Circulation, not cooperation: towards a new understanding of intelligence agencies as transnationally constituted knowledge providers

ABSTRACT Theoretical approaches to international intelligence relations primarily rely on the inadequate concept of cooperation. I offer an alternative analytical framework based on the concept of knowledge circulation. Knowledge circulation highlights that intelligence agencies are not highly bounded, national actors, butconstituted by transnationally circulated knowledge. Knowledge circulation draws attention to an autonomous realm of transnational intelligence relations, and thus offers insights for scholars, policy makers and practitioners. The article references understudied examples of intelligence relations and closes with an example of how the story of US-WestGerman intelligence relations could be retold via a knowledge circulation framework.

relations have become ever more visible, while remaining theoretically misunderstood and politically underregulated. The lack of regulation and oversight is moving into the political foreground, as policy-makers and the public are becoming increasingly aware of the impact of international intelligence ties. Thus, in May 2020, the German Supreme Court ruled that German intelligence's foreign relations required new controls, as the status quo violated the constitution. 11 The ruling came after journalist organizations had sued, arguing that the way German intelligence was sharing knowledge with foreign counterparts could endanger journalists working in countries lacking press freedom and a prevailing climate of political oppression. Since the Supreme Court's ruling, German lawmakers have struggled to develop a new intelligence law that is adequate, amid controversial public debate. 12 In Britain, the country's departure from the EU has shredded many transnational, institutional ties, yet research shows that UK intelligence relations with the EU remain fluid and durable, a matter which raises the question of how far international intelligence relations constitute an autonomous, transnational realm. 13 The gap in scholarly attention to international intelligence relations begs important questions: do intelligence agencies develop autonomous power that they exercise internationally? What makes an intelligence agency powerful within the international intelligence realm? How does the international power of an intelligence agency relate to the power of its home state? Do intelligence agencies influence each other and, if so, how? How important are different factors, such as individuals, institutions, technology and information, in these relations? Studying only specific instances of intelligence alliances, as the current literature mainly does, is of little help here. What is needed instead is better general theory. In this article, I advance two main theoretical arguments: firstly, I argue that international intelligence relations can be better understood via the concept of knowledge circulation. 14 Secondly, I argue that, considered via the prism of knowledge circulation, modern intelligence agencies emerge as fundamentally transnational actors, who have always been constituted via foreign relations. This re-conceptualisation of intelligence as transnational knowledge circulation carries implications for scholarship, but also offers benefits for policy-makers and practitioners, who may find its premise eye-opening in terms of how to understand and approach their own, 'national' intelligence providers.
To reach my arguments, I provide a thorough critique of the cooperation concept, which is the dominant model currently applied to international intelligence relations. While the cooperation literature has produced some important conceptual improvements to intelligence theory overall, which I also highlight, it relies on an overly static rational-choice model and, most importantly, proposes that intelligence agencies are inherently hostile to foreign relations with each other, applying them only as a last resort when all other options have failed. 15 This new research paradigm may even reveal that intelligence practitioners themselves consider their work as deeply transnational in character, rather than as a process carried out primarily within national boundaries.
Building on a 1996 intervention by leading intelligence scholar Michael Herman, 16 and a set of intelligence literature influenced by sociological theory, 17 I argue that the theory of knowledge circulation offers a strong theoretical and methodological basis to improve research and theory regarding international intelligence relations. Knowledge circulation argues that intelligence is a transnational process, product and institution, 18 and focuses attention on the moments, people and practices, which set knowledge in motion. It differentiates between moments of knowledge communication, reception and processing. Knowledge circulation allows us to posit the hypothesis that intelligence agencies and their knowledge are 'the outcome of an intellectual assembly line that is internationalized from the outset'. 19 Proceeding from this hypothesis means abandoning the currently prevalent assumption that intelligence agencies are ontologically national actors for whom international contact with each other is the exception, rather than the rule. Instead, it becomes an empirical question of how transnationally open, or inward looking any particular agency is. Herman, in a hitherto underappreciated chapter of his otherwise influential book Intelligence Power in Peace and War, argues as much when he states that in certain areas, intelligence 'has become a kind of international knowledge system', akin to academia's transnational invisible colleges and 'other "expert" intergovernmental relationships'. 20 Despite Herman's influential position in other areas of intelligence studies, cooperation scholars have ignored Herman's position, set out here, on international intelligence relations. In all probability, this is because Herman's points upend the assumptions of the cooperation model, which has functioned as a powerful, normative 'discursive resource', as has been shown for other analytical short hands. 21 This article proceeds as follows. Firstly, I critique the cooperation concept. Secondly, I develop knowledge circulation as a useful theoretical framework and detail three concrete ways in which knowledge circulation helps to highlight under-theorised aspects of intelligence relations that have fallen by the wayside under currently dominant thinking. Thirdly, I use the history of US-West German intelligence relations as a brief example to show how knowledge circulation may refocus IR scholars' attention on hitherto underappreciated aspects of this relationship.

