The Limits of Epistemic Communities : EU Security Agencies Mai ' a

This article examines the cases of the European Defence Agency (EDA) and EU Intelligence Analysis Centre (IntCen) to argue that although they are comprised of high-level security experts, they do not constitute epistemic communities. Research on other groups of security experts based in Brussels has shown that epistemic communities of diplomats, military experts, security researchers, and civilian crisis management experts, among others, have been able to influence the trajectory of security integration by virtue of their shared knowledge. Importantly, these security epistemic communities have been shown to significantly impact outcomes of EU security policy beyond what would be expected by looking only at member-states’ initial preferences. In exploring two examples of “non-cases” that are at the same time very similar to the other examples, the author seeks to shed light on why some expert groups do not form epistemic communities, and how this changes the nature of their influence. In so doing, the goal is to sharpen the parameters of what constitutes epistemic communities, and to add to our understanding of why they emerge. The argument advanced in this article is that institutional context and the nature of the profession matter as preconditions for epistemic community emergence.


Introduction
Over the past two decades, scholarly research has revealed the existence of numerous highly influential epistemic communities-basically defined as knowledgebased networks-that have swayed the trajectory of international cooperation by virtue of their shared professional expertise (Adler, 1992;Cross 2013a;Drake & Nicolaïdis, 1992;Gough & Shackley, 2001;Haas, 1989Haas, , 1992b;;Ikenberry, 1992;Kapstein, 1992;Peterson, 1992;Sandal, 2011;Youde, 2005).In particular, this research has shown that the European region-with its ongoing processes of integration, shared democratic values, supranational institutions, and transnational interactions-is highly conducive to the formation of ep-istemic communities (Cross, 2011;Howorth, 2004;Peterson & Bomberg, 1999;Radaelli, 1999b;Verdun, 1999;Zito, 2001).In previous work, I have examined nine separate case studies of epistemic communities in the European context, both historical and contemporary (Cross, 2007(Cross, , 2011)).This has enabled me to compare the ways in which epistemic communities work, determining whether they are nascent, emerging, weak, or strong.This previous research also demonstrates how epistemic communities can be located both inside and outside of formal institutions, and can be comprised both of scientists and of other kinds of experts.
My aim here is to compliment this existing research-without repeating the analysis in the introduc-tion to this special issue-through considering "noncases", that is, examples that bring to light the limits of the epistemic community concept.1 Thus, this article is inherently restricted in scope, and seeks to shed light on a specific dimension of the concept, rather than providing a comprehensive survey of the literature and theoretical debates that have come before.Under what conditions might we expect expert groups not to constitute epistemic communities?Does this have bearing on the expert group's future potential as an epistemic community?By necessity, these questions are actually prior to most research on epistemic communities.Rather than focusing on identifying epistemic communities, and investigating the nature of their influence, I explore some of the conditions that are conducive for an epistemic community to emerge in the first place, and the conditions that make their existence less likely.At a certain point on the strong-weak spectrum of epistemic community influence, the idea that an expert group could constitute an epistemic community drops away entirely.
This article examines the cases of the European Defence Agency (EDA) and EU Intelligence Analysis Centre (IntCen) to argue that although they are comprised of high-level security experts, they do not constitute (or contain) epistemic communities.I argue that the experts that populate both the EDA and IntCen lack the qualities that make them more than the sum of their parts, and as a result, they do not thus possess the political will to exercise collective agency beyond their formal mandate.In considering why this is the case, I seek to sharpen the parameters of what constitutes epistemic communities, and to add to our understanding of why they emerge.
Specifically, I hypothesize that if a group of experts work together within a highly formal institutional setting that is characterized by hierarchy and a limited mandate, epistemic community emergence is less likely.Second, I hypothesize that if the expertise of a group stems from professional backgrounds that value secrecy-such as in the fields of intelligence, certain corporate sectors, computer technology, journalism, and so on-epistemic community emergence is also more likely restricted because these professions are less open to informal interaction, information-sharing, deliberation, and networked communications.Thus, in some cases, institutional context and the nature of the profession make epistemic community emergence more challenging.
