Al-Muhajiroun and the terrorist groups Provisional Irish Republican Army (PIRA) and ... organizational goals in light of their opponentsâ actions. ..... groups,â we are also interested in studying how traditional social metrics differ across ..... also provide us with material for timeline construction, as well as event data generation.
Adversarial Behavior in Complex Adaptive Systems: An Overview of ICST’s Research on Competitive Adaptation in Militant Networks1 John Horgan, Michael Kenney, Kathleen Carley, Mia Bloom, Cale Horne, Kurt Braddock, Peter Vining and Nicole Zinni Pennsylvania State University Carnegie Mellon University Prepared for delivery at the 2010 Annual Meeting of the American Political Science Association, September 2-5, 2010. Copyright by the American Political Science Association. Key words: terrorism, Al-Muhajiroun, competitive adaptation, learning ABSTRACT There is widespread agreement among scholars and practitioners that the counterterrorism literature suffers from a lack of primary-source field research.2 The absence of solid ethnographic research has yielded studies that suffer from a lack of rigorous analysis and often result in opinion masquerading as analysis. The lack of field research is also due to a failure to integrate ethnographic research into modeling efforts, as well as a failure more broadly to appreciate the significance of ethnographically valid data in human, social, cultural, and behavioral studies in a systematic investigation of terrorist behavior. The project briefly outlined in this paper seeks to redress this deficiency by combining the strengths of ethnographic field research (collected by social scientists at Penn State) with the sophisticated modeling capabilities of computer scientists (at Carnegie Mellon University). Specifically, we are analyzing data from interview transcripts, news reports, and other open sources concerning the militant activist group Al-Muhajiroun and the terrorist groups Provisional Irish Republican Army (PIRA) and Revolutionary Armed Forces of Colombia (FARC). Using competitive adaptation as a comparative organizational framework, this project focuses on the process by which adversaries learn from each other in complex adaptive systems and tailor their activities to achieve their organizational goals in light of their opponents‟ actions. Ultimately, we will develop a mesolevel model of militant networks that combines insights from political science, organizational theory, psychology, network science, and computational modeling.
1
The work presented in this paper was supported (in part) by the Office of Naval Research under Grant #N0001409-1-0667. 2 Horgan, J. (in press). “Interviewing the Terrorists: Reflections on Fieldwork and Implications for Psychological Research.” Political Psychology. 1
INTRODUCTION A growing number of scholars and practitioners recognize the value of mixed-methods and interdisciplinary approaches to studying militant and terrorist3 organizations. Furthermore, there is growing recognition that the terrorism and counterterrorism literatures suffer from a lack of primary-source field research. Much of our knowledge and understanding of militant movements comes solely from news reports and other secondary sources, contributing to a systematic bias in available data. As a result of this bias, naïve and impractical policy recommendations are often generated from findings that do not necessarily reflect the reality of militant behaviors. This paper addresses these issues and provides an interdisciplinary framework from which to study the behavior of militant groups that may use or support terrorism. Building on the concept of competitive adaptation, our research team 4 investigates how militant groups learn and adapt when interacting with governments, civilians and other militant groups, as well as how they alter their subsequent behavior. Some guiding questions behind our work include:
How do militant (“red team”) and government (“blue team”) networks learn from one another and adapt (or fail to adapt) over time?
3
Whereas we recognize that “terrorist” and “terrorism” are often loaded terms, we use them in this paper primarily to distinguish violent militant groups (such as the PIRA or FARC) from non-violent militant activist groups (such as Al-Muhajiroun). Thus we refer to LaFree & Dugan‟s (2007) relatively broad definition of terrorism as „„the threatened or actual use of illegal force and violence to attain a political, economic, religious or social goal through fear, coercion or intimidation (p. 184).‟‟ Both the PIRA and FARC have carried out activities which clearly fit this definition, whereas Al-Muhajiroun has not, according to data available in the Global Terrorism Database (GTD). 4 Lead Principal Investigator – Dr. John Horgan, Penn State; Co-Lead PI – Dr. Michael Kenney, Penn State Harrisburg; Co-Investigator – Professor Kathleen Carley, Carnegie Mellon University; Co-Investigator – Dr. Mia Bloom, Penn State. 2
How do major events (terrorist attacks, counterterrorism (CT) operations, death of a group leader, government policy shifts) affect the structure of militant networks?
Does the structure of militant networks differ in models based on open-source information (news account, court transcripts, official statements) versus private information (interviews with de-radicalized or disengaged militants)?
This paper combines the analytical richness of ethnographic research with sophisticated computer modeling techniques to provide a meso-level model of militant networks that function in complex-adaptive systems. More specifically, we have used data from interview transcripts in addition to news reports, public statements and other open sources in order to study the competitive adaptation of three militant groups across four countries. This preliminary paper discusses the project‟s theoretical approach and initial focus on Al-Muhajiroun, a non-violent militant activist group, based in the United Kingdom and Lebanon. Several of Al-Muhajiroun‟s former members and associates, including Asif Hanif, Mohammed Junaid Babar, and Richard Reid, have been implicated in terrorist attacks and its leadership has been under investigation for militant activities in the UK and abroad.5 This distinction is important to make as subsequent studies of Al-Muhajiroun and interviews of its members are permitted under US law. First, this paper presents the theoretical framework of competitive adaptation and its usefulness in the investigation of our research questions. Second, the paper discusses the development of the project‟s data collection procedures, methods of analysis and how this 5
For an interview with the former leader of Al-Muhajiroun, see Horgan, J. (2009). Walking Away from Terrorism: Accounts of Disengagement from Radical and Extremist Movements. New York: Routledge.
3
„mixed-method‟ approach is expected to generate innovative and policy-relevant findings. Third, we briefly discuss some preliminary findings of interest for Al-Muhajiroun. Finally, we conclude with a discussion of challenges thus far, lessons learned and how the project is expected to evolve over the course of the next few years.