'Cooperation': the dominant model for understanding international intelligence relations
When modern intelligence was reordered after the Second World War, agencies across the globe established international relations, building on historical ties, many of them colonial or post-colonial, or creating new ones. 22 Both IR scholars such as Jennifer Sims and intelligence scholars such as Adam Svendsen or Bradford Westerfield have developed analytical categories for different types of international intelligence relations. 23 They identify as important divisions those between formal or informal contacts, bilateral or multilateral contacts and institutional or ad-hoc contacts. Welldocumented examples of formal and informal multilateral intelligence relations include the UKUSA alliance and its consequent development into the 5 Eyes intelligence network between Commonwealth members, 24 the Club de Berne between primarily European domestic intelligence agencies, 25 EU intelligence-sharing mechanisms 26 and, historically, the cooperation between the intelligence agencies of the Warsaw Pact. Bilateral relations between both foreign and domestic intelligence agencies have also been extensively researched historically, 27 including beyond the Western world, 28 both during war and peace, 29 between allies and enemies, 30 as institutional arrangements based on written agreements, or as ad-hoc contacts arising out of momentary needs, personal ties or even coincidence. 31 Thus, the historical evidence that intelligence agencies maintain significant foreign relations among each other is large and extensively documented, however also hampered by a strong bias towards Europe and the Anglo-Saxon world, which is only slowly changing. 32 Shiraz and Aldrich highlight the important work done on South America; for the Middle East there exist a number of memoirs by intelligence officials; this is also the case for France. 33 African intelligence is moving increasingly into the focus of research. 34 This wealth of empirical and historical labour on intelligence relations is not matched by similar efforts to analyze its meaning for international politics. Rather, the theoretical literature on international intelligence relations is thin, and dominated by a single, underdeveloped concept, that of cooperation, also referred to as liaison. 35

Strengths of the cooperation concept
For the sake of theoretical clarity, it is important to first highlight how the cooperation concept has improved general understandings of intelligence. Firstly, cooperation scholars argue, and have shown, that intelligence agencies are actors with interests, and not just neutral service-providers to the government. This is a significant analytical step, because it counters the assumptions of the Intelligence Cycle, another central model for intelligence studies, that explains the institutional position of intelligence agencies within their national government and intelligence's relationship with the national executive. 36 The Intelligence Cycle models the work of intelligence agencies as five (sometimes more) interlinked functions: 1) planning/direction from the executive; 2) collection; 3) processing; 4) analysis; 5) dissemination back to the executive, which re-starts the cycle. 37 The intelligence cycle has a critical flaw: it offers no conceptual space to analyze agencies as actors in their own right, and thus can provide no substantial guidance as to how agencies' own properties may shape the intelligence process, and its relationship to politics. Here, the cooperation literature offers a much more robust understanding, as it clearly conceptualizes intelligence agencies as bounded actors, whose internal politics, knowledge-production and institutional apparatus differentiate them from other state institutions, with whom they may form specific relations.
Regarding interests, the cooperation concept gives analytical room to analyze intelligence agencies as actors that may develop interests independent of those of the executive. Once again, this is a significant improvement over the purely functional understanding of intelligence agencies offered by the intelligence cycle and is an important step towards a more inclusive intelligence theory that moves beyond the Anglo-Saxon world. This is because in many countries of the Global South experiencing authoritarian rule the idea of a disinterested intelligence agency is a concept far off the mark: instead, governments frequently engage in coup-proofing their intelligence agencies to prevent the emergence of independence. 38 Under authoritarian rule, countries may develop highly fractured intelligence systems, with high levels of internal competition; one extreme example of this is Syrian intelligence. 39 Yet it is important to note that European and Western agencies may also develop divergent convictions from their political masters about what constitutes the national interest. For example, Stefanie Waske's research on West Germany in the 1970s shows that there existed a secret alliance between the foreign intelligence agency BND and conservative politicians, to covertly act against West Germany's first social democratic government, elected in 1969. 40 Also, the various Western intelligence failures in the run-up to the 2003 Iraq war provide further examples for what Gordon Mitchell has referred to as 'team B intelligence coups', in which interested 'intelligence boutiques stovepiped misleading intelligence assessments directly to policy-makers', undercutting the usual intelligence process. 41 Cooperation scholars link the existence of independent intelligence interests to research about why some forms of international intelligence relations flourish while others flounder, despite evident political desire to make them happen. One emerging finding from several projects is that international intelligence relations are most stable when they are initiated due to internal agency interests, rather than on command from external policy makers. 42 For example, Björn Fägersten writes: Motivated by interests as well as ideas, national bureaucrats have consistently shirked from designated tasks and failed to respond to their masters' exhortations. By doing so, they have effectively separated government ambitions from actual outcomes in the area of international intelligence cooperation. Hence current cooperation only imperfectly reflects the preferences of participating states. 43 Fägersten argues that intelligence agencies' high levels of investment in informal networks reduces their interest in participating in formal exchange agreements, even though the latter may be the preferred option of their governments. Jennifer Sims states it even more clearly, when she speaks of 'intelligence interests' that determine the nature of a foreign intelligence relationship, and which may run counter diplomatic or other political interests: When an intelligence service is particularly imbued with an apolitical culture, as are the best ones, its calculation of whether to exploit, strongarm, or deeply penetrate a liaison partner may turn less on the risks of disrupting the overall political relationship than on the intelligence gains or losses.
