1

Issues in the theorisation of intelligence Paper presented to an Intelligence Studies Panel, International Studies Association Conference, Montreal, March 2004 Peter Gill, Reader in Politics and Security, Liverpool John Moores University, UK p.gill@livjm.ac.uk, and Mark Phythian, Professor of Politics and History, University of Wolverhampton, UK m.phythian@wlv.ac.uk

Introduction This paper discusses the role of theory in intelligence studies. It does so by reviewing central issues within the more general body of social and political theory that we must confront in order to enhance our ability to understand and explain intelligence processes including their role in contemporary governance. For present purposes we define intelligence as the umbrella term referring to the range of activities from targeting through information gathering to analysis and dissemination that are conducted in secret and aimed at maintaining or enhancing security by providing forewarning of threats or potential threats in a manner that allows for the timely implementation of a preventive policy or strategy. First, we need to ask whether we need to be concerned with theory in studying intelligence. This question may sound odd, especially to those considering themselves social scientists, because of the indispensable role of theory in the generation and organisation of knowledge. But, for a field of studies in which historical accounts have always constituted a significant portion of the literature, it does need addressing. The memoirs of former intelligence officers and, increasingly, the reconstruction of past episodes from released official files is intended to fill the gaps in the hitherto missing dimension of historical studies.1 Of course, some implicit theorising lies behind this but the self-conscious development of theory has not been an aim. Thus, if the work has made any contribution to theory it is via induction: the accumulation of examples of the behaviour of intelligence personnel within particular organisational settings such that (meso level) generalisations might be constructed. More general
1

Andrew C. and Dilks D, The Missing Dimension: Governments and Intelligence Communities in the Twentieth Century, London: Macmillan, 1984; Hoare O. (ed.) British Intelligence in the Twentieth Century: a missing dimension? Special Issue of Intelligence and National Security 17(1), 2002.

2 or macro - theoretical assumptions behind this work tend to be those of the international relations school of realism so that the great game was played out between states, where threats could be objectively measured and the truth of what happened could be discovered by the accumulation of oral and written evidence. There is an argument that an emphasis or focus on intelligence inevitably supports Realist/Neo-Realist analyses because the core concern of them all is security, notwithstanding national and regional variations in the ways in which security is defined.2 In the US ostensibly greater efforts have been made to theorise intelligence at the organisational level, especially the relation between intelligence, secrecy and policy including covert action, but they are often too enmeshed in the debates around the organisation of the US intelligence community to move towards more generalisable findings.3 David Kahn recently set out principles that, he argues, a theory of intelligence should offer in explaining the relation between intelligence and policy but is, for our purposes, too narrowly focused on the issue of states and war though modification to take greater account of the new significance of non-state actors would make it stronger.4 We would suggest that theory is essential if intelligence studies is to move beyond story-telling. Second, what kind of theory is most likely to be productive? The mainstream within Anglo-American social science since the 1950s has been behaviouralism with the following significant features: that general law-like statements can be generated inductively from empirical research and observation of the international system; that there are regularities over time in political behaviour; that, as in the natural sciences, there is no dichotomy between appearance and reality; and that neutral value-free research is possible. Knowledge claims are then based on a cycle in which generalisations are tested by subsequent research and confirmed, modified or abandoned in the light of the findings.5 This approach has its roots in positivism that

For a recent strong re-statement of the neo-realist case and the primacy of security in an anarchic world system see Mearsheimer J., The Tragedy of Great Power Politics, New York: Norton, 2002. 3 For example, Hulnick A.S., Fixing the Spy Machine, Westport: Praeger, 1999; Lowenthal M.M., Intelligence: from secrets to policy, 2nd. Edn., Washington DC: CQ Press, 2003; Odom W.E., Fixing Intelligence, New Haven: Yale UP, 2003. 4 Kahn D., An Historical Theory of Intelligence, Intelligence and National Security 16(3) 2001, 79-92. 5 Hay C., Political Analysis, 2002, 10-13; Sanders D., Behaviouralism, in Marsh D. and Stoker G., Theory and Methods in Political Science, 2nd edn., Basingstoke: Palgrave Macmillan, 2002, 45-64.

3 is based on the foundationalist ontology, that is, the real world exists independently of our knowledge, our knowledge of this world is developed by observation and the aim of social science is to generate explanations of what is, not to be concerned with philosophical or normative questions of what ought to be. Positivism incorporates a powerful preference for quantitative research methods and the ultimate goal is prediction.6 In practice, the record of prediction is poor.

The critique of behaviouralism Positivism and behaviouralism have been subject to numerous criticisms. Some can be reiterated a fortiori with respect to studying intelligence. For example, the positivist proposition that there are facts - a truth - to be discovered independent of the values and theoretical assumptions of the analyst must be counter posed by the proposition that knowledge is socially constructed, in other words, that theory has a crucial role in determining what are the facts. Since theory cannot be developed simply by accumulating observations of reality but itself plays a part in determining what are facts, an approach relying solely on induction cannot suffice. To embark on research about the intelligence process (or analysis within it) without some prior conceptual framework, model or theory is to invite death by drowning or at least mental collapse in a deep sea of information.7 Of course, this fate is routinely avoided but only because of the implicit frameworks we employ. For the study of intelligence to become systematic, we must be more explicit about our frameworks; to do otherwise is at best nave or, worse, dishonest if our work is based on the presentation as scientific of assumptions that are actually highly value-laden and contestable. To pursue the watery metaphor, as analysts we must remember Graham Allisons crucial point that we select the ponds in which we fish8.

Cf. Marsh D. and Furlong P., A Skin not a sweater: ontology and epistemology in political science, in Marsh and Stoker, 17-41; Bottoms A., The Relationship between Theory and Research in Criminology, in King R.D. and Wincup E., Doing Research on Crime and Justice, Oxford: Oxford UP, 2000, 15-60. 7 Sheila Kerr noted that Roger Hilsmans 1958 critique of US strategic intelligence found that contemporary US doctrine put too much emphasis on the accumulation of descriptive facts as the essence of knowledge Turning Knowledge into Wisdom: British and American approaches, Paper given to Intelligence Studies Panel, International Studies Association, Portland, Oregon, March, 2003, p.9. 8 Allison G., Essence of Decision, 1971; Allison & Zelikow P., 2nd edn., New York, Longman, 1999, p.4.