The Importance of Studying Militant Group Dynamics Scholars have found that pathological characterizations of those who participate in militant and/or terrorist groups are inconsistent with empirical realities. Works by Horgan (2008), Taylor & Horgan (2006), Sageman (2004) and Kimhi & Even (2004) demonstrate that militant activists and those who commit or support acts of terrorism tend not to be psychologically abnormal, disturbed or desperate individuals as they are often popularly characterized. To the contrary, these scholars have found that profiling militant activists and terrorists is extremely difficult and often unproductive. In fact, individuals who participate in militant organizations as well as those who go on to commit or support acts of terrorism often reflect a broad diversity of backgrounds. Existing ethnographic fieldwork on our first case study (the militant activist group Al-Muhajiroun in the United Kingdom) supports these findings. Wiktorowicz (2005: 91), for example, notes that Al-Muhajiroun activists tend to be collegeeducated and often have aspirations of upward mobility, thus debunking generalizations of a prototypically depraved militant activist background and corroborating findings by Krueger & Malekova (2003) regarding the relatively high levels of education among militant activists. A review of the political science literature on terrorist groups suggests that organizational approaches in particular can offer valuable insight when investigating the behaviors of militant 4
and terrorist groups. Organizational perspectives of terrorism are broadly outlined in Crenshaw‟s (1987) discussion of why instrumentalist approaches cannot account for many unsolved puzzles regarding terrorist behavior. For example, strategic approaches to the study of terrorism usually focus on how the use of terrorism can potentially coerce target states into making policy concessions (Crenshaw, 1981; Dershowitz, 2002; Lake 2002; Pape, 2003; Brym and Araj, 2006; Freedman, 2007). Whereas strategic approaches have indeed advanced our understanding of how terrorism might be logically employed within the context of an asymmetric conflict, it is clearly necessary to also dissect terrorist organizations themselves in order to understand the many instances in which violent militant groups‟ use of terrorism seems to contradict strategic models. Noting this problem, Crenshaw (1987) suggests that the use of terrorism itself may offer distinct internal benefits to the organizations that employ it as they pursue group survival and cohesiveness, in addition to the achievement of stated political goals. Examples of recent work building on these organizational perspectives of terrorism include Abrahms (2008; 2006) and Freedman (2007); both of these authors observe that the use of terrorism often appears to be, at best, unsuccessful and, at worst, counterproductive with respect to achieving the organization‟s stated strategic goals. These observations again suggest that there might be internal benefits and/or interim advantages conferred to a group by carrying out acts of terrorism. Thus the relationship between a militant group‟s activism, its use or support of terrorism, and the achievement of strategic organizational goals is more complex than often conceptualized by strategic approaches. Moreover, it becomes clear that explaining militant behavior requires an in-depth examination of
5
militant groups themselves, in addition to the broader strategic environment in which they operate. A major goal of this research is to learn about the specific mechanisms that drive militant organizational behavior as a result of both the internal organizational processes, in addition to the strategic interactions between militant organizations and state governments. Specific research questions we have posed to study these mechanisms include:
What are the social network properties of the relationships within the militant organization as well as between one militant organization and another?
What is the cultural significance of familial kinship and friendship networks in relation to all stages of organizational involvement (e.g. pre-socialization and cultural risk precursors, indoctrination, recruitment, initial involvement, sustained involvement, engagement in violent activity, disillusionment, disengagement, and de-radicalization)?
How have individual and organization profiles evolved from the militant movement‟s inception to today?
How does loyalty and group power work in the organization, and what role is played by ideology in this (both in terms of ideological content, and ideology as a process)?
What factors influence the nature of the relationship between the organization and the community it claims to represent, and the organization‟s use of violence to achieve its aims?
How do militant organizations successfully rejuvenate themselves in the wake of external successes against them?
6
Organizational Learning and Competitive Adaptation in Militant Groups Scholars have reasoned that militant organizations may not be fundamentally different in structure and membership from other organizations, such as governments, political parties or corporations (Weinberg, 1992). Indeed, militant groups often exhibit attributes of conventional organizations, such as merit-based hierarchal structures, public relations and media branches, or even the use of non-violent political activism and social engagement. The groups Hamas and Hezbollah, for example, are both known to engage in terrorism while also pursuing roles of governance and provisioning social welfare to supporters (Berman 2009). Wiktorowicz (2005) writes extensively about the religious services provided by Al-Muhajiroun to its members and potential activists, such as Islamic adjudication and Qur‟anic study groups. A large amount of work already exists describing organizational processes in conventional political entities (see Breslauer & Tetlock, 1991; Eden, 2004; Etheridge, 1985; Goldgeier, 1994; Haas, 1990; Haas, 1992; Hall, 1993; Heclo, 1974; Khong, 1992; Nye, 1987; Reiter, 1996; Sagan, 1993; or Weir & Skocpol, 1985 for just a few prominent examples from this expansive literature). Some terrorism scholars have adopted organizational perspectives to the study of militant groups, using a broad array of methodologies including ethnographic research, rational choice modeling and social network analysis (Abrahms 2008, 2006; Asal & Rethemeyer, 2008; Freedman, 2007; Helfstein, 2009). Whereas considering militant and terrorist organizations as comparable to more conventional organizations is fruitful, there likely remain key differences between conventional organizations (as they are often conceptualized in the organizational theory literature) and militant or terrorist groups, especially with respect to organizational learning. For example, while
7
we understand how learning functions through multiple iterations and trial and error in conventional organizations, learning within militant organizations likely differs due to of their oftentimes covert nature and illicit status within the states that they operate. There are few studies that specifically examine learning in nonconventional, covert, illegal or violent groups such as militant or terrorist organizations. Nevertheless, some recent work has offered useful insights into how militant and terrorist organizations learn and adapt within the adversarial environments that they operate. A team of RAND researchers, led by Jackson (2005), examined organizational learning in several violent militant groups, including the Provisional Irish Republican Army, Aum Shinrikyo, Jemaah Islamiyah, Hezbollah, and the radical environmentalist Animal Liberation Front and Earth Liberation Front. Jackson models organizational learning in militant groups as a four-part process involving the acquisition, interpretation, distribution and storage of information and knowledge. This knowledge may be explicitly expressed in knowledge artifacts and organizational routines, or tacitly expressed in the skills and experience of organization members. A separate study by Hamm (2007; 2005) draws on court documents contained in the American Terrorism Study database and the criminological literature on social learning to explore how some violent political extremists acquire the skills to perform their violent tradecraft. Leweling & Nissen (2007) provide conceptualizations of violent militant groups as industries that attempt to perform the “work” of terrorism while interacting and competing with similarly-conceptualized counterterrorism and governance industries. While these studies offer insights into how numerous militant groups train their members and develop certain technological innovations, they do not systematically examine the internal processes of group learning and interpretation, as experienced by militants 8