The article is organized as follows.The next section briefly reviews some of the literature on epistemic communities, and situates my argument about the limits of epistemic communities within this.Subsequently, I examine the case studies of the EDA and IntCen, and explain why they do not constitute or contain epistemic communities, despite being comprised of experts.Finally, I conclude that expert groups that do not constitute epistemic communities are not merely weak or nascent cases.Rather, they are fundamentally different kinds of actors, and are unlikely to emerge as epistemic communities without fundamental change to either the bureaucracies or professions to which they belong.

The Concept of Epistemic Communities and Its Limits
In a nutshell, epistemic communities have been defined as networks of experts who persuade others of their shared norms and policy goals by virtue of their professional knowledge.An epistemic community is rarely so broad as to include an entire discipline.Rather, all of its members must have the expertise necessary to understand the issues at stake, to interpret the information similarly, and then to form the same goals about what should be done.The group's policy aims have to reflect their expert knowledge-and not some other motivation-otherwise they lose authority with their target audience, which in the area of security policy is usually elite decision-makers.A strong epistemic community seeks to go beyond their formal professional role as a group, and is often able to persuade decision-makers to fundamentally change the nature of their policy aims.A weak or nascent epistemic community may be able to achieve incremental change over time, but only on occasion or without a high level of ambition.
Haas defines the concept as, "a network of professionals with recognized expertise and competence in a particular domain and an authoritative claim to policyrelevant knowledge within that domain or issue area" (Haas, 1992a, p. 3).In other words, epistemic communities must have an authoritative claim on knowledge to impact policy outcomes.Naturally, knowledge itself may be socially constructed, but epistemic communities must nonetheless have a means of objectively recognizing the validity of their knowledge.As Haas describes, epistemic communities (1) share professional judgment on a policy issue, (2) weigh the validity of their policy goals in their area of expertise, (3) engage in a common set of practices with respect to the problem area with the goal of improving human welfare, and (4) share principled beliefs (Haas, 2001, pp. 11578-11579).Thus, there is an ideational core that brings together these components of professional expertise.
Peter Haas, Amy Verdun, Claudio Radaelli, and others have applied the concept of epistemic community to empirical case studies of environmentalists, economists, and scientists.All agree that epistemic commu-nities exercise agency only when there is a kind of contextual gap-uncertainty-that allows them to do so (Haas, 1992a;Radaelli, 1999a;Verdun, 1999).To determine when epistemic communities will likely have the most influence, Radaelli argues that it is important to find these "conditions of radical uncertainty and political visibility" (Radaelli, 1999a, p. 763).Usually this comes in the wake of some kind of crisis or triggering event.Epistemic communities may exist prior to being "called" to action, but according to these scholars their impact is contingent upon this new source of uncertainty.
Adding to this literature, I have highlighted a number of key points that enable us to more clearly recognize epistemic communities when we look for them (Cross, 2013a).First, it should be understood that the process of professionalization and professionalism itself are at the heart of epistemic community cohesion.This is often where common points of reference and behavioral rules arise and are internalized.Second, I argue that epistemic communities do not simply exist or not exist, but can be characterized as strong or weak.Once an epistemic community comes together, it can exercise varying degrees of agency.Third, it should be acknowledged that non-scientific knowledge can be just as influential as scientific knowledge.Diplomats, judges, defence experts, high-ranking military officials, bankers, and international lawyers, among others, all have just as much of a claim to authoritative knowledge as scientists.And finally, I contend that epistemic communities are often always at work, and thus it is not necessary to look for a triggering crisis or episodes of uncertainty to observe their influence, as Haas and other contributors to the 1992 special issue have argued.While it is true that a crisis of uncertainty may provide an opportune time for epistemic communities to gain a voice, they are often quite influential even in the absence of such a triggering event.After all, uncertainty is a built-in feature of the international system, especially in the EU, which by its nature is a workin-progress with no agreed-upon end-goal.