4 Another very specific reason for the inadequacy of behaviouralism as a theoretical approach to intelligence studies is its concentration on observability as a criterion for evidence and the search for actors who cause events. The former is a problem in political analysis generally but it would be entirely disabling if the observability that is normally absent from intelligence were to be defined as an indispensable element for a theory of intelligence! The intelligence literature is replete with accounts of individuals who apparently had great impact on events, either in the formation and operation of agencies or as agents working for some agency (and sometimes for several). Such historical accounts are a rich source of material but, of themselves, can provide only part of the basis for more general statements about intelligence processes. Any attempt to devise a grand, all-encompassing theory of intelligence is doomed if we can only theorise on the basis of that which we can observe. With intelligence, there is too much unseen or at least we assume there is, a further handicap is that we do not know just how great this amount is for us to be able to confidently theorise in a way that behaviouralists would regard as methodologically credible. During the first UK inquiry into the intelligence morass of Iraq that held by Lord Justice Scott between 1992-96 into arms sales to Iraq a British civil servant sought to defend the practice of proving answers to Parliamentary questions that were less than complete: Of course half a picture can be accurate.9 Though criticised in the eventual report,10 this is so the problem for the student of intelligence is knowing the extent to which the partial pictures we develop are accurate. The critique of positivism has been informed by a range of what have been described as post-positivist approaches.11 Some of these are diametrically opposed to positivism: at the ontological level interpretist approaches are anti-foundational, denying that the world exists independently and therefore that our knowledge of it is entirely socially constructed. Therefore, research cannot be objective or value-

Norton-Taylor R., Truth is a Difficult Concept: inside the Scott Inquiry, London: Fourth Estate, 1995, 86 10 The misleading of Parliament by Governments is detailed in Scott R., Report of the Inquiry into the Export of Defence Equipment and Dual-Use Goods to Iraq and Related Prosecutions, 115, Volume 1, Part 2, chapter 4, London: HMSO, 1996, 11 Terriff et al Security Studies Today, 1999, 99-114; Morrow R.A. with Brown D.D., Critical Ttheory and Methodology London: Sage, 1994, 62-82; Marsh and Furlong, 2002, 25-26.

5 free since it is theoretically-laden and is directed, not at scientific explanation but at how our understanding and interpretation affect outcomes.12 One major strand of interpretist thinking is feminism though this follows a number of different routes, for example, liberal, radical, Marxist. Although some feminists share a positivist approach by arguing that certain aspects of the nature and experience of women are universally true, others point to the different experiences of women as mediated by class, ethnicity, culture and sexuality.13 There has been little if any direct feminist critique of the dominant ways of studying intelligence per se but more of the identification of security with that of states found in traditional international relations.14 The connection between states and sovereignty backed up by the states claim to the internal monopoly of legitimate violence and readiness to use force externally in pursuit of its interests has led to national or state security becoming the central analytical concern.15 The steady entrenchment of patriarchal norms through the process of state formation marginalises the experience of many women for whom (in)security means something quite different: A more comprehensive view of security, which begins by asking what, or who, most threatens particular groups of people, will disrupt any notion of national security, for the greatest threats to peoples security in many cases are local state agents or military personnel, or home men who are constructed as soldier-protectors of the very people they endanger.16 To the extent that most intelligence literature has spun-off from IRs concern with state or national security, then it is equally vulnerable to this critique. But if intelligence has been the missing dimension from much IR history, have women been missing from intelligence writing? In some respects, yes. For example, Helen Laville has noted that the historical discussion of the furore that followed the Ramparts exposure of CIA funding of US political groups ignored the funding of the womens Committee of Correspondence between 1953 and 1967.17 But perhaps

12 13

Marsh and Furlong, 2002, 26-27. E.g. Randall V., Feminism, in Marsh and Stoker, 2002, 109-130. 14 A partial exception is Enloe C., Bananas, Beaches and Bases: Making Feminist Sense of International Politics, University of California Press, 1990. 15 Pettman J.J., Worlding Women: a feminist international politics, London: Routledge, 1996, 4. 16 Pettman, Worlding Women, 105. 17 Laville H., The Committee of Correspondence: CIA funding of Womens Groups 19521967, Intelligence and National Security 12(1), 1997, 104-21.

6 agencies have been less short sighted to the role of women in intelligence than have historians! As might be predicted, the role of women in intelligence has mirrored their broader involvement in politics and government. For example, where intelligence is bureaucratised as part of the state, then women have been underrepresented and have been concentrated in clerical and administrative tasks,18 though some graduated from secretarial to operational duties.19 In other contexts, the role of women has been entirely operational, for example, in the peoples war against the US the fact that women were the primary workers in market places made their role as information gatherers and couriers crucial to the Viet Cong.20 In Western states, the exception to the relative invisibility of women in intelligence has been their representation as sexual bait to be deployed in a honey-trap. Certainly these were used, for example, by the East German Stasi21 it was reported that a dozen US embassy staff in Moscow in the 1950s were seduced by intelligence swallows in an attempt to recruit them via blackmail,22 but the real extent of their use is hard to gauge when the profile of the female spy as a beautiful seductress resounded through popular culture in films and fiction.23 However, the most radical departure from positivism is represented by postmodernism. Since intelligence is about the production of knowledge, with agencies operating at the cutting edge of deploying new ICT, it seems entirely appropriate to explain it with an approach to social theory that itself emphasises the significance of new information technologies in re-shaping and subverting modernist methods of generating knowledge. But postmodernism goes much further in its radical epistemological claim that there is no single rationality by which knowledge can be generated; there are no means of establishing truth that transcend the location of the observer and so there can be only competing discourses or ways of representing or narrating events. Der Derian provides an argument for this
18

Gill P., Policing Politics: security intelligence and the liberal democratic state, London: Cass, 1994, 232-34. This was equally the case during the 1980s in the foreign intelligence directorate of the KGB Andrew C. and Gordievsky O., KGB: the inside story, New York: Harper Perennial, 1991, 616. 19 For example, Miller J., One Girls War: personal exploits in MI5s most secret station, Dingle, Co. Derry: Brandon Books, 1986. 20 Taylor S.C., Long-haired women, short-haired spies: gender, espionage, and America war in Vietnam, Intelligence and National Security 13(2), 1998, 61-70. 21 Wolf M. and McElvoy A., Memoirs of a Spymaster: the man who waged a secret war against the West, Pimlico, 1998. 22 Andrew and Gordievsky, KGB, 452. 23 For example, see van Seters D., The Munsinger Affair: images of espionage and security in 1960s Canada, Intelligence and National Security 13(2), 1998, 71-84.