themselves. Moreover, these studies also do not take into account the broader competitive environments in which militant groups operate. Drawing on organizational and complexity theory (see, for example Fiol & Lyles, 1985; Dodgeson, 1993; Argyris & Schon, 1978; Argyris & Schon, 1996; Hedberg, 1981; Tetlock, 1991; Levitt & March, 1988; Epstein & Axtell, 1996; Axelrod, 1997; Axelrod & Cohen, 1999; Carley & Prietula, 1994; Carley, 2002; Holland, 1995; Jervis, 1997) in addition to fieldwork in Colombia, Kenney (2007) describes how organizational knowledge is leveraged by competing networks that interact in complex adaptive environments. Kenney dubs this process competitive adaptation, which explains how organizational learning occurs within an environment that is typically (though not always) characterized by hostility and multiple actors pursuing opposing goals. Often, organizational networks that exhibit learning and adaptation in competitive environments will vie for a common, finite resource (such as the support of a population or control of territory). Thus, competitive adaptation may be described as the tit-for-tat process of organizational learning that is associated with the pressures of a typically zero-sum contest between organizations, in addition to the pursuit of broader organizational goals that are often discussed by conventional organizational theorists. We believe that a network-based theoretical approach to the study of militant groups will allow us to effectively model both the internal organizational dynamics of militant groups, in addition to broader strategic interactions between militant groups and governments. Competitive adaptation is thus the framework from which we approach our study of militant groups, and we employ ethnographically-based network analysis as our primary tool when modeling this framework. This approach, along with findings in the
9
organization theory literature provides us with guidance in developing our research and expectations. Accounting for Overlapping Social Networks of Militants While existing social network analyses of militant and terrorist groups tend to focus on the roles of individuals for the purpose of comparing and contrasting different organizations (Leweling & Nissen, 2007; Sageman, 2008; Asal & Rethmeyer, 2008), we are interested in studying both intraorganizational and interorganizational associations that may be betteridentified by expanding social network analysis beyond a single-role social context. While assigning human agents to a single role (such as job title or primary task within the organization) can simplify a network analysis for inter-organizational comparisons, it is well-known that within organizations, humans can belong to many different groups simultaneously and at varying levels of association. Employees may be assigned to or form multiple formal or ad-hoc working groups. Individuals with proximate offices or similar backgrounds may form cliques. Movement of personnel within and between organizations lead to networks of association that are far more complicated than those modeled with single role humans. Davis & Carley (2008) develop an algorithm to explore these “fuzzy overlapping groups” from network data and find that emerging group dynamics may be better-accounted for by quantifying these groups using ethnographic data. We are extending this finding to militant and terrorist organizations in our research design when addressing our first research question regarding the nature of relationships within and between militant social networks. We expect that: H1: Participants in militant movements are organized into overlapping social and organizational networks 10
H2: Social networks are a primary mechanism by which participants in militant groups and organizational
networks
share
information
about
their
activities
and
government
counterterrorism efforts
Comparing Metrics of Social and Organizational Networks of Militant Organizations In addition to studying intraorganizational dynamics emerging from “fuzzy overlapping groups,” we are also interested in studying how traditional social metrics differ across militant groups. Analyses of the militant group Al-Muhajiroun by Wiktorowicz (2005) and the terrorist groups FARC and Provisional IRA, touched upon by Kenney (2007) and described by Horgan & Taylor (1997) (respectively), provide qualitative descriptions of these groups‟ structures and how those structures may account for group behavior. We seek to expand on these findings using ethnographic data in order to compare and contrast these groups from both a qualitative perspective and from using quantitative metrics of social network analysis. Measures of closeness, for example, allow for quantitative comparison of the distances between actors within networks, whereas measures of centrality enable us to compare the density of militant social networks (Hanneman & Riddle, 2005). Other metrics permit us to compare additional network attributes such as hierarchies and emerging leadership, thus enabling us to study organizational attributes and the evolution of militant groups over time based on the analysis of primary and secondary ethnographic data. We expect that: H3: Social and organizational network properties, such as centralization, hierarchy, and density of network ties, vary across different militant movements
11
Specifically, we expect that social network metrics calculated from ethnographic data, such as those briefly described above, will allow us to arrive at substantive conclusions when comparing the structures, decision-making, learning and adaptation of militant and terrorist groups. In comparing Colombian drug cartels and the terrorist organization al Qaeda with state governments, Kenney (2007) finds that structural attributes of militant and terrorist organizations (such as levels of management, centralization of decision-making and institutional rules and regulations) influence the ability of those organizations to learn and evolve as they compete with governments. Moreover, these structural attributes are associated with such phenomena as organizational splits and proliferation. When considering relationships between structures of management and organizational learning, it has been suggested that bureaucracy may be both beneficial and detrimental. Autonomous decision-making structures and limited management are often associated with diffuse “cell” networks of organizations, a relevant example being al Qaeda (as discussed by McCallister, 2004). Kollman, Miller and Page (2000) argue that autonomous decision-making and management may benefit organizations as they seek to accomplish simple tasks by facilitating innovation and adaptation. For example, a particular cell within a diffuse organization that lacks rigid standards or protocols to follow might develop a more efficient method of bomb assembly that cuts down on materials used. This method might then be shared throughout the organization, thus facilitating an improvement in efficiency. In contrast to autonomous organizations, larger bureaucracies are more likely to fall into “competency traps” with respect to task performance, whereas organizational inertia associated with a large organization, such as a government, may result in the continuation of routines and strategies that are suboptimal (Levitt & March, 1988). Levitt and Marsh discuss the “QWERTY” 12
arrangement of keys on a keyboard, originally designed to avoid typewriter jams, as an example of such a competency trap; the keyboard‟s keys could be rearranged to facilitate faster typing, however doing so would require employees to re-learn how to type. The costs of adopting such a change are expected increase as a function of organization size and internal centralization; thus a smaller firm with fewer internal regulations might be more likely to adopt a superior routine bit by bit, whereas such a change would be a major undertaking for a larger firm. Similar dilemmas likely affect both governments and militant organizations as they refine and improve their routines, designs and tactics in pursuit of their strategic goals. Kenney (2007) maintains that while the Colombian government enjoys more resources than the militant groups and drug cartels that it has fought for decades, its size inhibits its ability to change in tandem with those groups, thus inhibiting its ability to eradicate them. This finding can be corroborated with findings by Jackson (2005) from the Provisional Irish Republican Army, and is consistent with broader theories of asymmetric conflict proposing that a weak actor can potentially defeat a strong actor by emphasizing speed, flexibility and intelligence-gathering when carrying out attacks, as well as by prompting a stronger enemy to expend resources inefficiently during the course of the conflict (Arreguin-Toft, 2001; Slantchev, 2003; Bennett, 2008; Butler & Gates, 2009). From this discussion, we expect that: H4: Networks with fewer management levels and decentralized decision-making will process information and make decisions more quickly than those with greater management levels and centralized decision-making.