In addition, there are at least three indications (see Table 1) that a group of experts might constitute an epistemic community or not.A first indication of the presence of an epistemic community is whether an institutional group or committee seems to act as more than the sum of its parts.Is it producing outcomes that go beyond the expectations of its formal functions?A second indication is whether its members know each other or have worked with each other in previous settings.If so, they might meet often outside of work and in informal settings.Such interaction, going beyond the call of duty, would suggest that they have opportunities to really deliberate on issues important to them.For this to have significance, such interactions naturally benefit from smaller numbers of individuals involved.A third indication is whether or not the group shares a particular culture and professional norms that are in-dependent of their formal function.Together these three criteria are helpful in distinguishing not only the difference between a strong or weak epistemic community, but also between a weak epistemic community, and an ordinary bureaucratic committee or other type of actor.Much work has been done in explaining what to look for in recognizing an epistemic community, as well as anticipating how persuasive it is.But what preconditions contribute to or detract from the existence of an epistemic community in the first place?Looking at the qualities that do not define epistemic communities is useful in ensuring that the framework is not overdetermined.I argue that institutional and professional context, in particular, have a direct bearing on the presence or absence of the three indicators outlined above.First, if a group of experts is housed within a highly formal institutional setting, with strict hierarchy and goals that explicitly limit its mandate from the outset, this is a difficult environment to foster the existence of epistemic communities.However, a limited mandate on its own is not enough to preclude epistemic community emergence.The example of the Committee of Permanent Representatives (Coreper) within the Council of the EU illustrates this well.Coreper was cre-ated with the limited mandate to prepare meetings for the Council of Ministers, but over time Coreper quickly became "a de facto decision making body" (Heinisch & Mesner, 2005, p. 1) as it emerged as a powerful epistemic community in its own right (Cross, 2007(Cross, , 2011)).When a limited mandate is combined with a highly formal and hierarchical institutional setting, this restricts the existence of epistemic communities quite significantly, as discussed in the EDA case study.Even though Coreper is formally an intergovernmental institution, with each member representing each of the member states, it nonetheless has benefitted from being a horizontal body with a sense of equality across the diplomatic experts who comprise it.No such truly transnational space exists within the EDA, despite the fact that it is a supranational bureaucracy; although, as I will discuss, it does emanate outward from it.
Second, if a group of experts comes from a profession in which the qualities of expertise explicitly work against information-sharing, informality, deliberation, persuasion, and transparency among people in the same profession, this also creates limitations for epistemic community emergence.Such professions can be wide-ranging.Intelligence professionals or spies-as in the case of IntCen-value secrecy as an integral part of their expertise.Defence industry experts have traditionally guarded their technological advancements so that they can be at the cutting edge of their own fields and distinguish themselves from those in other countries or corporations.Indeed, in many corporate environments, intellectual capital is valued specifically because you are the only one to possess it.Similarly, experts in the computer and information technology fields typically keep their breakthroughs secret-either because they use this knowledge for nefarious activities like hacking or spying, or to effectively counteract those engaged in such activities.Inventors, as well, will keep their discoveries under wraps until they can protect their discoveries with patents.And of course journalists, both in their relationship with sources and competition with peers, must learn the value of secrecy as a central part of their professional expertise.While there are always exceptions, for the most part, a built-in quality of these professions and others is the ability to safeguard against the kind of informationsharing, deliberation, persuasion, and informality that can lead to networked interactions, and to the formation of epistemic communities.
At the same time, it is important to note that just because an expert group does not constitute an epistemic community does not mean that it is somehow ineffective or falling below its potential.It may simply be a different sort of actor.There are many kinds of transnational actors or networks-often comprised of professionals-that scholars have identified and researched in depth.Beyond epistemic communities, these include communities of practice (Adler & Pouliot, 2011), business networks, advocacy networks (Keck & Sikkink, 1998), interpretive communities (Johnstone, 2005), and argumentative communities (Collins, 1998), among others.Many of these groups are held together by shared values and a common motivation to achieve specific goals in the international arena, whether to improve environmental regulation, protect human rights, or promote EU integration in new policy areas.