It is for this reason, writes Sims, that 'policy oversight of liaison practices is so important, and so dangerous if done poorly'. 44 Secondly, cooperation scholars have usefully applied economic models to highlight that intelligence agencies trade with each other -in information, technology or other resources. In his classic article on intelligence liaison, Bradford Westerfield applies an international political economy model to cooperation, arguing that certain traits of US intelligence give it comparative advantages and disadvantages vis-à-vis other countries, and that there results a particular 'bargaining power in a global intelligence market'. 45 In this market, secrecy, as a professional practice, does not create barriers between intelligence agencies, but functions as a shared expertise, strengthening a sense of connection between international intelligence professionals. 46 James Igoe Walsh, author of one of the rare US university press books on international intelligence relations, pushes the market analogy further by applying 'relational contracting' as an explanatory framework. 47 Relational contracting is a concept borrowed from transaction cost economics. To explain why states engage in high-risk international intelligence relations Walsh argues that 'relational contracting draws attention to how states can incorporate hierarchical control, oversight, and monitoring mechanisms in their [intelligence] cooperative agreements in order to minimize defection by participants'. 48 By creating a hierarchical intelligence relationship, in which the dominant state obtains intelligence in return for providing a variety of goods to the subordinate state, both partners benefit while minimizing the risk that either partner will renege on the arrangement. 49 Unlike Westerfield, Walsh considers secrecy as a major hurdle to intelligence sharing, to which hierarchies offer one solution and intelligence specialization another. In a further analogy to international markets, Walsh argues that states can encourage cooperation by developing special skills in a certain area of intelligence collection, as this 'allows each participant to focus on the intelligence collection and analysis of targets and functions to which it can add the most value', and thus the overall intelligence value for each partner can be maximized. 50 This idea echoes theories about how an international division of labour between countries minimizes opportunity costs for each.
Thirdly, cooperation scholars highlight that beyond interests, intelligence agencies produce power, which may be harnessed by states, but may also be exercised with varying degrees of autonomy. 51 As mentioned above, highly autonomous intelligence agencies are frequently associated with post-colonial and authoritarian states. But intelligence agencies of democratic states may also engage in autonomous exercises of power. Perhaps the most extraordinary example to be recently revealed is Operation Rubicon , in which the CIA and the BND jointly owned the Swiss encryption company Crypto AG which, at the height of the operation, was the world-leading producer of encryption machines. 52 In the late 1960s, the two agencies bought Crypto AG, both paying US 8.5 USD million. From then on, Crypto AG's machines, sold to intelligence agencies and governments worldwide, were rigged: the company's secret owners could now easily read their encrypted messages. 53 While top West German government figures knew of the operation, secret files show that the BND agreed to far wider targeting than West German government positions would have allowed. 54

Weaknesses of the cooperation concept
Despite the above-listed strengths of the dominant cooperation literature, with regards to its central aim, namely explaining the reasons and outcomes of international intelligence relations, cooperation runs into significant problems. Although cooperation scholars generally share a similar, positivist and rational-choice framework, 55 disagreement continues on whether cooperation depends on state interests, 56 threat perceptions, 57 cultural and sociological factors, 58 technological and legal changes, trust, or even the individual preferences of intelligence professionals. 59 Contradictory statements are made about whether intelligence cooperation is embedded in and dependent on a state's diplomatic relations 60 or whether the opposite is the case, i.e., that (good) cooperation remains independent of diplomacy and politics. 61 A further problem is that cooperation scholars address cooperation on all kinds of levels of analysis: some treat states as unitary actors, others focus on the institutional level, others on individual officers, and others mix all levels together without offering guidance as to how they relate to each other.