7 approach: noting the relative under-theorisation of intelligence and arguing that what there is is too positivistic in its attempt to discipline global disorder. Rather, he suggests, what is needed is a meta-theory that would take into account the fact that ambiguous discourse, not objective truth, is the fluctuating currency of intelligence. The indeterminacy of what is seen or heard, aggravated by encoding, decoding and possibly by deception plus the gulf between what is said and what is meant requires an approach rooted more in rhetoric than reason. This approach intertextualism aptly covers the field of intelligence, where there is no final arbiter of truth, meaning is derived from an interrelationship of texts and power is implicated by the contingent nature and ambiguity of language and other signifying practices.24 Further, the texts to be analysed are not just the factive ones of national security studies but also the fictive literature of international intrigue that produce meaning and legitimate particular forms of power in their relation to each other.25 Andrew Rathmell identifies five core postmodern themes: the end of grand narratives, the end of the search for absolute truths, the fracturing of identities, increased fluidity of boundaries between approaches to explanation and states, cultures etc., and the emergence of the knowledge economy with its implications for organisation shifts from hierarchies to networks.26 He goes on to relate these to contemporary shifts in the intelligence business: first, the collapse of the grand narrative of the Soviet threat and its replacement by a variety of rapidly-changing and loosely organised targets such as transnational organised crime, proliferation and terrorism. Second, unlike the objective reality of the Cold War, intelligence now does not even know if there is a single objective reality out there. Third, identity is challenged, technologically by the development of human/machine systems, organisationally by doubts as to for whom intelligence is being produced is it states, corporations, non-governmental organisations? and by the collapse of old certainties about who is on whose side and with respect to what. Fourth, the previously impermeable boundaries within the intelligence community and between it and outside (state and private) organisations are dissolving in the face of the decline in the
24 25

Der Derian J., Antidiplomacy: spies, terror, speed, and war, Oxford:Blackwell, 1992, 27. Ibid. 46 26 Rathmell A., Towards Postmodern Intelligence, Intelligence and National Security 17(3) 2002, 95-96.

8 states monopoly of expertise and consequently greater interchanges of personnel and technologies. Fifth, the knowledge economy has hastened the demise of factories for the production of intelligence just as much as it has elsewhere in commerce and government.27 Rathmell suggests that the key implications of this for intelligence organisations are that they should acknowledge their essential nature as a knowledge industry with the concomitant delayering, embrace horizontal knowledge networks and accept change as dynamic, non-linear and accelerating.28 Now, there is much here that can be agreed upon. Positivist methods cannot possibly capture the uncertainties, complexities and ambiguities of the world and this applies to the work of intelligence as much as to those studying intelligence. Both groups do seek to derive meaning from comparing texts, for example, the Hutton Inquiry in the UK spent much energy on just this, but the object of the exercise is not necessarily to discover truth. Intelligence analysts seek knowledge with a degree of certainty sufficient to satisfy and inform those who wish to act upon it; academics are not seeking truth but knowledge with a degree of reliability that will satisfy peer reviewers and standards of intersubjectivity.29 Thus while a central part of the postmodern critique is that there are no objective truths towards which the social analyst can seek to progress, it is not clear that the work of intelligence (or its scholars) usually thought there were. Indeed, one of the reasons for the tension between intelligence professionals and political executives is the very fact that the former do deal in ambiguities and probabilities and recoil from the certainties that the latter wish to be able to deploy in their attempts to persuade mass publics as to the rightness of their cause. But there is something of the counsel of despair in der Derians argument: yes, making sense and explaining the world of intelligence is very difficult for all the reasons he enumerates. And any scholar who claimed to have ascertained the truth of it would be a fool. But it is far from clear that his alternative methodology would

Ibid. 97-98. Ibid. 101. David Lyon puts it stronger in his discussion of postmodernists embracing chaos, Postmodernity, Minneapolis: University of Minnesota Press, 1994, 74-78. 29 Objectivity is not possible in social and behavioural sciences but intersubjectivity is its pragmatic surrogate and ensures that an observation could have been made by any other observer in the same situation. Kaplan A., The Conduct of Inquiry: methodology for behavioral science, New York: Transaction Publishers, 1998, 127-28. Similarly, Bevir and Rhodes argue that the quality of narratives produced by interpretist work can be evaluated by comparisons of comprehensiveness, consistency and heuristic value. Bevir M. and Rhodes R., Interpretive Theory, in Marsh and Stoker, 2002, 142.
28

27

9 improve our understanding. The fictional literature of intelligence may well serve useful purposes (apart from provide enjoyment) when it is a vehicle for the description of events that could not be recounted as fact30 and for its ability to explore and reveal aspects of life as do art and fiction in general. But it may be highly misleading and, taken literally, might lead to the absurdity of Baudrillards claim that the first Gulf War was not real.31 Rather, we would argue, our purpose must be to seek ways of understanding and explaining intelligence, including by way of analysing texts, believing that useful knowledge (that which has some real existence beyond the text) can be ascertained and made use of by those seeking to improve the human condition. Der Derian is right when he reminds us that the piling up of documents does not lead to the truth32 but the doubts, enigmas and complex moral dilemmas that may accumulate can be tackled by the slow, steady development of critical theory and its modification in the light of research. Although postmodernism is not helpful to us, we do need to pay attention to key elements of postmodernity as a description of the social and political conditions at the turn of the Millennium notably postindustrialism and globalisation.33 It is important, however, to consider just what is new here: it seems to us that there are some continuities which caution against any wholesale ditching of modernist methods. For example, the extent to which the threat emanating from the Soviet Union during the Cold War was represented as an objective reality reflected the success of specific organisational and political interests in institutionalising that representation. To the extent that this actually misconstrued what the Soviet Union was seeking to achieve, then it was poor intelligence that can be blamed on a number of factors such as politicisation, mirror-imaging and groupthink. It may well be that, in the more fluid conditions of post- or late-modernity, the intelligence business has become even more complex but it does not necessarily follow that our criteria for judging its effectiveness or otherwise have been transformed, as appears to be the postmodernist claim. Rathmell points to one of the major reasons why intelligence may be slow to embrace the confused realities of post modernity: the fact that agencies
30