Moreover, it is still possible that a committee or expert group could achieve the outcomes we might expect of an epistemic community, but without actually being one.After all, groups of individuals can exercise independent agency for a variety of reasons and with many kinds of motivations.Transgovernmental networks, communities of practice, and advocacy networks, may not push for policies based on their expert knowledge, but may still achieve outcomes that go beyond expectations if we were to only consider member-state preferences alone.

Two "Non-Cases"
EU security policy is an area in which there are at least several Brussels-based epistemic communities.EU security epistemic communities are comprised of diplomats, military experts, security researchers, and civilian crisis management experts, among others, with a recognizable claim to expertise in both internal and external security policy (Cross, 2011).They have been shown to significantly impact outcomes of EU security policy beyond what would be expected by looking only at member-states' initial preferences.But just because there are many examples of security epistemic communities in the EU, does not mean that all groups of security experts comprise epistemic communities.Indeed, they may not even constitute weak or nascent epistemic communities, despite working together on a daily basis and making decisions that influence policy in some way.By focusing on an issue area that is heavily populated with experts, and within which epistemic communities are numerous, it is easier to see what differentiates certain groups of experts from those that form epistemic communities.
Both the European Defence Agency (EDA) and the Intelligence Analysis Centre (IntCen) are EU agencies, based in Brussels, and populated with security experts.Although there are two possibilities-the agency as a whole could function as an epistemic community, or the agency may contain one or more epistemic communities within it-I will argue that the professional and institutional contexts in both cases are not conducive to the existence of epistemic communities.As such, there is little evidence of the three indications discussed in the previous section, i.e. outcomes that go beyond formal agency functions, frequent informal deliberation, and culture and professional norms that are independent of the agency.It should be noted that these cases are not intended to be exhaustive analyses of the EDA and IntCen, but rather test cases for my hypotheses that professional and institutional context matter in fostering the growth of epistemic communities.Table 2 below summarizes these findings.

The European Defence Agency
The EDA was founded in 2004 with the aim of improving the EU's defence capabilities through promoting collaboration, common initiatives, and innovative solutions to the EU's security needs.As an intergovernmental agency, the EDA is designed to bring member states and their priorities together in the area of armaments and defence research, including investment and procurement, with the overarching objective of improving member states' collective military capability over the longer term.Rather than seeking a particular goal in a particular form, it searches for synergies across priorities, member-states, and projects.EDA policy aims mainly include pooling and sharing of resources, achieving interoperability, diminishing duplication of spending, emphasizing civil-military strategies, and agreeing on best practices.
The EDA's steering board, comprised of the Defence ministers of the member states, has the responsibility to make decisions about the overall guidelines for EDA operations.Defence ministers meet twice per year, but specific national representatives meet more frequently.In terms of the permanent staff, there are around 130 professionals working in the EDA.Each staff member possesses a high level of expertise, with extensive previous background in the defence field or in the military.The EDA's professional staff is selected based on merit, not a quota system or an effort to represent each member state equally (Giergerich, 2009).This is in contrast to other previously identified epistemic communities in the security area, such as the EU Military Committee or Civilian Crisis Management Committee (Cross, 2011).These committees, while housed within the formal Council hierarchy, are still horizontal in and of themselves.Each member state has a representative who sits in the committee, and they operate on the principle of one member state, one voice.Thus, the sense of equality among these military or civilian crisis professionals is strong, enabling more deliberation, informality, and informationsharing.Within the EDA, however, department sizes tend to be small, and the space for transnationalism within the EDA is not as strong as in an institution like the Council of Ministers, which is comprised of committees and working groups with a representative from each member state, all of whom have equal standing.