A good example of this problem is provided by a quick comparative look at the works of Richard Aldrich and James Igoe Walsh, two scholars who have written influential works on intelligence and international relations. Both are realists and both broadly argue that states engage in intelligence relations to better counter national security threats. Both scholars pay nuanced attention to theory and the work of both is heavily empirical, drawing on historical methods to analyze richly researched case studies from the world of intelligence. Yet while Aldrich argues that 'the realm of intelligence is . . . perhaps the most human of all aspects of government and consists to a large degree of personal relationships. The universal currency is trust', Walsh argues that 'countries may share intelligence even when they do not have much trust in each other. They do so by substituting a hierarchical relationship for trust'. While Walsh's core argument is that countries overcome inherent risks of intelligence relations by establishing mutually beneficial hierarchical relationships, and by specializing in particular intelligence domains, Aldrich confidently states that 'the architecture of transatlantic intelligence co-operation remains a complex network with few key nodes or hierarchies and not a little duplication'. This type of exercise could be carried out between any number of cooperation scholars, highlighting that the cooperation concept has failed to provide a strong, overall theoretical framework to analyze international intelligence relations.
One root cause of this failing is that the cooperation concept includes several implicit assumptions about the actor-properties of intelligence agencies which need to be made explicit and critically examined, firstly in order to eliminate confusion within the cooperation literature, and secondly to move beyond cooperation's narrow lens, which remains influenced by the particular concerns of Western countries during the Cold War era. One implicit assumption of cooperation is the problematic premise that engaging in international relations is a necessary evil for intelligence agencies, which will only unwillingly maintain relations with each other for the purpose of compensating for gaps in their own collection abilities. This idea pervades the literature and is captured by the frequently quoted dictum, generally attributed to Henry Kissinger, that 'there is no such thing as friendly intelligence agencies. There are only the intelligence agencies of friendly powers'. 62 Due to an unproven assumption that intelligence agencies are by nature wary of maintaining relations with their counterparts, cooperation is conceptualized as a circumscribed, rational decision, which is strictly determined by an agencies' collection needs. Cooperation is considered as determined by strategic interests, understood as collection needs, which are determined by national security priorities and threat assessments. From this perspective, the widespread historical evidence of international intelligence relations must appear as a paradox, or as an 'oxymoron', 63 even though the same authors acknowledge that 'contact is routine', and 'the volume of exchange is large'. 64 Thus, the cooperation concept relies on an unproven definition of intelligence agencies as actors with a highly national perspective on security and international affairs, who consider foreign intelligence agencies as enemies, rather than collegial institutions. From this perspective, intelligence agencies are highly unwilling to share information, and will do so only if they receive something concrete in return, in a 'quid pro quo' fashion. 65 Shiraz and Aldrich highlight that this premise is linked to how 'Western liberal democratic states have . . . conceptualized their secret services as defenders of their freedoms against foreign adversaries', while in an earlier paper Aldrich emphasizes the impact of US intelligence's 'us vs. them' mentality on Western approaches to cooperation. 66 As a consequence, the literature on international intelligence relations is dominated by an ontological nationalism. Scholars mostly investigate how a single national agency maintains contact with other nationally-bounded agencies, rather than researching how agencies are institutionally rooted in foreign relations amid a constant, transnational flow of intelligence-knowledge. 67 The cooperation literature's broad commitment to problem-solving and policy-improvement has produced a strong emphasis on normative and practical questions such as what kind of cooperation should (not) be sought, or how agencies can get more out of cooperation, or reduce its risks. In principle this is a valuable approach. However, it risks ignoring a large amount of historically documented instances of international intelligence relations that are not directly linked to the sharing of information (or other operational resources), such as training, conferences, staffexchanges or jointly organized covert operations. Perhaps even more importantly, it risks ignoring those practices of foreign interaction which cannot be directly traced to strategic interest at all, but can better be explained by routine, bureaucratic institutions, personal relations, changes in diplomatic status or technological change. These include the longstanding use of automated signals intelligence sharing, highly institutionalized networks of foreign intelligence residencies abroad, requirements to upgrade technology as powerful allies improve their systems and other processes of organizational routine, which are well-known from other forms of international institutions 68 but have only rarely been researched in the field of intelligence. Further, it ignores the documented instances, in which intelligence officials engage in genuine diplomatic activity outside of a strict intelligence remit. Egypt's former intelligence chief, the late Omar Suleiman, who headed Egypt's General Intelligence Department for 18 years and simultaneously acted as one of Egypt's most important diplomats is only the most obvious example.