Discussing the writing of his latest novel, Absolute Friends, John Le Carr said: you have got to write a novel to tell the truth. Front Row, broadcast BBC Radio 4, January 1, 2004. 31 Lyon, Postmodernity, 48-52. 32 Der Derian, Antidiplomacy, 59. 33 Lyon, Postmodernity, 6-7; Rathmell, Towards Postmodern Intelligence, 93

10 tend still to be bastions of modernist meta-narratives of state power, state sovereignty and national security as well as being formalized modern hierarchies.34 It is somewhat easier to envisage how those charged with producing strategic intelligence will adapt to the chaos of global conditions since so much of the information on which it will be based is available from open sources. The only place to work is within horizontal knowledge networks where there is little substantive difference between analysts and researchers working in intelligence from those in academia or business. However, when it comes to tactical intelligence (for military, security or law enforcement) then the adaptation will be harder. Here, if there is any value-added by state intelligence, it is the information gathered covertly with the aid of special legal powers that are not possessed by business (though that does not mean they will not deploy similar methods without or against the law). Also, not all knowledge networks are open: corporations go to great lengths to protect their intellectual property and to privatise knowledge as can be seen in the major battles over software, pharmaceuticals and the attempt to patent the exploitation of the genome. The decline of state sovereignty may well be a central feature of post modernity but in the area of national security we see the last refuge of the spook. Paradoxically, or perhaps not, as the greater fluidity of global change and relationships seems to demand the recognition by intelligence that it is just another knowledge industry, the concomitant increase in perceptions of insecurity since 9/11 see it acquiring yet more special powers for the greater penetration of privacy and maintenance of secrecy for its operations.35 This tension is going to dominate developments in intelligence for the foreseeable future. In order to generate our own signposts for this future and the tools with which we examine intelligence, we would agree with Rathmell that postmodern perspectives do capture important elements of the contemporary intelligence environment but we do not accept that they provide a satisfactory meta-theory for the explanation of intelligence.

Neither positivist nor postmodernist

34 35

Rathmell, Towards Postmodern Intelligence, 102. For example, Lyon D., Surveillance after September 11, Cambridge: Polity, 2003, 40-55.

11 In seeking to further the task of theorising intelligence, it will be most fruitful to identify a path that avoids the major pitfalls of both positivism and postmodernism. Neither is able to develop knowledge that can be regarded as accurate, leading to understanding, and generalisable beyond the particularities of time and place. We must make our theoretical considerations explicit if for no other reason than intellectual honesty. Value-free social science is impossible because analysts are embedded within the socio-political context that is the subject of their study. Since

analysts cannot claim value-freedom for their findings, they must acknowledge what their value-assumptions are so that their arguments can be evaluated in that context. We return to this in the Conclusion. Analysts cannot claim superiority for their views simply because they occupy a privileged scientific viewpoint from which to observe but they can make their reasoning, methods and sources transparent to others so that the validity of their arguments can be judged. This relates also to the issue of ethics. For positivists this is minimised, if not entirely banished by their insistence that analysts can suspend their own values while engaged in scientific observation; the observer can separate herself from the observed. This is not to say that positivism is inherently unethical but it can lead to this conclusion. If positivism is rejected then the inevitable ethical implications of political analysis must be accepted and analysts must be honest that their work is intended to contribute to making a difference36. If analysts are unhappy with this and insist that their work only describes and/or explains, then they have to remember that their work will contain implicit assumptions and that its publication will have consequences, even if they are unintended and/or simply re-affirm the status quo. Reflexive consideration of the role of intelligence in society is required: intelligence is the handmaiden of public and private political power with potentially great consequences for peoples lives. Theory must be developed in such a way that is does not become a simple apologist for that power. If we presume to tell it like it is without making our assumptions clear, then we may fool some of the people for some of the time but we should not fool ourselves. In order to be useful, our intention must be to generate knowledge and ideas that transcend particular times and places; otherwise we should admit that we
36

cf. Hay, Political Analysis, 86-88 and Barbara Hudson: For critical research, since standpoint is inevitable, it had better be overt. Critical Reflection as Research Methodology, in Jupp V. et al (eds.) Doing Criminological Research London: Sage, 2000, 184

12 are just story-tellers. But we must be realistic and accept that there can be no overarching universal theory of intelligence because of complexity and the inevitability of the social construction of knowledge. What we can strive for is greater shared understanding of the central questions to be asked about intelligence and what items from the conceptual toolkit will be useful and/or appropriate for the analysis of the phenomenon at the various social levels at which it operates. Thus our objective is to establish some principles upon which progress can be made rather than arguing for the necessary superiority of any particular conceptual framework; ultimately choice will be finalised on the basis of the personal beliefs and objectives of the analyst. First, it is recognised that there is some material basis to understanding but how the material is understood is crucial. Thus, there is some reality in the world37 but the process of understanding it requires critical selfreflection on how we understand. Theory and empirical work are inextricably linked and neither deduction nor induction are adequate alone: theory is a guide to empirical exploration, a means of reflecting more or less abstractly upon complex processes of institutional evolution and

transformation in order to highlight key periods or phrases of change which warrant closer empirical scrutiny. Theory sensitises the analyst to the causal processes being elucidated, selecting from the rich complexity of events the underlying mechanisms and processes of change.38 Realism thus seeks to avoid the hobbling effect both of positivism and its antifoundationalist critics. It distinguishes elements of reality that are relatively unchanging and exist independently of the scientific process from those that change more frequently, being produced (socially-constructed) as part of the scientific process.39 Further, with positivism, it believes that causal statements can be made while, against positivism, it accepts that not all social phenomena can be observed and therefore research must seek also for underlying mechanisms of events.40

cf. Richard J. Evans objections to post-modernism in both In Defence of History, London:, Granta Books, 1997 and Lying About Hitler: History, Holocaust, and the David Irving Trial, New York: Basic Books, 2001. 38 Hay, Political Analysis, 47 39 Bhaskar R., Scientific Realism and Human Emancipation, London: Verso, 1986, 51 as cited in Morrow, 1994, 78. 40 Marsh and Furlong, 2002, 30-31.