Beyond the institutional structure within the EDA, these professionals interact with around 4,000 defence specialists from across the participating member states (EDA, n.d.).EDA experts coordinate the formation of so-called Integrated Development Teams tasked with the work of determining defence capability needs in a wide range of areas.They also form around 20 Project Teams with national experts, focusing on specific defence initiatives that member states wish to pursue together, such as the helicopter initiative.Finally, they form Capability Technology Groups (CapTech) that work on collaborative research and technology projects with experts from participating member states.Sometimes the experts in these groups are also drawn from academia, industry, or other research groups.With the proliferation of so many expert groups in the field of defence, and all with connections to the EDA, it may be surprising that the EDA itself does not house epistemic communities of some kind.
The key distinction is that EDA experts are fulfilling a coordinating role, rather than becoming agents for change in their own right.The EDA serves as a kind of hub for these larger, defence-expert networks.This is actually quite similar to the ways in which European think tanks devoted to EU issues work as well.EU think tanks host numerous meetings, seminars, and conferences on the topic of EU security, bringing hundreds of people together to discuss important security topics.However, these think tanks do not have close ties with one another, tend not to advance shared policy goals, and generally serve more as forums for other professionals to meet rather than acting as policy leaders in their own right.Similarly, EDA experts also serve in this function vis-à-vis the numerous networks of defence experts that are connected to their activities.While some of these networks may very well constitute epistemic communities, EDA experts serve to coordinate and manage the network of national defence experts, as well as to get a better sense of what member states want through these interactions.EDA experts themselves do not comprise epistemic communities in their own right.In the words of the agency itself, these networks of experts "are crucial for EDA's work as they ensure coherence with national priorities" (EDA, n.d.).
The EDA's activities have mushroomed in recent years, but not because of any specific push from EDA staff.Rather, the EDA's "way of working" is to face outwards instead of inwards.That is, it is a facilitator, information supplier, and momentum generator, bringing member states together on goals they have agreed to, but may not be able to achieve without the help of the EDA.The agency exerts light pressure on memberstates to achieve follow-through, without any specific effort to persuade them to do something that they are not already comfortable with (Arnould, 2011).The reason for this is that the EDA as whole takes a more pragmatic and piece meal approach, rather than pushing for an overarching goal, like achieving more integration among member states.The agency may ultimately encourage more integration, but this is more indirect, rather than purposeful.As an information supplier, it serves as a clearinghouse for information that member-states would not otherwise have about each other (Giergerich, 2009).On its own, this information can "name & shame" member states while keeping the EDA neutral.But EDA staff also have a norm of being flexible in creating a framework for armaments integration so they can adjust the level and nature of ambition based on member-state needs (Trybus, 2006) and desires to participate (EDA, n.d.).
Ultimately, it is clear that the overarching impetus behind EDA initiatives comes mainly from the member states, and not from epistemic community activity stemming from the EDA.Indeed, even in recent years, member states' representatives have spoken strongly about the need to create a common European Defence, characterizing integration as a "non-choice".Wolfgang Ischinger, former state secretary of the German foreign office, writes, "starting to Europeanize our defence is the only reasonable way forward" (Ischinger, 2012).Belgian Defence minister Pieter de Crem said, "it is better to have collective capabilities rather than non-existent national ones" (de Crem, 2012, p. 5).The Franco-German Declaration of February 6, 2012 states, "In times of strategic uncertainty and limited resources, strengthened defence requires common procurement" (Franco-German Declaration, 2012, p. 2).An Italian document proposing goals for the December 2013 European Council on Defence states, "If EU member states do not pool their efforts, where appropriate on certain common requirements or capabilities, none of them, nor Europe as a whole, will be able to guarantee its own security" (Italian Ministry of Foreign Affairs & Italian Ministry of Defence, 2013).Through an analysis of EDA public statements, André Barrinha writes that the EDA: has expanded the "existential condition" justification, and included all the other arguments within this one.The EDA has, in that matter, become the organising core within the European Union regarding a European discourse on defence industries.More than the ESDP, more than the potential positive or negative relations with the United States, and more than the mere economic rationale, Europe must unite for its own survival-this is no longer an option (Barrinha, 2010, p. 481).
The image of the EDA as an "organizing core" is telling.Indeed, as the above quotations indicate, member states' statements about the importance of the work of the EDA are typically stronger than what EDA officials themselves say.