Approaches beyond cooperation
The weaknesses of the cooperation literature have not gone unnoticed, and a number of scholars have raised criticisms or attempted to introduce alternative approaches, without hitherto having a significant impact. Richard Aldrich has frequently written about the empirical complexities of intelligence relations, mentioning many of the factors that cooperation theory ignores, but without coming up with alternative explanations. 69 Together with Aldrich, Zakia Shiraz and John Kasuku have emphasized the ethnocentric nature of intelligence theory; for example, asking 'why do we think like Americans when we think about intelligence?' While these articles do not explicitly focus on intelligence relations, they raise important criticisms that can be applied to the cooperation literature as well. Yet, while doing so, they fail to offer a coherent suggestion for alternative frameworks that could be applied. In the following paragraphs I set out alternative suggestions before moving on to offer the knowledge circulation concept as a novel idea.
In a 2017 article, Jonathan Brown and Alex Farringdon addressed the puzzle that neither a shared strategic rationale (i.e., the standard cooperation argument), nor shared governance norms explain the large variety of international intelligence relations. 70 Answering Adam Svendsen's call 'to turn to less obvious bodies of ideas that focus on information exchange', Brown and Farringdon use sociological research on communication to argue that the depth of intelligence relations hinges on personal ties and on how these develop in terms of mutual expectations: Tying everything together, our framework suggests intelligence-sharing relationships are most likely to attain deep and enduring levels of cooperation when all three elements exist: a reason to exchange (information and friendship), a route for exchange (close personal ties), and a record of exchange (reliably meeting expectations). Relationships missing one or more are likely to remain shallow and short-lived. 71 Brown and Farringdon's piece is innovative, because it first turns to archival evidence to establish how intelligence is communicated, and only then proceeds to develop a cause and effect model of this communication. Unfortunately, their article is but a short intervention and focuses on a small set of European intelligence agencies during the Second World War.
The doyen of intelligence studies, Michael Herman, in a remarkable but underappreciated chapter of his influential 1996 book Intelligence Power in Peace and War takes a more global perspective, when he speaks of 'the fabric of the international intelligence system', 72 which 'became a form of international diplomacy in its own right'. 73 International relations are so ingrained into the intelligence process that 'most intelligence agencies are producing partly for an international audience of fellow-professionals, as well as for their primary national recipients', argues Herman. 74 He agrees that there 'is usually a bottom line of national professional-technical self-interest' and that intelligence relations 'are ultimately sets of professional bargains', but that these are embedded in historical, social and professional ties, 'which have their own momentum; once well established, considerable political weight is needed to disrupt them'. Herman describes how intelligence relations weave an increasingly integrated global system of knowledge production, as they require international standardization of technology and information security. 75 'Big nations' [intelligence] communities consciously and unconsciously shape the intelligence structures of the smaller partners with whom they deal' 76 and 'mutual education also promotes common action, just as having a common base of intelligence knowledge promotes agreed decisions'. 77 As deeply internationalized national-agencies become closely linked, 'national users' requirements cease to be the only basis for intelligence priorities. Agencies also need to maximize their value to foreign partners', which may create problems for the domestic political management of these agencies. 78 Didier Bigo, particularly known for his work on transnational policing in Europe, 79 takes this argument about a knowledge-based integration of allied intelligence agencies further, via the concept of shared secrecy. Bigo argues that 'national exclusivity on the means and construction of intelligence data' has never existed and that intelligence agencies maintain communication channels, via which 'segmented information that nevertheless travels globally as secret information' is shared by many. Both Bigo and Herman thus stress the notion of 'travelling knowledge', or 'knowledge in motion' as a key aspect via which to understand international intelligence relations.
A number of authors have called for the application of neo-realist IR theories on state alliances to international intelligence relations. 80 However, these approaches are hampered by their understanding of states as largely unitary and rational actors, which is not a good fit to research and analyse the 'low politics' and 'fissiparous activity'. 81 Rather, the sub-state nature of these relations contradicts neo-realism's basic assumptions. In fact, a theoretical engagement with the literature on interorganizational collaboration within management, administration and organizational studies may offer a more fruitful way ahead, as intelligence agencies' relations can, on the institutional level, be indeed compared to those between firms or non-governmental organizations. 82 While the quantitative methods and large datasets used in these fields are generally not available to intelligence researchers, 83 there exist examples of smaller-scale, qualitative work. 84 For example, in a 2003 article, Hardy et al. carefully examine the collaborations of a single NGO to understand what characteristics determine a collaboration's effect. 85 Engaging these authors' broader concerns about how collaborations shape an organization's power within a network, and how knowledge and resources flow via such networks from an intelligence perspective, may yield useful research questions that could realistically be investigated within intelligence studies' more limited empirical access.