13
H5: Militant networks, with their flat decision making hierarchies and informal rules of engagement, adapt quickly in response to external pressures. While many levels of management and centralization may inhibit innovation within an organization as it carries out simple tasks, the organizational reach, resources and connectivity associated with larger bureaucracies may be better-suited to organizations that pursue complex tasks, such as developing chemical weapons (Kollman, Miller and Page, 2000). As an example, Jackson‟s (2005) study of the Japanese terrorist group Aum Shinrikyo notes that the group‟s centralized bureaucracy and decision-making structures facilitated its coordinated and wellfunded chemical weapons research program. This effort eventually led to the group‟s acquisition and use of sarin gas against Japanese subway commuters in 1995 with tragic consequences. In contrast, scholars have suggested that al Qaeda‟s evolution into a diffuse, cell-based network following the September 11, 2001 World Trade Center and Pentagon attacks might partially explain why the organization has not yet successfully employed chemical, biological, radiological or nuclear (CBRN) weapons against Western targets, despite expressed interest in doing so (McCallister, 2004). We expect that: H6: Networks with decentralized and dispersed authority structures will be more susceptible to fragmented, localized learning than organizations with centralized, tightly coordinated administrative structures. Examining Networks of Locations, Events, Knowledge, Resources and Tasks When studying organizational learning and competitive adaptation in militant and terrorist networks, we seek to examine militant networks in their entirety, beyond the social context. Specifically, we recognize that in order to understand group dynamics, learning, 14
evolution, decision-making and emergent behavior, it is necessary not only to examine the roles and relationships of individual agents and groups within organizations, but also how those agents relate to locations in space, as well as the knowledge and resources leveraged by agents within organizations in order to fulfill group tasks. Carley (1999) writes that organizations can be described as an “ecology of networks” that continually evolve as agents within the organization learn, move and interact. A network of social roles within an organization might appear very different from a network of knowledge and expertise, which in turn might be very different from the network of resources or geographic proximity. Kenney‟s (2007) work on competitive adaptation similarly emphasizes the importance of organizational properties beyond those associated with individual human agents, arguing that the flow of knowledge, routines and artifacts within organizations is as important as the flow of personnel. We conceptualize militant networks consistently with these arguments and expect that:
H7: Militant networks “learn” when their participants receive information about their activities, process this information through knowledge-based artifacts, and apply the information to their practices and activities
Methodology, Case Selection and Data Collection A major limitation to the study of competitive adaptation in militant groups is the lack of available primary data. Jackson (2005: 200) describes this problem in his conclusions, noting that “[t]he lack of data on group decision making and the factors that drive groups‟ interpretation
15
of their own successes and learning opportunities is a major handicap to complete understanding of learning in these organizations. Additional study and exploration of specific elements of organizational learning in these and other terrorist groups is therefore warranted.” Our own research is opening this “black box” of militant groups by capturing them in their own words through leveraging available primary and secondary data sources. Specifically, we combine ethnographic data collected from existing interviews with data derived from public statements, media interviews, newspaper articles and other secondary sources in order to provide an extensive database of information about militant behavior, learning and adaptation. We then use state of the art computational modeling techniques to analyze these data and thus augment our understanding of how participants in militant networks acquire knowledge and adapt their operations in response to law enforcement agencies with which they share complex adaptive systems. This interdisciplinary and mixed-methods approach thus combines the subject matter expertise of social scientists and social science theory with the analytical power of computational modeling and network analysis.
Developing an Interdisciplinary Framework to Study Competitive Adaptation
A first step in our research was the development of an interdisciplinary framework in order to arrive at mutual understandings of the important terms and concepts associated with competitive adaptation. This framework was developed through a series of meetings between our ethnographic research team, led by Drs. John Horgan, Mia Bloom, and Michael Kenney at Penn State University and Penn State Harrisburg (respectively) and our computational modeling team, 16
led by Dr. Kathleen Carley at Carnegie Mellon University. The development of our framework was a major undertaking and is described in more detail by Braddock et al. (2010, forthcoming). Gruber (1993), Musen (1992), and Uschold and Gruninger (1996) assert that shared understanding of commonly-used terms and concepts is paramount for interdisciplinary research. A primary motivation for developing a shared conceptual framework was the reduction of conceptual and terminological confusion that emerges as a natural byproduct of interdisciplinary research. In addition to the communicative benefits associated with a shared framework, Noy and McGuinness (2001) claim that making assumptions about a particular domain explicit affords researchers the ability to alter their assumptions should their knowledge about that particular subject domain change in any way.