I argue that a significant reason why the EDA lacks epistemic community activity, and as such, does not seem to go beyond its formal mandate as an EU agency, is that it is an example of a bureaucracy that houses experts in a highly formalized and hierarchical institu-tional structure.From the outset, the EDA was designed to serve the member states rather than to direct them.The fact that the internal workings of the EDA do not rely on the principle of one representative from each of the member states in each of the various directorates, also adds to the sense of hierarchy, instead of allowing for horizontal and transnational spaces within the agency.Moreover, defence integration is voluntary on the part of the member states, and the professionals within the EDA do not aspire to change this.Rather, they allow the member states to take the lead in crafting the direction and degree of defence integration (Arnould, 2011).To be sure, in its young history, the work of the EDA has grown in scope and scale, adding dozens of projects and goals.There is nothing in its record thus far that would indicate any serious failings.Yet, it has adhered closely to its original mandate, and has stuck to its role of generating momentum behind the expressed political will of the member states.This contrasts with similar expert groups housed within more horizontal institutional structures-like the EUMC and Civcom-that do act as epistemic communities.

The EU Intelligence Analysis Centre
IntCen's chief mandate is to provide intelligence analysis and strategic assessments to EU decision-makers, especially in the area of counter-terrorism.The agency operates twenty-four hours a day and seven days a week to ensure that it is able to provide rapid updates, especially to the High Representative.IntCen intelligence experts work closely with the EU Military Staff, External Action Service, and to some extent, the European Defence Agency (Cross, 2013b).They routinely provide "flash reports" on international crises as they develop, and may issue early warnings in particularly urgent cases.They also constantly monitor potential terrorist threats, the proliferation of weapons of mass destruction, and conflict-prone locations around the globe, so that they are prepared to respond immediately in the event of a crisis (Cross, 2013b).On a medium-term basis, they provide several services involving the Common Security and Defence Policy (CSDP), such as recommending procedures for crisis management, risk and situation assessments, and crisis response facilities.This kind of readily accessible intelligence at the EU level is a crucial component of the EU's ability to speak with one voice in terms of common foreign policy, and to respond quickly to events.If analysts are onlocation they may serve as the operational contact for the high representative.On a longer-term basis, IntCen experts focus on strategic assessments that can build stronger resistance to terrorist attacks over time.For example, analysts deal with aviation security, cybersecurity, and problems of radicalization and recruitment (Hertzberger, 2007, p. 68).Their priority is to gain a better understanding of the internal dynamics, fi-nancing, ideology, and potential targets of terrorist networks (Duke, 2006, p. 607).
The internal structure of IntCen is not excessively hierarchical, but the professional environment is highly formal and restrictive, as it is in the nature of the intelligence profession to avoid transparency and information sharing.Even as far as intelligence goes, IntCen itself is very secretive and closed off from public scrutiny, making it difficult even to pinpoint basic elements of its structure, responsibilities, and evolving role.We do know that IntCen is comprised of both decisionmaking bodies and implementation bodies.IntCen is comprised of experts who have autonomy, and a mandate to improve the EU's intelligence-sharing sphere.It has a staff of around 80 experts,2 which includes analysts of both civilian and military backgrounds, as well as other support staff.The analysts are typically seconded from national intelligence services, and are double-hatted to both (Hertzberger, 2007, p. 69).After the Lisbon Treaty, the number of analysts within IntCen increased (Council of the EU, 2011).Each year, IntCen intelligence experts produce some 100 intelligence reports, 40% of which deal with terrorism assessments (Hertzberger, 2007, p. 66).