Knowledge circulation as an alternative analytical framework
Knowledge circulation as a concept was developed with the aim of better understanding why certain knowledge spreads quickly to become part of an internationally accepted stock of knowledge while other knowledge does not. In an influential 2004 essay, the historian of science James Secord argued that his discipline's main concern with the historical moments at which new knowledge emerged was insufficient to understand the contemporary epistemological landscape. 86 Instead, scholars should 'give interaction between agents a central role in epistemology' 87 and re-conceptualize knowledge-making as communicative action, rather than isolated discoveries. The reason for this was that an analysis of the conditions and practices that allowed certain knowledge to travel and be received by different audiences around the world was crucial to explaining today's knowledge landscape. Knowledge, argued Secord, was always in transit, and "an explanation of 'what' is being said can only be answered through a simultaneous understanding of 'how', "where', "when" and "for whom"'. 88 Nicolas Guilhot, who has used knowledge circulation theory to explain the origins of IR knowledge, emphasizes the material embeddedness of travelling knowledge, which circulates via people and as part of material transactions. 89 Following this thought, Wiebke Keim writes that 'in the end, ideas . . . circulate . . . in the form of books, articles, data sets and documents, or through persons, who need visa and flight tickets: their circulation costs money'. 90 Knowledge circulation, thus, focuses firstly on the ways and moments in which knowledge within a specific field is communicated and secondly, on how the arrival and negotiation of new knowledge changes both the communicator and recipient. Less focus is placed on the question of why knowledge was communicated in the first place. Since the mid-2000s, historians of Asia, Africa and South America have used knowledge circulation to highlight globally entangled systems of knowledge production. 91 Applying a 'knowledge circulation' lens emphasizes the role played by translation and inter-mediation for the development of only apparently national and homogeneous landscapes of expertise and power. This has helped to illuminate the veiled silences around, for example, the impact of non-Western knowledge and experts in the development of the modern sciences and has brought to the fore the diverse media via which knowledge travels and the role that certain social formations, such as cities, but also plantations or industries, play in the spreading and reforming of knowledge. 92 Knowledge circulation theory differentiates between three modes of circulation. 93 Firstly, reception, where actors simply absorb existing knowledge from elsewhere; secondly, exchange, where actors co-create new knowledge via mutual communication; and thirdly, negotiation, where actors from different institutional settings and different epistemological horizons discuss and debate knowledge. Scholars have theorized each of these steps in detail, a close reading of which holds promise for intelligence scholarship. For example, the assumption that reception is most likely to take place between foreign actors in similar, structural positions, 94 speaks to Herman's proposition that in intelligence relations 'like talks to like, whether Huminter, Siginter, counterespionage specialist, defence intelligencer, national assessor or whatever', as a result of which, 'centrifugal tendencies within national intelligence communities are reinforced', and 'investments in national intelligence organizations are, in some degree, subscriptions to the variety of specialist international clubs to which these organizations belong'. 95 Similarly, insights about the workings of negotiation, where experts communicating with the wider public 'need to take responsibility for the results they feed into public debates, as these may serve as orientation for social action', and where the responses of non-scientists serve as a 'reality-check', may provide useful models for understanding the communication process between intelligence professionals and their audience in the executive or the diplomatic corps (the difficulties of which have been much described and analysed). 96 In the following part of the article I provide three examples that highlight why knowledge circulation provides a strong methodology to kick off a new research agenda to empirically map the international landscape of intelligence relations and theorize by what processes this landscape is reproduced. By moving beyond strategic interest as the only driver of intelligence relations and by creating theoretical openness about why agencies engage in foreign relations, knowledge circulation also offers methodological flexibility to encompass intelligence as a global, rather than a Western or Anglo-Saxon, phenomenon.

Knowledge relations beyond strategic information
The cooperation literature conceptualizes knowledge too narrowly: knowledge transactions comprise specific, strategic or operational knowledge, required to meet circumscribed security goals. Knowledge circulation, on the other hand, widens the scholarly perspective to the broad body of knowledge that is set in motion when intelligence agencies liaise. Knowledge circulation's lack of a premise about why intelligence agencies contact each other here becomes a strength: it allows for an unbiased gaze onto all forms of knowledge that are received, exchanged and negotiated during liaison. Specifically, I mean here implicit knowledge related to threat and risk perceptions, organizational and bureaucratic procedures, standards of secrecy and security, but also social and cultural aspects like knowledge about each other's 'ways of doing things', ideas and biases, language, information about each other's leaders -any type of internationally traveling knowledge around which a theory about the political impact of international intelligence relations may be built. Research on the biographies of intelligence leaders 97 and influential agents, 98 intelligence culture, 99 the transformation of (post) colonial intelligence, 100 and comparisons of intelligence agencies, 101 highlights a broad, diffuse and indeed constant exchange of international knowledge between agencies, but intelligence theory has not captured this. Focusing on this implicit, diffuse knowledge may offer research designs that can address the enduring puzzle of why modern intelligence agencies the world over are organized in a broadly similar institutional manner, independent of regime type. 102 Should we perhaps not speak of an international intelligence system, rather than of disconnected, discrete agencies? Importantly, knowledge circulation does not replace or denigrate strategic interest as an explanation for intelligence relations, but simply 'shines the light elsewhere', onto those elements of travelling intelligence knowledge that have yet to be taken into account. 103

Levels of analysis
Knowledge circulation offers a more robust framework to clearly distinguish between intelligence knowledge that travels 1) via the level of the state, e.g., in the case of formal intelligence agreements, official intelligence residencies, the diplomatic corps, SIGINT stations abroad and so on, 2) via the level of the institution, e.g., communication that takes place between offices (regardless of who staffs them), institutional communiques (such as letters or telegrams issued within intelligence alliances), agency meetings or staff exchanges, and 3) the level of the individual, e.g., individual staff members but also agents, informers and even whistleblowers, whose actions may lead to a transmission of formerly secret knowledge between agencies, such as in the case of the Snowden revelations and their impact on the US-German intelligence relationship. Here, knowledge circulation can make explicit the different levels of intelligence activity that together shape the success, the content and form of an intelligence relationship, and aid researchers by clarifying whether to analyze all, or some of these analytical levels. Making these analytical choices explicit would present an important improvement over the currently implicit and unstructured theorizing about where and how intelligence relations take place.