37

13 One central methodological issue in exploring causal processes is that of agency and structure. The former refers to action or conduct based on the existence of free will so that actors have some degree of choice or autonomy. Note that actors may be individual or collective political analysis contains many references to groups, classes or states as actors. Analyses that emphasise the role of agency (sometimes to the exclusion of any role for structure) are referred to as intentionalist. For example, Dennis Wrong argues that intentionality is the indispensable defining characteristic of political power.41 defined as context, that is: to the fact that political institutions, practices, routines and conventions appear to exhibit some regularity or structure over time.42 Sociologists are more likely to define structure in terms of class, ethnicity, gender, sexuality and these may be deeper structures than institutions and practices but the central point is that they provide the context or arena within which people do or do not act. Again, the object here is not to explore the long-running debate between the proponents of intentionalism and structuralism as alternatives.43 There is nothing essential about the distinction: in some analyses the same political phenomenon, say, a Cabinet, an organisation or a class will constitute the context (structure) within which particular political action occurs but in other situations will be the actor whose impact is studied by the analyst. The (in)actions of any individual actor can only be understood within a range of structural contexts (group, organisational, social etc.) and the constraints that these impose. For example, if our focus is on a specific organisation then we need to understand both its impact as a structure within which people and colleague groups interact, that is, within which bureaucratic politics44 occurs and as an agent acting within the broader context of state and society. In a further refinement to the basic agency/structure relationship, Jessop suggests the incorporation of strategy. Thus agents are reflexive, capable of thinking strategically about their situation and reformulating their interests within the existing structural constraints. These, in turn, are not determined absolutely but operate selectively and
Wrong, Power, ch.1. Hay, Political Analysis, 94 43 Hay, Political Analysis, ch 3 for summary; nb. see also McAnulla S., Structure and Agency, in Marsh D. and Stoker G. (eds.) Theory and Methods in Political Science, Basingstoke: Palgrave Macmillan, 2nd edn., 2002, 271-91. 44 Allison G., Essence of Decision, Boston: Little Brown, 1971 (2nd edn. ?)
42 41

Structure has also been

14 are always specific to the situation of specific agents.45 Thus, certain strategies will be favoured over others the environment is not a level playing field. Outcomes are not structurally determined but, over time, certain outcomes are more likely to occur than others. This strategic-relational approach can be summarised in the following way: the strategic actor makes (explicit or intuitive) choices within the strategically selective context that imposes limits on behaviour. The strategic action that results then produces some direct effect on the structural context (though it may be very slight and not as intended) and provides some potential learning for the actor.46 Thus, the analysis of intelligence can proceed best if agency and structure are viewed as existing in a dialectical relationship. Therefore the task for the analyst of intelligence becomes to develop a way to generalise about how agents understand and are thus influenced by their structural context, on the one hand, and how their strategic actions (or inactions) impact upon that structural context. The context may be either empowering or constraining for agents and their actions may amend or reinforce that structure. For example, in examining the actions of intelligence officers one issue is how they view procedural rules and laws. Whether they view them as empowering or constraining will have some effect on their conduct their original intention to act or view of what is appropriate in a given situation may be reinforced, they may decide to act differently within the rules, to act outside of the rules or not to act at all. In turn, these (in)actions may have (short or long) term implications for

the rules they may be seen as providing useful legal protection for officers or as unwieldy or as so restrictive as to require amendment. So, how to proceed? The social sciences (indeed, sciences in general) have been characterised by increasing specialisation. It is far from clear that this is entirely beneficial (sometimes it feels as though we acquire ever more information about less and less) for it means that much research and analysis focuses on very narrowlydefined issues or problems. We specialise in ways that, not surprisingly, mirror the most obvious divisions in the world we study. So, for example, reflecting the dominance of the nation-state since the seventeenth century, political science is dominated by national studies, sub-national (or, regional) studies, international
45

Jessop B., Interpretive Sociology and the Dialectic of Structure and Agency, Theory, Culture and Society 13(1) 1996, 119-28. This is an equivalent approach to role theory in foreign policy analysis: Kegley and Wittkopf, American Foreign Policy, 5th edn., New York, St. Martins Press, 46465. 46 Hay, Political Analysis, 126-34.

15 relations and comparative politics (as between nations).47 Intelligence studies have

clearly not escaped this dominance. Arguably, for the purposes of seeking a greater generalisability, a more helpful way of characterising the field of interest will be to talk of levels of intelligence. This less institutional and more conceptual technique can already be found, for example, in psychology (individual and social) and economics (micro- and macro-). Specifically, it is suggested that four distinct levels be deployed in order to organise our thinking about the key issues and appropriate analytic strategies.48 The first three can be labelled micro, meso and macro or individual/small group, organisational and societal. The fourth refers to inter-societal as in international relations or economics. These are convenient because they are commonly-used categories that require distinct study. But we must not forget that this is an analytical device; these social levels actually co-exist within great complexity. Yet they enable us to draw up a framework for the analysis of intelligence that alerts us to identify factors or processes that characterise intelligence at each of these levels (thus assisting the goal of generalisability). Before examining the utility of this approach in more detail, however, we should discuss the central concept of surveillance.

Surveillance: knowledge and power It is proposed that we adopt surveillance as the core concept because of its importance in explaining modern governance including the behaviour of agents and development of structures at each of the four levels identified above. Though discussed in different ways by social theorists such as Dandeker, Giddens and Foucault, there is a core of similarity in their definition of surveillance as constituted by two primary components: first, the gathering and storing of information and, second, the supervision of peoples behaviour. In other words, it is concerned with knowledge

and power. In contemporary Western social theory, surveillance is seen both as the central aspect of the establishment of modern sovereign state forms49 and of the
Phil Cerny makes the point that the structural dominance of the nation-state is reflected throughout social science. Cerny P.G., Globalization and the Disarticulation of Political Power: towards a new middle ages? in Goverde H., et al (eds.) Power in Contemporary Politics, London: Sage, 2000, 17-86. 48 This might be compared with the approach in Holsti K.J., International Politics: a framework for analysis, New Jersey: Prentice-Hall, 1967, ch. 1. 49 For example, Giddens A., The Nation State and Violence, 1985, 181-92.
47