By explicitly stating assumptions regarding
particular facets of competitive adaptation, the research team is given the flexibility to change those assumptions in accordance to a dynamic academic environment that is informed by multiple disciplines.
Finally, any conceptual framework may be generalizable beyond its
original intent. In the event that an investigation related to the one described here be undertaken, an explicitly developed framework could be reused to advocate efficiency and conceptual consistency.
In this sense, a shared framework like the one described here is central to
cumulative research and knowledge building (Kuhn, 1996). First Case Study: AL-Muhajiroun in the United Kingdom Following the development of our interdisciplinary framework, the research team began to collect primary and secondary data related to the first militant groups we chose to study: AlMuhajiroun (henceforth referred to as AM) in the United Kingdom. While Al-Muhajiroun is not
17
a terrorist group per se, recent work by Pantucci (2010: 226) claims that AM and another UK militant group, „Supporters of Shariah‟, “have been the connective thread through most Islamist terrorist plots that have emanated from the United Kingdom.” Founded in 1996 by Omar Bakri Mohammed and officially disbanded in 2004, AM members continue to operate in the United Kingdom under several splinter organizations that we seek to include in our study. AM was appropriate as a first choice for developing our data collection procedures because information about the group is relatively available and accessible and members of our own research team are able to leverage available transcripts of interviews with AM members that were relevant to the analysis. Work by Wiktorowicz (2005) indicates that AM not only exhibits adaptive behavior as it interacts with UK authorities, but that the relatively liberal political and social environments of the United Kingdom often condition these interactions, providing both advantages and disadvantages to each side. As an example, Wiktorowicz shows how freedom of the press in the United Kingdom has been a double-edged sword for Al-Muhajiroun, both assisting the group in publicizing its ideologies to potential recruits, but also resulting in a broader societal ostracization of the group in such a way that has had direct negative ramifications on its operations. These negative ramifications have included the groups‟ banning from the use of public venues and increased police scrutiny of group activities, resulting in arrests and the costly loss of charitable organization status in the United Kingdom (p. 126). In response to the negative ramifications of publicity, AM has adapted a strategy of organizational proliferation, diversification and obfuscation in order to continue spreading the group‟s ideology and connect with potential recruits without suffering the costs that are now 18
associated with the AM label, and without risking organizational death in the event of a police crackdown (p. 124). Indeed, later work by Pantucci (2010) has shown that despite exactly such a crackdown in 2004 which resulted in the organization‟s abolishment, AM members continue to operate in the United Kingdom and abroad under a variety of alternative platforms, fronts and splinter groups. Recent work by Kenney (2009) discusses this tit-for-tat dynamic between the British government and Al-Muhajiroun‟s in depth. Following the disbanding of AM in 2004, the group‟s leadership established two new groups called Al Ghurabaa (The Strangers) and the Saved Sect, both of which attracted many of the same members as Al-Muhajiroun. Furthermore, when these splinter groups faced the threat of disbandment, former AM leadership created the „Ahlus Sunnah wal Jamaah‟, an invitation-only Internet discussion forum (p. 124). Whereas AM‟s organizational expansion is a clear example of adaptive behavior within the competitive environment in which it operates, it is also equally interesting that militant organizations such as AM often fail to adapt, or learn the “wrong lessons,” despite their experiences. Kenney (2010, forthcoming) explains that militant groups might fail to adapt within their environments not only due to simple mistakes and human error, but potentially due to the underlying structures, ideologies or rules guiding an organization‟s behavior. The religious underpinnings of Al-Muhajiroun, that clearly condition the incentives structure of individual AM members and leaders (Wiktorowicz, 2005: 127). Kenney (2010, forthcoming) notes that AM‟s religiosity has led its leadership to occasionally leave important tactical decisions, such as behavior that may result in imprisonment, to “God‟s predetermined fate for them.” Thus from these examples of both adaptive and non-adaptive behaviors, AM is an interesting and relevant
19
case study of how a militant group evolves and adapts (or has fails to evolve and adapt) in a Western democracy. Collection of Primary and Secondary Source Data for Network Analysis When collecting data for our analysis of AM, we have diversified among a broad range of primary and secondary sources. Thus far, these sources have included:
Publically
available
media
interview
transcripts,
court
transcripts,
autobiographical sources and documentary accounts of AM members. These sources are being analyzed for content and direct quotes for the purpose of allowing comparative analysis between “private” and “public” statements of group members. As we will discuss, preliminary findings suggest that there are indeed stark differences between the two.
Press releases, website op-eds, magazine articles and other materials written and published by AM members. These materials provide us with supplementary evidence of AM‟s interactions in British society, as well as their organizational goals and approaches.
Newspaper articles, blog postings and other media accounts of AM. We are currently analyzing a database of over 1,000 newspaper articles published about the group. In addition to providing more direct quotes from AM members, these articles also provide us with material for timeline construction, as well as event data generation and analysis, which is a recent addition to the project.