Materiality of international intelligence relations
Knowledge circulation's emphasis on the materiality of knowledge transmission is one of the theory's great strengths. Knowledge does not circulate in a vacuum but is connected to the circulation of material ('stuff') and people. What does this mean for research? It means, to an extent, to 'follow the material' or to 'follow the people', rather than merely 'follow the information', in the search for explanations about intelligence relations. 104 For example, current scholarship acknowledges that wealthy countries' ability to offer money or aid in exchange for intelligence gives them an advantage, but fails to consider how other elements of a national economy influence intelligence relations. While during the Cold War, political opposition hampered West German government aid to violent regimes, the existence of private German technology companies such as Siemens, Rohde & Schwartz, Daimler Benz, and Messerschmid-Bölkow-Blohm boosted Germany's intelligence relationships, as agencies worldwide sought to buy their products. 105 West Germany's foreign intelligence agency BND's influence on which countries could receive German security technology raised the BND's attraction to foreign agencies, who sometimes even approached the BND directly with requests. 106 In this manner, national resources beyond financial clout may also shape smaller intelligence players' positions within the international intelligence landscape. Following the people on the other hand, means conceptualizing the role of individuals for intelligence relations far beyond the issue of personal trust, the single factor relating to individuals referred to in the cooperation literature. Instead, the importance of the biographies, interests, ideas and activities of key individuals shaping certain intelligence relations needs to be systematically acknowledged. Such individuals may include intelligence officers, but also individuals outside of the formal intelligence apparatus, such as the recurring figure of the 'intelligence-entrepreneur', who successfully trades intelligence across and between international agencies. More than anything, thus analyzing the material embeddedness of travelling intelligence knowledge could provide useful parameters along which intelligence relations could be distinguished and categorised, not least via the development of a political economy of transnational intelligence relations.
Beyond these three factors -a broader conceptualization of knowledge, a better disaggregation of levels of analysis, and a focus on materiality -knowledge circulation offers additional, methodological advantages to help push intelligence scholarship onto a much stronger analytical platform. These include a reflexive epistemology in line with intelligence scholarship's general commitment to qualitative research and an ontology with room for both positivist and constructivist research designs. Via knowledge circulation, intelligence's transnational entanglements move to the forefront, placing intelligence scholarship in an interdisciplinary conversation with current research trends.
To conclude this primarily theoretical contribution, in the following section I offer a brief, empirical example of how knowledge circulation could be applied to a concrete research puzzle. As a full case study is beyond the scope of this article, this exercise is limited, but nevertheless indicates how to move beyond a strictly national view of intelligence and highlights how post-war German and US intelligence knowledge may have emerged in tandem, in a situation of perpetual exchange and circulation.