16 more recent decline of sovereignty as it is replaced by governance (or, for Foucault, governmentality50) including the concomitant recognition of the significance of private forms of governance. Furthermore, studies of non-Western societies show that surveillance is similarly central there: its philosophical basis may be crucially different, for example, the rejection of individualism, but its core goals understanding and control pertain.51 So, not surprisingly, global surveillance is argued to be an intrinsic part of the general economic restructuring of capitalism that is referred to as globalisation,52 and post-9/11 developments have served only to accelerate this already existing trend.53 We suggest that the security intelligence process with which we are specifically concerned is essentially a sub-set of the more general surveillance that constitutes contemporary governance. The latter is defined in terms of the generation of knowledge for the purposes of supervision and regulation; we defined the former at the outset as the generation of knowledge that can inform the formation and implementation of security policy. States clearly depend as much on internal as external security for their wellbeing but the greatest part of writing on intelligence at the national or macro level has concerned external intelligence within the context of international relations.54 This was established a half century ago when Sherman Kent reflected on the agenda for post-war US strategic intelligence and excluded from his discussion internal or police intelligence.55 More recently we have seen the beginnings of discussions as to the possibility of actually organising intelligence at a transnational level. Initially this was limited to the discussion of transnational sharing agreements56 a much

Foucault M., Governmentality, in Burchell G. et al (eds.) The Foucault Effect: Studies in Governmentality, London: Harvester Wheatsheaf, 1991. 51 Bozeman A., Knowledge and Comparative Method in Comparative Intelligence Studies, in Bozeman, Strategic Intelligence and Statecraft, Washington DC: Brasseys, 1992, 198-205. 52 Lyon D., Surveillance Society, 2001, 103. Cf. also der Derian, Antidiplomacy, 1992, 46. Whitaker R., The End of Privacy: how total surveillance is becoming a reality, New York: The New Press, 1999 is an excellent survey of contemporary developments. 53 Lyon D., Surveillance after September 11, Cambridge: Polity, 2003. 54 Of course, weak states may be compared with strong states in terms of the formers greater concern with internal security. Buzan B., People, States and Fear, 2nd edn., Hemel Hempstead: Harvester Wheatsheaf, 1991, esp. 96-107 55 Kent S., Strategic Intelligence for American World Policy, Princeton: Princeton UP, 1966 (first pub. 1949) p.3. 56 For example, Richelson J.T. and Ball D., The Ties that Bind, 2nd edn., Boston, Unwin Hyman, 1990 on the UKUSA agreement.

50

17 enhanced concern since 9/11 - but now includes also consideration of how intelligence can support multi-lateral peacekeeping.57 Surveillance is equally central to the meso level body of literature that has sought to explain, more specifically, how intelligence works (or not) in organisations. All organisations spend some resources on seeking information about the environment within which they operate, though in many this will not be specifically organised or even referred to as intelligence. Even the most humble organisations will engage in information gathering regarding their strengths, weaknesses, opportunities and threats (SWOT analysis) and will then seek to translate the findings into action aimed at better achieving their goals. But the concern here is with a more specific sub-set of the literature that is concerned with the examination of various INTS - foreign, military, security, criminal, environmental, economic, business and how they contribute to the goal of increasing security. At the micro level of individuals (cognitive psychology) and small groups (social psychology) there are extensive literatures on the processes involved in information gathering, problem solving and decision making. For example, cognitive psychology is characterised by an information processing approach that makes much of the computer metaphor. The means by which information acquired via the senses is transformed into experience is described as perception which, according to some, is direct or bottom-up while for others constructivists it is indirect or top-down; it: occurs as the end-product of the interactive influences of the presented stimulus and internal hypotheses, expectations, and knowledge, as well as motivational and emotional factors.58 Attention is the term used to refer to the selectivity involved in processing; again, this may be top-down when determined by the individuals goals rather than directly in response to the external stimulus. Memory involves both structure and the processes operating within the structure: encoding, storing and retrieval. This knowledge has to be organised or categorised in a way that is both economical yet sufficiently detailed to be informative and may be represented internally or mentally
57

For example, Nomikos J.M. Intelligence Requirements for Peacekeeping Operations, Research Institute for European and American Studies, 2003, www.rieas.gr/papers.html July 7. 58 Eysenck M.W. and Keane M.T., Cognitive Psychology: a students handbook, Hove: Psychology Press, 2000, 4th edn., 54.

18 as well as externally through language, for example, speaking, writing or drawing. The basis for this thinking and reflection is to plan and solve problems: the research and theorising here is typically organised into puzzle solving - creativity and discovery; reasoning both inductive and deductive and judgement and decisionmaking.59 In terms of our overall argument, it is significant to note that, even in this self-proclaimed hardest of the social sciences, processes of selectivity and construction demonstrate the limits of behaviouralism. Thus, for present purposes we might safely assume that the desire for information is ubiquitous among individuals, groups, organisations, states and societies. Faced with uncertainty, risk, feelings of insecurity or in search of some other goal, all human entities face a knowledge problem60 and seek information that (they hope) will reduce uncertainty and insecurity or enable them to address their vulnerabilities and advance their interests. This search is necessarily selective: complete scanning is unrealistic (which is why rational actor theories can exist only as ideal-types see further below) and therefore the criteria guiding the search are crucial. How do agents define problems in order to render them thinkable for the purposes of governance doing something about them?61 We return to this below also. Information gathered can be classified as one of two main types: defensive and offensive. The object of the former is to identify ones own (personal, organisational or societal) vulnerabilities or weaknesses and the threats emanating from either external or internal sources in order to maintain the existing state of affairs or to defend oneself. Risk analysis provides much of the contemporary language used to discuss threats and normally seeks to calculate two main dimensions: the magnitude of the risk and the likelihood of its occurrence. The object of offensive information is to bring about some change in the state of affairs. At the personal

level this may be either intentional or habitual action to bring about some improvement in personal prospects or feelings of security, an organisation might perceive the need to develop new markets (in the private sector) or improve its
Eysenck and Keane, Cognitive Psychology, passim. Mayntz R., Governing Failures and the Problem of Governability in Kooiman J. (ed.) Modern Governance: new government-society interactions, London: Sage, 9-20. 61 Edwards A. and Gill P., After transnational organised crime? The politics of public safety, in Edwards and Gill, Transnational Organised Crime: perspectives on global security, London: Routledge, 2003, 265.
60 59