Peer-reviewed articles, books and other major publications about AM. While existing ethnographic fieldwork on AM is limited to only a few studies, these studies 20
already provide evidence of adaptive behavior. We will thus build on findings by Wiktorowicz (2005), Pantucci (2010) and other scholars that are relevant to our research. From these primary and secondary sources, our research team constructed thesauri of known AM members (agents) and aliases with their associated roles in the organization. In addition, we have created thesauri of AM front groups and splinter organizations (of which there were many), timelines of AM activities and events. In addition “blue team” equivalents of these were compiled, documenting the activities of the UK government related to AM. Researchers extracted lists of relevant locations which will be used for the dynamic network analysis as well as public and private quotes of AM members. These data are being collected by a trained and supervised team of undergraduate interns in consultation with the computational modeling team at Carnegie Mellon University to ensure that the requirements of both social and computer scientists were met for the overall analysis. These data are coded using the program NVivo for qualitative analysis. These data were also provided to the computational modeling team for extraction of key semantic and network properties, using the text analysis program AutoMap, and for semantic and network analyses using the dynamic network analysis (DNA) software Organizational Risk Analyzer (ORA) (Carley et al., 2007). We also adapted the agent and organization thesauri that we provided to CMU for semantic and network analysis for the purpose of event data generation and analysis. The use of automated programs, such as the open-source Textual Analysis by Augmented Replacement Instructions (TABARI) in conjunction with high-quality actor and verb dictionaries that are
21
custom-designed for a conflict can provide accurate and detailed event data for statistical analysis (Schrodt, 2009). Indeed, the use of our actor and organizations thesauri in conjunction with a pre-existing coding scheme such as Conflict and Mediation Event Observations (CAMEO) is expected to generate a high-quality dataset of interactions between AM and UK authorities, possibly even at daily levels of aggregation (Schrodt et al., 2008). Moreover, the statistical aggregation and analysis of these event data are expected to provide interesting insights into the nature and evolution of the overall adversarial relationship between AM and UK authorities. In addition to these substantive analyses, we are working with our partners at CMU to develop practical procedures for the documentation and storage of primary source data in digital format, as well as determining what additional data should be collected in upcoming research. These procedures will be followed as we complete our analysis of AM and move on to our other case studies, which include the Irish Republican Army (IRA) in Northern Ireland and the Revolutionary Armed Forces of Colombia (FARC) in Colombia. Thus our current work on AlMuhajiroun is both providing us with interesting substantive insights into the conflict between AM and the UK, in addition to paving the way for similar analyses of larger conflicts for which far more source materials are available.
Preliminary Findings for Al-Muhajiroun As of August 2010 we are still completing the construction and verification of our thesauri from news articles and court transcripts. Collection of additional primary source material is expected to begin in fall, 2010 and will continue through the following year. Whereas 22
the project is still evolving, we have already generated some interesting findings from our first case study, Al-Muhajiroun. First among our findings is that differences do indeed exist between the content of public and private statements made by organization leaders. Private statements by AM leaders, for example, suggest an importance associated with the conflict in Kashmir and Muslim issues in Uzbekistan that is not apparent in public statements (Horgan & Horne, forthcoming). Findings such as these reinforce our notion that collecting additional primary source material from former members of groups like AM will produce valuable insights. In addition, our preliminary semantic and network analyses offer insight about the organizational roles of group members and how those individual roles interact within the overall organizational structure. This includes the identification of network nodes according to specialized knowledge, skill sets and leadership, as well as connections to other organizations. We are finding that the comparison of public and private statements together can help clarify these relationships. In the case of Al-Muhajiroun, our preliminary analyses indicate the importance of one particular individual as an emerging leader in the network, despite his relative lack of coverage in newspaper articles (Horgan & Horne, 2010) Furthermore, the results of our analysis are highly predicated on the quality of thesauri used and source matter. By constructing our thesauri in such a way that excludes “general terms” (examples being terms like “children”) this increases our ability to focus specifically on the authority structures of the network. Regarding source matter, we are (perhaps not surprisingly) finding that the origin of reporters has a clear effect on the directions of interviews. For example,
23
comparing US and UK news articles shows a clear reporter focus on US and UK issues, respectively (Horgan & Horne, 2010). Thus as we continue to build our dataset and begin further substantive analysis, we are learning important methodological “lessons” which will help us increase the accuracy and yield of our research.
Challenges, Lessons Learned and Next Steps Several challenges have presented themselves over the course of the project that we are currently addressing. One of these challenges is associated with the interdisciplinary, mixedmethods approach that we have chosen to adopt. Our work has required reviewing literature in multiple fields and has required researchers to explore concepts, ideas, literature and knowledge in fields outside of their specialties in order to establish baselines of common understanding. To address these challenges, the research team has implemented several measures to ensure that the project retains this interdisciplinary character throughout its lifespan. The construction of our framework was a major first step in achieving this goal. Moreover, our ethnographic researchers at Penn State have received training in the use of network analysis software that our computational modeling partners at Carnegie Mellon University will be drawing on to analyze our data. As previously discussed, we have consulted our partnering institutions in data collection and formatting procedures and will be cooperating to develop a common coding scheme which will be used for our interview transcripts. Among the technological tools that we have found to be helpful in collaboration thus far include a collaborative wiki and the use of shared electronic documents and spreadsheets that permit multiple users to edit and comment on project materials in real time. 24
Another major challenge to our research is associated with our gathering of data. Collecting rich, primary source information about militant organizations requires a careful approach that balances researcher‟s questions with the realities of a conflict. Interviewing militants, for example, can be done safely, ethically and does produce valid, policy-relevant data. However, during such interviews, several important concerns affect the sorts of questions that should be asked and the extent to which detail can realistically be pursued. Many of these concerns originate in the ongoing conflicts between militant groups and states, which will likely cause an active militant to limit the type and extent of information they are willing to provide. Moreover, it is typical for active members of political organizations to respond to interviewer questions with ideology and propaganda. While such responses are inherently valuable to researchers, they might potentially prevent our own researchers from gathering the processrelated information they are seeking. Nevertheless, the questions that are asked during an interview and the way in which they are asked will often determine the type of answer that is provided. Horgan (2008) finds, for example, that asking militants “how” they became involved in terrorism often results in very different responses than those provided when asking “why” they became involved. We are addressing the difficulties associated with gathering primary source data from available former militants by carefully developing a rubric of questions to be covered during planned interviews, while leaving the interviews themselves semi-structured in order to maximize the relevant information provided. Moreover, the primary investigator has provided all members of the research team with additional training in interview techniques meant to address these issues.
25
As our computational modeling team continues to analyze the data we have provided to them on Al-Muhajiroun, our researchers aim to collect new primary source data from former members of AM as well as from other groups. Using our theoretical framework of competitive adaptation, our research team will analyze these data with the objective of understanding how many of the internal processes affecting militant organizations affect their behavior, their use of terrorism (or lack thereof) and how they learn and adapt within their complex-adaptive environments. The resulting model of competitive adaptation will contribute an evidence-base to inform decision-making and law enforcement training, in addition to evaluating the impact of specific policy interventions and policy forecasting.