IntCen professionals have no formal mandate to engage in intelligence gathering, as traditionally understood, and rely to a significant extent on intelligence provided by member states on a voluntary basis.For example, they receive information from the French, German, and Italian spy satellites for imagery, as well as from member states' diplomatic reports.Between seventeen and twenty EU member states provide national intelligence to IntCen, so not all member states participate, but all twenty-seven do receive IntCen's reports.Each member state can even stipulate who is allowed to see information they provide to IntCen, beyond those who regularly consume the reports, under the so-called "originator principle" (Rettman, 2010, November 18).To the extent that IntCen experts do originate intelligence themselves, it usually comes from open-source information, or on-the-ground observations in crises.For example, IntCen professionals can use US commercial satellite imagery, Internet chatroom intelligence, media reports, and information gathered from within the European External Action Service.In addition, these intelligence experts routinely travel to crisis zones and CSDP operation locations to gain a better sense of real conditions on the ground.
As with the case of the EDA, there is little evidence that the intelligence experts within IntCen have exercised agency as an epistemic community.Intrinsic to the intelligence profession is the ability to maintain a high degree of secrecy, especially when it comes to transnational interactions.Moreover, for IntCen professionals, the primary goal is to do their work better rather than redefine their work to change the direction of policy in the intelligence area (Cross, 2013b).However, rather than comprising epistemic communities, I argue that the professional staff within IntCen is actually part of a larger transgovernmental network of intelligence experts across Europe.
Transgovernmental cooperation more generally is the process by which sub-units of governments engage in direct and autonomous interaction separate from nation states (Keohane & Nye, 1974).Transgovernmental networks can be quite informal, and do not necessarily have a specific agenda or policy goal in mind (Grevi, 2008;Thurner & Binder, 2009).Rather, they are more focused on processes of governance.Anne-Marie Slaughter (2004) has documented how transgovernmental networks across the globe-judges, legislators, regulators, and so on-are increasingly coming together in this way to share best practices and knowhow.This is also increasingly true of intelligence experts across Europe, of which IntCen professionals are a central part.A key example of this sharing of best practices is in the area of open-source intelligence, which is of growing importance in the intelligence profession with the widespread use of the internet and social media (Pallaris, 2009).As a result, increasing numbers of intelligence professionals participate in informal networks that enhance their ability to do their job well.
For example, Eurosint Forum, founded in 2006, is a non-governmental, non-profit organization based in Brussels that holds around five workshops a year and comprises a network of around 400 intelligence professionals, at all ranks, from member states' intelligence agencies, private-sector organizations, and EU institutions such as the EU Military Staff, SitCen, and Europol.
Each workshop usually consists of around 35 participants, but Eurosint also holds one or two larger conferences each year with more than 100 participants.According to Eurosint General Manager, Axel Dyèvre, these workshops and conferences have many opportunities for informal interactions that clearly create an atmosphere of trust, emphasize an exchange of ideas, and allow for brainstorming (Dyèvre, 2011).Rather than discussing topical and potentially confidential issues, the focus is on getting to know each other, finding areas of potential collaboration, and discussing practices.Several shared projects have emerged from these Eurosint gatherings (Dyèvre, 2011).Besides the Eurosint Forum, EU intelligence experts have long met in more informal settings to foster transgovernmental cooperation, such as in the so-called Berne Group or Budapest Club.This kind of interaction is leading to the creation of a kind of European intelligence space (Cross, 2013b), but one that still carefully protects information from crossing borders.Thus, IntCen is an example of an agency that is comprised of security experts whose profession is highly limited in terms of fostering epistemic community emergence.Since the need to keep information secret is wrapped into the very expertise of spies, in many ways deliberation, transparency, and informationsharing is highly restricted.Nonetheless, IntCen experts have found ways of networking despite this.They focus more on professionalism and best practices rather than the real substance of their knowledge.In her study, Hertzberger (2007) finds that personal contacts among IntCen experts have actually led to better intelligence cooperation in Europe over time, and towards an emerging institutional culture.Thus, like in the case of the EDA, these experts comprise a different kind of actor, and one that is larger than the Intelligence Agency itself.But the nature of their profession prevents the emergence of an epistemic community because sharing their knowledge goes against the nature of their expertise.Table 2 summarizes the suggested findings from these two illustrative cases.