19 performance (in the public sector) while states may seek information with a view to the maintenance of public order or acquisition of new territory. There are also significant grey areas of intelligence, for example, in the growing area of economic intelligence the efforts of states to gather information on competitors within the context of market economies may be defined as offensive (in terms of increasing market share) or defensive (in terms of domestic jobs). Information gathering takes place in two broad ways: openly through the collation of already published material and relatively passive environmental scanning (overt) or more aggressive techniques of penetrating the secrecy and privacy of others (covert). The first of these may be conducted more or less self-

consciously and the amounts of potentially relevant material available on any issue are prodigious and, thanks to the Internet, very quickly available. Covert information gathering requires the commitment of greater resources. Expertise in the covert gathering of information (or the gathering of information held covertly by others) is significant within intelligence but it is best not to reify covert information gathering since, though it may be crucial, it is not necessarily more useful as a basis for action than what can be obtained from open sources. What is to be done with information gathered? It is common (and useful) to draw a distinction between information and intelligence and the difference is made by a process of analysis or evaluation of what is gathered. Again, this process may be more or less conscious: in the case of

individuals the act of obtaining some information is often accompanied immediately by an intuitive evaluation of its meaning and significance based on the credibility or otherwise of its source and substance. On other occasions a more explicit process is undertaken in which we are forced to confront information from various sources that does not apparently make sense. Even in organisational settings those involved in collecting information will often immediately evaluate its meaning if only to decide whether further information needs to be sought. Thus collection and analysis often occur in a rapid cycle; indeed there is a danger that process and cycle both present an image of linear neatness that rarely occurs in practice.62 However, it remains helpful for research purposes to view stages of the process such as gathering and analysis as distinct activities. Facts do not speak for themselves; if they seem to
62

Arthur Hulnick, for example, suggests that in reality it is best to see the intelligence process as a matrix of interconnected, mostly autonomous functions (Controlling Intelligence Estimates, in Hastedt G. (ed.) Controlling Intelligence, London: Cass, 1991, 84).

20 there is the danger that they will mislead. Analysis is the crucial process through which appearances are subjected to critical scrutiny. Memory plays a key role in this collection/analysis cycle: again, this is a feature that exists at all levels. What is or is not consigned to an individuals conscious memory is itself a highly complex process involving not just the significance of the information but also the will to remember. Evaluation is carried out with respect to, for example, the veracity of sources and the consistency or otherwise of fresh information with what is in the store. Organisational memories are, in part, more formal than those of individuals. One of the characteristics of the bureaucracies that came to signify the emergence of modern state forms was the maintenance of records that could provide the basis for greater consistency of decisions over time. In specialised intelligence agencies the files represent unambiguously the organisational memory that provides the structural context for current analysis. Yet, it must be remembered that organisational memories also have an informal dimension and, to the regular frustration of those who would reform and reorganise bureaucracies, intelligence and otherwise, organisational sub-cultures can persist fuelled by employees memories and preferred ways of working. This collection and analysis process may or may not be followed by action; the crucial connection between knowledge and power (or, intelligence and policy) being provided by the dissemination of intelligence. If the knowledge problem is believed to have been solved, then what can be done about the threat, the vulnerability or the ambition? Can the power problem be solved? Two broad theoretical streams can be identified with respect to power. The first, sometimes referred to as the sovereign view of power is principally concerned with the power that one agent may exercise over another in what can be described as a constant sum relationship - pouvoir in French. The second stream is more concerned with agents ability to facilitate things, with the techniques of power and sees power as more widely dispersed in society rather than being concentrated in sovereign states or individuals. As such, power may facilitate as well as repress and therefore is variable, rather than constant sum - puissance in French63. Again, the point is not that

63

Scott J., Power, Cambridge: Polity, 2001, 6-12. Also Clegg S., Frameworks of Power, Sage,

1989

21 one of these streams is superior to the other; it is that the analyst needs to be aware that power is contingent and may be manifested as either constant or variable sum. Dissemination may take several forms but what they all have in common is the link with power. Even the simple decision to pass intelligence on to another person or organisation will have consequences for power. These may vary from the relatively neutral to the highly interventionist where the intelligence is calculated to have a major impact on the thoughts and/or behaviour of the recipient. Another useful distinction to make is where the intelligence disseminated is believed to be true and thus amounts to a form of expertise that may persuade because it is accepted on trust contrasted with the situation in which the intelligence disseminated is believed to be misleading in which case deception or manipulation are more appropriate terms64. But whether believed to be true or not, intelligence is itself a form of power: knowledge is power. Otherwise, information is not so much a form of power but is better described as a resource which can support the exercise of other forms of power, for example, material or coercive. This is the more common view that intelligence may provide the basis for policy or decisions: that people, organisations, states, if they are to act rationally, will do so after canvassing fully the alternative courses of action open to them and the costs and benefits of each. Of course, as the voluminous decision making literature indicates, rational action is really only useful as an ideal-type against which practice might be measured. Various phrases have been used to describe the messier realities of the relation between information and action: for example, bounded rationality, garbage can model and incrementalism.65 Key to these is selectivity a combination of prior assumptions and limited resources means that we are very rarely willing or able to conduct a full canvass of potentially relevant information. Indeed, we must be aware that, in practice, the knowledgepower connection might actually be reversed so that the will to act pre-exists the search for information and the significance of what is collected will be judged in terms of its ability to support our chosen course of action rather than to inform it. Here, the relation between knowledge and power is like that of lamppost and drunk, to provide support not illumination. There is not the space to develop the argument here
64 65

Cf. Scott, Power, 16-25. Wilsnack, Information Control For a summary of this debate see Hill M., The Policy Process in the Modern State, Hemel Hempstead: Harvester Wheatsheaf, 1997, 3rd edn., 98-126.

22 but there are significant signs that just such a relationship characterised the search for intelligence by US and UK administrations in the lead up the invasion of Iraq in 2003. But whether intelligence is the reason or the rationalisation for policy is not the only question to be asked about the knowledge-power relationship. Even if manifest action is not taken, it does not mean that the knowledge generated has had no impact or that power has not been exercised. For example, there is the law of anticipated reactions identified by Friedrich in which an agent behaves in a particular way anticipating how other participants would react were s/he to behave otherwise.66 More generally, the same idea is at the core of Foucaults suggestion that surveillance breeds self-regulating subjects: the principle of the panopticon is that those surveilled base their behaviour on their understanding that they may be under surveillance. Thus they regulate their own behaviour despite their ignorance as to whether they actually are under surveillance at any particular time. This creates difficulties for analysis since no observable behaviour on the part of those wielding power may be necessary; another illustration that understanding based entirely on empirical work will only ever be partial if not actually misleading. There are two additional variables that permeate the knowledge-power relation. The first is secrecy. Whether or not this is seen as the very essence of security intelligence, it certainly impacts on surveillance. Apart from the sheer complexity of modern society, one reason why knowledge problems exist is the conscious efforts people, organisations and states make to keep their affairs secret. For individuals we are more likely to refer to privacy rights and these are included in human rights declarations and conventions because privacy is seen as indispensable to peoples dignity and autonomy. Corporations seek to have the same privileges extended to themselves via notions of commercial confidentiality that are written into contracts of employment with their staffs and may also present a significant bulwark to the work of outside regulators. At the level of states, current ideas of official secrets developed with modern state bureaucracies but may be traced back further to the notion of arcana in pre-modern religious states.67 Thus, once open sources are exhausted, the privacy of subjects presents an obstacle to surveillance that provokes a whole panoply of covert gathering techniques. But secrecy is not just significant as a
66 67

