J, Cross R (2003) Structural properties of work groups and their consequences for performance.

Soc Networks 25(3):197–210 Dalton RJ, Scarrow SE, Cain BE (2003) Democracy transformed? Expanding political opportunities in advanced industrial democracies. Retrieved October 15, 2008 from http://repositories.cdlib.org/cgi/ viewcontent.cgi?article=1021&context=csd Darley V (1994) Emergent phenomena and complexity. In: Brooks RA, Maes P (eds) Artificial life IV, proceedings of the fourth international workshop on the synthesis and simulation of living systems.

MIT Press, Cambridge, MA, pp 411–416 Eisenberg EM (1990) Jamming: transcendence through organizing. Commun Res 17:139–164 Feld S (1982) Social structural determinants of similarity among associates. Am Sociol Rev 47:797–801 Festinger L (1950) Informal social communication. Psychol Rev 57:271–282 Friedkin NE (1993) An expected value model of social exchange outcomes. Adv Group Process 10:163–193 Fuchs C (2006) The self-organization of social movements. Syst Pract Action Res 19(1):101–137 Gates BH (1999) Business at the speed of thought: using a digital nervous system. Grand Central Publishing Gilbert N (1995) Emergence in social simulations. In: Gilbert N, Conte R (eds) Artificial societies: the computer simulation of social life. University College London Press, London, pp 144–156 Gleick J (1988) Chaos: the making of a new science. Heinemann, London Goldstein J (1999) Emergence as a construct: history and issues. Emergence 1:49–72 Granovetter MS (1973) The strength of weak ties. Am J Sociol 78:1360–1380 Hodgson G (2000) The concept of emergence in social science: its history and importance. Emergence 2(4):65–77 Ibarra H (1993) Network centrality, power, and innovation involvement. Acad Manage J 36(3):471–501 Ibarra H (1995) Race, opportunity, and diversity of social circles in managerial networks. Acad Manage Rev 38:673–703 Kalmijn M (1998) Intermarriage and homogamy: causes, patterns and trends. Ann Rev Sociol 24:395–421 Kincaid DL (1988) The convergence theory and intercultural communication. In: Kim YY, Gudykunst WB (eds) Theories in intercultural communication. Sage, Newbury Park, pp 280–298 Korea Herald (2008) Fear over U.S. beef. http://www.koreaherald.co.kr/. Accessed 5 Jan 2009 Krippendorf K (1991) Society as self-referential. J Commun 41:136–140 Lawrence BS (2006) Organizational reference groups: a missing perspective on social context. Organ Sci 17:80–100

123

Syst Pract Action Res Lewin AY, Stephens CU (1995) Designing postindustrial organizations: combining theory and practice.

In: Huber GP, Glick WH (eds) Organizational change and redesign: ideas and insights for improving performance. Oxford University Press, New York, pp 393–409 Loewenstein GF, Weber EU, Hsee CK, Welch N (2001) Risk as feelings. Psychol Bull 127:267–286 Luhmann N (1995) Social systems. Stanford University Press, Stanford Marsden PV (1987) Core discussion networks of Americans. Am Sociol Rev 52:122–313 Matheson D (2004) Weblogs and the epistemology of the news: some trends in online journalism. N Media Soc 6(4):443–468 Maturana H (1975) The organization of the living: a theory of the living organization. Int J Man-Machine Stud 7:313–332 Maturana HR, Varela FJ (1980) Autopoiesis and cognition: the realization of the living. D. Reidel, The Netherlands Maturana HR, Varela FJ (1987) The tree of knowledge: the biological roots of human understanding.

Shambhala, Boston McLeod PL, Lobel SA, Cox TH Jr (1996) Ethnic diversity and creativity in small groups. Small Group Res 25:267–293 McPherson M, Smith-Lovin L, Cook J (2001) Birds of a feather: homophily in social networks. Annu Rev Sociol 27:415–444 Monge PR, Contractor N (2001) Emergence of communication networks. In: Jablin F, Putnam L (eds) Handbook of organizational communication, 2nd edn. Sage, Thousand Oaks, CA. Monge PR, Contractor N (2003) Theories of communication networks. Oxford University Press, New York Monge PR, Eisenberg EM (1987) Emergent communication networks. In: Jablin F, Putnam L, Roberts K, Porter L (eds) Handbook of organizational communication. Sage Publications, Newbury Park, pp 304– 342 Nicolis G (1989) Physics of far-from-equilibrium systems and self-organization. In: Davis P (ed) The new physics. Cambridge University Press, Cambridge, pp 316–347 Orbell JM, Schwartz-Shea P, Simmons RT (1984) Do cooperators exit more readily than defectors?

Am Polit Sci Rev 78:163–178 Pe´rez MC (2008) Facebook becomes a force for good. The International Herald Tribune:13 Prigogine I, Stengers I (1984) Order out of chaos: man’s new dialogue with nature. Heinemann, London Reynolds C (1987) Flocks, herds, and schools: a distributed behavioral model. Comput Graph 21:25–34 Rheingold H (2003) Smart mobs. Perseus Publishing, Cambridge Rosen D (2002) Flock theory: cooperative evolution and self-organization of social systems. In: Proceedings of the 2002 CASOS (computational analysis of social and organizational systems) conference.

Carnegie Mellon University, Pittsburgh, PA. Rosen D (2008) Cooperation and coordination in decentralized communication networks. In: Proceedings of the 41st Hawai’i international conference on system sciences. Institute of Electrical and Electronics Engineers, Inc. (IEEE), New Brunswick.

Rozin P, Haidt J, McCauley CR (1993) Disgust. In: Lewis M, Haviland JM (eds) Handbook of emotions.

Guilford Press, New York, pp 575–594 Sa´nchez F, Palau M (2006) Conflict, decentralization and local governance in Colombia, 1974–2004. CEDE document 2006–20. Economics Department, Universidad de los Andes, Bogota´ Shaw M (1964) Communication networks. In: Berkowitz L (ed) Advancements in experimental psychology.

Academic Press, New York Shifter M (1999) Colombia on the brink. Foreign Aff 78(4):14–20 Simmel G (1950) Individual and society. In: Wolf KH (ed) The sociology of George Simmel. Free Press, New York Sinaceur M, Heath C, Cole S (2005) Emotional and deliberative reactions to a public crisis. Am Psychol Soc 16(3):247–254 Slovic P (1987) Perception of risk. Science 236:280–285 Sparrowe RT, Linden RC, Wayne SJ, Kraimer ML (2001) Social networks and the performance of individuals and groups. Acad Manage J 44:316–325 Sperry RW (1986) Discussion: macro- versus micro-determinism. Philos Sci 53:265–270 Stark DC (1999) Heterarchy: distributing authority and organizing diversity. In: Clippinger JH (ed) The biology of business: decoding the natural laws of the enterprise. Jossey-Bass, San Francisco Tjosvold D, Andrews IR, Struthers JT (1991) Power and interdependence in workgroups. Group Organ Stud 16(3):285–299 Varela F (1979) Principles of biological autonomy. Elsevier, New York

123

Syst Pract Action Res Washer P (2006) Representations of mad cow disease. Soc Sci Med 62:457–466 Watts DJ (2003) Six degrees. W. W. Norton and Co, New York Wheeler WM (1928) Emergent evolution and the development of societies. Norton, New York Yuan YC, Gay G (2006) Homophily of network ties, and bonding and bridging social capital in distributed teams. J Comput Mediat Commun 11(4)

123