Conclusion
These two "non-cases" show that it is not really the policy area itself that determines the emergence of epistemic communities.After all, numerous epistemic communities exist in the same policy area as the experts that work in the EDA and IntCen.For example, the EU Military Committee, Civilian Crisis Management Committee, and Political and Security Committee constitute epistemic communities of varying degrees of influence in the area of external security policy.
Rather, these case studies suggest that bureaucratic structure and the nature of the profession of those involved can serve to limit epistemic community formation.Much of the empirical research on epistemic communities focuses on single case studies of epistemic communities and traces their role in influencing policy choices.As such, the literature tends to take for granted that epistemic communities either exist or they do not.My aim in this article is to take this reasoning a step back to explore the conditions that might limit epistemic community emergence in the first place.I argue that institutions whose internal structure is hierarchical and formal, like in the case of the EDA, typically do not enable enough of a critical mass at any single horizontal level to encourage epistemic community formation.They also tend to advance a strict mandate from the start.Thus, they operate as a traditional bureaucracy, only carrying out the autonomy granted to them, rather than trying to go beyond this.
Secondly, professions that require secrecy and nontransparency with others in the same profession, or as an integral quality of their professional expertise, are also likely to limit epistemic community formation.This is certainly true in the case of the intelligence profession.By way of contrast, the diplomatic profession does tend to foster the existence of epistemic communities (Cross, 2007).Even though diplomats must main-tain a degree of secrecy, deliberation is a fundamental part of their profession and the ability to share knowledge about possible areas of agreement is crucial.The "art of compromise" is cultivated within diplomatic corps, and networks of professional diplomats understand that they can enhance trust by knowing when and how to share secret information with each other.Diplomatic deliberation often occurs without public scrutiny, but part of what enables fruitful, informal discussion, especially in the European context, is the ability for diplomats to speak frankly about where they can find common ground.
Many kinds of networks, groups, institutions, and bureaucracies have influence on policies and outcomes in international relations.However, it is important to recognize that they also have different kinds of influence.Thus, I am not arguing that the EDA and IntCen are weak or ineffective agencies, but rather that they seek influence in ways that conform closely to their mandate, and do not stretch the boundaries of this.By contrast, epistemic community influence often involves changing the very basis of the way states operate in the international system, as well as the rules they follow in their interactions.This requires high levels of status and persuasion in the eyes of state leaders.Thus, in the cases of IntCen and the EDA, as long as professional and institutional context remain unchanged, it is unlikely that epistemic communities would form in the future.
These arguments are based on just two illustrative cases, but future research could further explore the limits of epistemic community emergence with a specific focus on which kinds of professions and bureaucratic structures might be more or less conducive to this.For example, the cases of the EU Military Staff and the Civilian Planning and Conduct Capability are similar cases of supranational bureaucracies that are also hierarchical.In particular, it would be valuable to explore the preconditions for epistemic community emergence in other countries and regions of the world with varying types of regimes.In authoritarian states, for example, bureaucracies are typically extremely hierarchical and formal, and a greater number of professions may be prevented from engaging in deliberation and networking.Epistemic community emergence may be virtually impossible in these more extreme cases.And in other democracies, certain professions may have different norms and practices than in the European region.Such comparative work is useful in mapping out the various conditions that limit or encourage the existence and influence of epistemic communities.
The decision-making bodies consist of the Intelligence Steering Board, chaired by the High Representative and Vice-President of the Commission, and the Intelligence Working Group, chaired by the directors of IntCen and the EU Military Staff's Intelligence division.The implementation bodies consist of IntCen's expert staff itself, and the intelligence directorate of the EU military staff.Unlike Europol, IntCen prepares intelligence analyses for EU decision-makers, rather than authorities in the member states.Its target audience includes High Representative Federica Mogherini, Counter-Terrorism Coordinator Gilles de Kerchove, Coreper II, PSC, the Working Party on Terrorism, the Article 36 Committee, the Policy Unit, and decision-makers in the area of police and judicial cooperation (Müller-Wille, 2008, p. 59).

Table 1 .
Three indications of the presence of an epistemic community.

Table 2 .
Summary of findings in the cases of the EDA and IntCen.