Friedrich C.J., Man and His Government, New York: McGraw-Hill, 1963, 201-02. Bok S., Secrets: on the ethics of concealment and revelation, Oxford: Oxford UP, 1986, 6-7

23 barrier to surveillance; it permeates aspects of the process itself. Certainly in areas regarding security the need to protect the information gathered, methods and the identity of sources leads to elaborate internal procedures including the restricted circulation of documents on a need-to-know basis. The consequent limitation on the availability of information can, in turn, hinder the free flow of ideas and quality of analysis. Secrecy may also apply to power: some actions make no sense unless carried out with an element of surprise such as arrests. But there are other, more controversial examples where actions are taken secretly in the hope that responsibility can be disguised or plausibly denied. The second variable is resistance. Secrecy is one form of resistance: attempts to maintain personal privacy or business confidentiality are forms of resistance to the efforts of others to collect information. But there are others: if privacy fails then lying and deception are other forms of resistance. Evaluation or analysis is, in turn, an attempt to resist the attempt of others to mislead. Resistance to other forms of power such as coercion may well take on a more physical aspect but often these will be intertwined with the use of information. The central point here is that the relation

between surveillance and its subjects is dialectical: efforts at gathering information and wielding power (in whatever form) will provoke greater or lesser attempts to resist. If resistance succeeds then fresh approaches to surveillance may be deployed and so on.68

Conclusion Any community of scholars and researchers shares certain assumptions about the way in which knowledge claims in their field are generated. As a result of academic specialisation and/or personal taste, some people choose to spend more time concerned with conceptual and theoretical issues while others prefer to get their hands dirty with empirical work. The democratisation of intelligence in many countries over the past quarter century and accompanying avalanche of released files and papers, not to mention the increased tendency of former officers to write their memoirs, has given an extraordinary boost to scholarship in the field but more of this

68

cf. Giddens, The Nation State and Violence, 1985, 10-11 regarding the dialectic of control in social systems. See also Clegg S., Power and Authority, Resistance and Legitimacy, in Goverde H. et al (eds.) Power in Contemporary Politics: theories. practices, globalizations, London: Sage, 77-92.

24 has been based on an urge to provide some historical accounting for the past the missing dimension than to reflect on how we study and write about intelligence. This is entirely understandable in former authoritarian regimes the unearthing of intelligence secrets has been a painful but necessary part of making political progress and even in liberal democracies it has contributed to reforms intended to reduce the likelihood of future abuses of state power. The production of detailed historical accounts is a necessary part of our work but more is needed if we are to develop understandings and explanations of intelligence that transcend particular times and places. First, too often the theoretical assumptions that precede research are implicit and thus disagreements between scholars and practitioners may be generated as much by different assumptions as by different empirical findings. Therefore, we must be explicit about the conceptual frameworks and theoretical assumptions we use. We have argued that the behaviouralist paradigm that has dominated Anglo-American social science for fifty years is inadequate because, if for no other reason, the insistence on observability is entirely disabling to the study of intelligence. Yet we have argued also against switching wholesale to an interpretist approach. This can seem very tempting when trying to make sense of the wilderness of mirrors that is intelligence. Indeed, the dichotomies of appearance and reality are precisely what the intelligence business is about and so any approach to theory must include an important place for the social construction of knowledge. We have emphasised issues of security and secrecy and identify the central significance of the unknown (and possibly unknowable) in analysing intelligence. But to deny that there are any realities independent of our research efforts is to disable ourselves in a different way, leaving us enmeshed in endless debates as to how we might describe the world with no chance of proceeding further. Postmodernism is a cautionary tale and we should heed it but it must not immobilise us. We suggest that understanding and explanation can best be furthered by the self-conscious development of theory that is reflexive and critical. Over time, this can proceed by informing the research agenda and methods deployed; in turn being subject to modification in the light of research findings. The object is not to seek some unattainable universally-applicable explanation of intelligence processes but to produce knowledge that is applicable beyond particular time/space dimensions and

25 that can serve the purposes of all those concerned with intelligence. Whether what we produce counts as knowledge will be determined by the judgements of the community (whether scholars or professionals). Not all will be equally interested in all aspects of the field but if theory is sound then it will be of value to all. For example, we do not accept that there is any necessary division between theory and practice in intelligence; laws, techniques, policies and practices are all informed by theoretical assumptions, better that they be explicit and generally-accepted than implicit and highly debateable once exposed. Different aspects of traditional frameworks may be of use if we are to seek to ensure that intelligence is conducted in proper and efficacious ways. Of course, work of different degrees of abstraction interests people in different positions and with varying interests but this makes it all the more important that there is some core understanding that all (or most) can acknowledge as establishing some basic principles of knowledge upon which more specific studies, reconstructions, policies, and explanations can be constructed. We have suggested that the concept of surveillance, understood as the generation of knowledge for the purposes of exercising power, provides the best place to start because there is already considerable work regarding its significance at all levels of society from the micro to the macro. As the study of intelligence delves ever deeper into its histories and practices, it must also look outwards to take notice of theoretical developments in sister social sciences. Surveillance is concerned with the crucial power-knowledge nexus and thus provides an excellent start point for the more particular questions of concern in intelligence, especially its relation to policy. We are most likely to generate theory that is relevant and interesting if our analyses check constantly between both the vertical social levels at which intelligence processes occur and horizontally between the worlds of intelligence and its social, economic and political contexts. The consequences of the use and abuse of intelligence are clearly profound, even more so now that, post-9/11, security concerns threaten to escalate into paranoia. The last few years have seen an unprecedented level of public controversy regarding intelligence. If we are to make a serious contribution to better explanation, understanding and public education on the key relationship of intelligence to law, politics, security and governance, then our work must be based on an appreciation of

26 central theoretical issues. Then we shall be able to speak truth unto power and not simply find ourselves conscripted as and when power finds it convenient.