REFERENCES Abrahms, Max, 2006. “Why Terrorism Does Not Work.” International Security 31(2): 42-78. Abrahms, Max, 2008. “What Terrorists Really Want: Terrorist Motives and Counterterrorism Strategy.” International Security 32(4): 78-105. Arreguín-Toft, Ivan, 2001. “How the weak win wars: A theory of asymmetric conflict.” International Security 26(1): 93-128. Asal, Victor and R. Karl Rethemeyer, 2008. “The Nature of the Beast: Organizational Structures and the Lethality of Terrorist Attacks.” Journal of Politics 70(2): 437-449. Axelrod, Robert, 1997. The Complexity of Cooperation: Agent-Based Models of Competition and Cooperation. Princeton: Princeton University Press. Axelrod, Robert and Michael D. Cohen, 1999. Harnessing Complexity: Organizational Implications for a Scientific Frontier. New York: Free Press. Bennett, Scott, 2008. “Governments, Civilians, and the Evolution of Insurgency: Modeling the Early Dynamics of Insurgencies.” Journal of Artificial Societies and Social Simulation 11(4): 7 http://jasss.soc.surrey.ac.uk/11/4/7.html. Berman, Eli, 2009. Radical, Religious and Violent: The New Economics of Terrorism (Cambridge, Mass.: The MIT Press). Braddock, Kurt H., John Horgan, Michael Kenney, and Kathleen Carley, 2010. “Developing a Multidisciplinary Framework: A Case Illustration from ICST‟s Research on Competitive Adaptation in Terrorist Networks.” Annual Meeting of the International Society of Political Psychology, San Francisco, California, July 7-10, 2010. Breslauer, George and Philip Tetlock (eds), 1991. Learning in U.S. and Soviet Foreign Policy. Boulder, CO: Westview Press. 26
Brym, Robert J. and Bader Araj, 2006. Suicide Bombing as Strategy and Interaction: The Case of the Second Intifada. Social Forces 84(4): 1969-1986. Butler, Christopher and Scott Gates, 2009. “Asymmetry, Parity, and (Civil) War: Can International Theories of Power Help us Understand Civil War?” International Interactions 35(3): 330-340. Carley, Kathleen, Jana Diesner, Jeffrey Reminga and Maksim Tsvetovat, 2007. “Toward an Interoperable Dynamic Network Analysis Toolkit.” Decision Support Systems 43: 13241337. Carley, Kathleen and Michael Prietula, eds, 1994. Computational Organization Theory. Hillsdale: Lawrence Erlbaum. Carley, Kathleen, 2002. “Intra-organizational Computation and Complexity,” in The Blackwell Companion to Organizations, edited by Joel A.C. Baum. Malden, Mass.: Blackwell: 208232. Crenshaw, Martha, 1981. “The Causes of Terrorism.” Comparative Politics 13(4): 379-399. Crenshaw, Martha, 1987. “Theories of Terrorism: Instrumental and Organizational Approaches” in Inside Terrorist Organizations ed. David C. Rappoport, 2001. London: Cass. Cyert, Richard and James March, 1963. A Behavioral Theory of the Firm. Englewood Cliffs, New Jersey: Prentice-Hall. Davis, G.B. and K.M Carley, 2008. “Clearing the FOG: Fuzzy, Overlapping Groups for Social Networks.” Social Networks 30(3): 201-212. Dershowitz, Alan, 2002. Why Terrorism Works: Understanding the Threat, Responding to Challenge. New Haven: Yale University Press. Eden, Lynn, 2004. Whole World on Fire: Organizations, Knowledge, & Nuclear Weapons Devastation. Ithaca, NY: Cornell University Press. Epstein, Joshua M. and Robert Axtell, 1996. Growing Artificial Societies: Social Science from the Bottom Up. Washington, D. C.: Brookings Institution Press. Etheridge, Lloyd, 1985. Can Governments Learn? New York: Pergamon. Freedman, Lawrence, 2007. “Terrorism as a Strategy.” Government and Opposition 42(3): 314339. Goldgeier, James, 1994. Leadership Style and Soviet Foreign Policy: Stalin, Krushchev, Brezhnev, Gorachev. Baltimore, MD: John Hopkins University Press. Gruber, T. R. (1993). A translation approach to portable ontology specification. Knowledge Acquisition, 5, 199-220. Haas Ernst, 1990. When Knowledge is Power: Three Models of Change in International Organizations Berkeley: University of California Press. Haas, Peter (ed), 1992. “Knowledge, Power, and International Policy Coordination,” special issue of International Organization 46 (1): 1-35. Hall, Peter, 1993. “Policy Paradigms, Social Learning, and the State.” Comparative Politics 25: 275-296. Hamm, Mark, 2005. “Crimes Committed by Terrorist Groups: Theory, Research and Prevention” U.S. Department of Justice.