US -West German intelligence relations: from cooperation to circulation
Post-War, West German intelligence has received some attention from historians and very little attention from political scientists. All accounts agree on the foundational role of US-West German intelligence relations for the development of West German intelligence. 107 Scholars concur that US interest in obtaining strategic security assets in the early years of the Cold War explains the massive US financial support for a relatively motley crew of former Wehrmacht spies, who offered themselves to the US army directly after the war. 108 The few political science analyses on US-West German intelligence relations argue that the US sought to develop West Germany as a dependent intelligence partner, to ensure access to strategic intelligence and share the burden of fighting a common enemy, the Soviet Union. 109 This narrative is fully in line with the cooperation framework, as it primarily focuses on why the US engaged in such close support for West German intelligence and identifies strategic interest as an explanation. How may research via the knowledge circulation framework reframe this work? Wolfgang Krieger, one of the foremost BND historians, argues that in 1945, 'the United States military had no specific agenda when it first took an interest in the remnants of German military intelligence', 110 and continues that 'sources and methods were on their minds when they got their hands on a printing press with Russian type, several hundred original pay books, and a pile of forged Soviet army documents and stamps -highly useful items for any intelligence operations against the Soviets'. 111 Instead of obtaining strategic information, and in an underappreciated historical twist, the US security establishment was firstly interested to learn tradecraft from the relatively wellestablished German foreign intelligence, whose roots stretch back to the 19 th century German empire: 'Amongst the first studies they commissioned was a history of German counterintelligence methods in dealing with the Soviets'. 112 How was such information absorbed into the then developing US foreign intelligence apparatus, and what impact did it have on the design of US intelligence and counter-intelligence? A further, much discussed but little understood, aspect of US-West German intelligence relations is that US intelligence obtained a large trove of knowledge about the Nazi-past of many post-war West German intelligence agents, including senior officers. While in other areas of occupation policy, the US acted on such knowledge to prosecute war criminals, in the area of intelligence cooperation it remained silent: not even the German government, nor public, were aware until decades later, when CIA files were made public and the German government commissioned an investigation of the BND's past, of how many violent Nazi criminals found shelter in West Germany's new intelligence establishment. 113 Taking this secret knowledge about BND officers' Nazi past, which circulated between US and German intelligence, seriously, sheds light on the important theoretical question of whether shared political values influence intelligence relations. Whereas the immediately apparent answer may be that US and German intelligence cared little about their ideological differences, other areas of the relationship indicate that, rather, the relationship included, perhaps even fomented, important ideological overlaps between the radical anticommunism of former Nazi officers and senior US intelligence leaders of the 1940s, 1950s and 1960s. In how far did circulating knowledge related to common 'images of the [Soviet] enemy' 114 and threat perceptions constitute US-West German intelligence sharing and shape mutual priorities? The fact that the BND's first chief, Reinhard Gehlen, received personal coaching from Allen Dulles, then director of the CIA, on a range of intelligence and administrative questions, and exchanged personal letters with J. Edgar Hoover, then director of the FBI, 115 points towards opportunities to develop compatible views among the top leadership. Contrary to the BND's mandate, Gehlen, just like Hoover, set up an infamous set of secret files with damaging information on leading German politicians, 116 and, until the late 1960s, the BND pursued an intensive, secret campaign against German citizens suspected of communist activities. 117 This campaign included the secret pursuit, over many years, of an ultimately imaginary, secret communist espionage organization referred to as the "red chapel", which matches similarly harebrained and illegal domestic anti-communist activities of the CIA, such as the infiltration of the National Student Association, at precisely the same time. 118 Did such similarities emerge due to structural factors, or did circulating knowledge between friendly agencies play a role? These observations support Bigo's argument that intelligence agencies' foreign relations may create a transnational architecture of shared secret knowledge, which serves as a structural influence on how intelligence works in national but connected contexts. It is this architecture, which currently remains almost entirely off the radar of IR scholarship, that the 'knowledge circulation' research agenda may make visible.

Conclusion
Modern intelligence agencies did not develop in isolated, national vacuums, but out of transnational exchange and via the international experience of their founders and staff. In this article I have argued for taking these transnational origins of modern intelligence as a basis for developing new analytical models to understand how international intelligence relations function, and what effect they may have on the content and form of intelligence itself, and on the domestic and international politics they intersect with. This approach can bring to the fore the role that transnationally constituted intelligenceknowledge plays for the constitution of national agencies, and thus, relatedly, for international affairs, on which these national agencies impinge. Proceeding from the assumption that intelligence agencies operate from a basis of transnationally constituted knowledge rather than conceptualizing them as mutually suspicious, nationally-bounded actors, allows intelligence scholarship to re-focus on the vast, historical material documenting events, people, practices and artifacts via which knowledge travels between agencies. Crucially, knowledge circulation does not mean that a focus on strategic interests must fall by the wayside. However, it becomes a question for research; on the basis of which circulating knowledge did agencies develop their strategic outlook in the first place? Was it a briefing from the executive? Or was it a piece of information obtained from an allied (or hostile) intelligence agency abroad? Knowledge circulation allows for a thorough investigation of whether and how intelligence agencies' interests and ideas about national security and threats to it emerge in concert with other foreign intelligence agencies, rather than in a strictly national arena. Most importantly, perhaps, it offers an open-minded research paradigm via which different intelligence contexts and cultures can be approached outside of the Anglo-Saxon and Cold War bias that currently bedevils intelligence scholarship. As more and more empirical material about intelligence in the Global South emerges, knowledge circulation may offer an apt theoretical framework for its analysis.