27
Hamm, Mark, 2007. Terrorism as Crime: From Oklahoma City to Al-Qaeda and Beyond. New York: New York University Press. Hanneman, Robert A. and Mark Riddle, 2005. Introduction to Social Network Methods. Riverside, CA: University of California, Riverside (published in digital form at http://faculty.ucr.edu/~hanneman/). Helfstein, Scott, 2009. “Governance of Terror: New Institutionalism and the Evolution of Terrorist Organizations.” Public Administration Review 69(4): 727-749. Heclo, Hugh, 1974. Modern Social Politics in Britain and Sweden: From Relief to Income Maintenance. New Haven, Yale University Press. Holland, John, 1995. Hidden Order. Cambridge: Perseus Books. Horgan, John and Max Taylor, 1997. “The Provisional Irish Republican Army: Command and Functional Structure.” Terrorism and Political Violence 9(3): 1-32. Horgan, John, 2008. “From Profiles to Pathways and from Roots to Routes: Perspectives from Psychology on Radicalization into Terrorism.” ANNALS, American Association of Political and Social Science 618: 80-94. Horgan, John, (in press). “Interviewing the Terrorists: Reflections on Fieldwork and Implications for Psychological Research.” Political Psychology. Horgan, John, 2009. Walking Away from Terrorism: Accounts of Disengagement from Radical and Extremist Movements. New York: Routledge. Horgan, John and Cale Horne, 2010. "Competitive Adaptation in Terrorist Networks: A Multidisciplinary Basic Research Project for Modeling the Social, Psychological and Cultural Properties of Adaptive Militant Networks, Year One Review." International Center for the Study of Terrorism. Presentation to the Office of Naval Research, May 15, 2010. Jackson, Brian A. et. al., 2005. Aptitude for Destruction Volume 1: Organizational Learning in Terrorist Groups and its Implication for Combating Terrorism. Santa Monica, CA: RAND. Jackson, Brian A. et. al., 2005. Aptitude for Destruction Volume 2: Case Studies of Organizational Learning in Five Terrorist Groups. Santa Monica, CA: RAND. Jervis, Robert, 1997. System Effects: Complexity in Political and Social Life. Princeton: Princeton University Press. Kenney, Michael, 2007. From Pablo to Osama: Trafficking and Terrorist Networks, Government Bureaucracies, and Competitive Adaptation. University Park, PA: Penn State Press. Kenney, Michael, 2010. “Dumb” yet Deadly: Local Knowledge and Poor Tradecraft among Islamist Militants in Britain and Spain.” Studies in Conflict and Terrorism 33, no. 10 (Forthcoming, October 2010): 1-22. Kenney, Michael, 2009. Organizational Learning and Islamist Militancy. National Institute of Justice, U.S. Department of Justice (Washington, D.C., May, 2009). . LaFree, Gary and Laura Dugan, 2007. “Introducing the Global Terrorism Database.” Terrorism and Political Violence 19:2, 181-204. Khong, Yuen Foong, 1992. Analogies at War. Princeton: Princeton University Press. Kimhi, Shaul and Shemaul Even, 2004. "Who Are the Palestinian Suicide Bombers?" Terrorism and Political Violence 16(4): 815-840. 28
Kollman, K, JH Miller and SE Page, 2000. “Decentralization and the Search for Policy Solutions.” Journal of Law, Economics and Organization 16(1): 102-128. Krueger, Alan and Jitka Malekova, “Education, Poverty and Terrorism: Is there a Causal Connection?” Journal of Economic Perspectives, Volume 17, Number 4, Fall 2003, pp.119–144 Kuhn, T. S. (1996). The structure of scientific revolutions. Chicago: University of Chicago Press. Lake, David A., 2002. “Rational Extremism: Understanding Terrorism in the Twenty First Century.” Dialogue, International Organization, Spring: 15-29. Levitt, B. and J. March, 1988. "Organizational Learning," Annual Review of Sociology 14(3): 319-340. Leweling, Tara and Mark E. Nissen, 2007. “Defining and exploring the terrorism field: Toward an intertheoretic, agent-based approach.” Technological Forecasting and Social Change 74: 165-192. McCallister, Brad, 2004. “Al Qaeda and the Innovative Firm: Demythologizing the Network.” Studies in Conflict and Terrorism 27: 297-319. Musen, M. A. (1992). Dimensions of knowledge sharing and reuse. Computers and Biomedical Research, 25, 435-467. Noy, N. F., & McGuinness, D. L. (2001). Ontology development 101: Creating your first ontology. Stanford Knowledge Systems Laboratory Technical Report KSL-01-05. Retrieved June 27, 2010, from http://www-ksl.stanford.edu/people/dlm/papers/ontologytutorial-noy-mcguinness.pdf Nye, Joseph, 1987. “Nuclear Learning and U.S.-Soviet Security Regimes.” International Organization 41 (3): 371-402. Pantucci, Raffaello, 2010. “The Tottenham Ayatollah and the Hook-Handed Cleric: An Examination of All their Jihadi Children.” Studies in Conflict & Terrorism 33: 226-235. Pape, Robert A., 2003. “The Strategic Logic of Suicide Terrorism.” American Political Science Review 93(3): 1-19. Reiter, Dan, 1996. Crucible of Beliefs: Learning, Alliances, and World Wars. Ithaca, N.Y.: Cornell University Press. Sagan, Scott, 1993. The Limits of Safety: Organizations, Accidents, and Nuclear Weapons. Princeton: Princeton University Press. Sageman, Marc, 2008. Understanding Terror Networks. Philadelphia: University of Pennsylvania Press. Schrodt, Phillip A., 2009. TABARI Version 0.7. Free software available at . Schrodt, Phillip, Omur Yilmaz, Deborah J. Gerner and Dennis Hermrick, 2008. “Coding Substate Actors Using the CAMEO (Conflict and Mediation Event Observations) Actor Coding Framework.” Presented at the 2008 Annual Meeting of the International Studies Association, 26 - 29 March. Slantchev, Branislav, 2003. “The Power to Hurt: Costly Conflict with Completely Informed States.” American Political Science Review 97(1): 123-133.
29
Taylor, Max and John Horgan, 2006. “A Conceptual Framework for Addressing Psychological Process in the Development of the Terrorist.” Terrorism and Political Violence 18: 1-17. Uschold, M., & Gruninger, M. (1996). Ontologies: Principles, methods, and applications. Knowledge Engineering Review, 11, 93-136. Weinberg, Leonard, 1992. Political Parties and Terrorist Groups. London: Frank Cass. Weir, Margaret and Theda Skocpol, 1985. “State Structures and the Possibilities for „Keynesian‟ Responses to the Great Depression in Sweden, Britain and the United States,” in Bringing the State Back In, edited by Evans et al. Cambridge: Cambridge University Press. Wiktorowicz, Quintan, 2005. Radical Islam Rising: Muslim Extremism in the West. Lanham, Maryland: Rowman & Littlefield.
30
