Launched in 1970, Europe’s common foreign policy has, to some degree, come of age. Because previous attempts to introduce cooperation in the field of foreign policy had failed, the cooperative enterprise was deliberately launched with very modest ambitions. Its development over the years came to include still more policy areas, and still more modules were added to its institutional infrastructure. When defence policy was included in the provisions of the Treaty on European Union (1993) it became, at least in principle, a tous azimut foreign policy institution. Thus, while observers of developments during the 1970s and 1980s noted a somewhat slow beginning (Wallace 1983a; Nuttall 1992; Nørgaard Pedersen and Petersen 1993), observers of the 1990s have noted that European decision-makers during the 1990s significantly accelerated their involvement in common foreign policy-making (Hill 1996a; Piening 1997; Regelsberger, de Schoutheete de Tervarent and Wessels 1997; Flockhart 1998; Peterson and Sjursen 1998; Bretherton and Vogler 1999; Cameron 1999).

What are the most significant areas to have been addressed by Europe’s common foreign policy? First of all, to European states’ traditional bilateral relations with the US has been added a European–US component, meaning that relations with the US increasingly have been cultivated by means of common European stances. Moreover, the US has largely recognised this change on the European side. Furthermore, US–European joint policies have increasingly been thoroughly coordinated and there have been numerous cases of mutual ‘backing’ (Featherstone and Ginsberg 1996).

Second, the European Union’s relations with what was formerly known as Eastern Europe and the Soviet Union have been thoroughly reconsidered and redesigned. In terms of resources spent on a European Ostpolitik, it clearly ranks among the first priorities of the Union. Policies towards the Soviet Union, and particularly towards Central and Eastern Europe, were initiated in the mid-1980s. These policies were somewhat contradictory during their formative years (Torreblanca Payá 1997) but from the Copenhagen Summit in June 1993 they were transformed into a perspective on enlargement of the European Union. A so-called Stability Pact was designed in order to de-escalate potential conflicts in the Central and Eastern parts of Europe. Furthermore, the Eastern enlargement can be regarded as a huge foreign policy initiative significantly increasing the scope of innumerable European regimes, from the entire acquis communautaire to administrative practices in the state administrations of future members. In reality, they are currently experiencing a process through which they change from being nation-states to being member states of the European Union.

Third, Europe’s relations with developing countries have since the very beginning of the joint European enterprise been a significant feature of Europe’s external relations. Until the June 2000 signature of the Cotonou Agreement, the regularly up-dated Lomé Conventions (until 1975 called the Yaounde Convention) framed EU relations with the developing world for nearly fifty years, with other policy tools adding to the overall picture (Ravenhill 1985; Grilli 1993; Cosgrove-Sacks 1999).

Fourth, a common trade policy has been in place for several decades. It has been tested during various rounds of the General Agreement on Tariffs and Trade (Grieco 1990), including the so-called Uruguay round, which resulted in the World Trade Organisation (WTO). Furthermore, it has been tested within specific trade issues, such as the EU–US banana dispute (Stevens 1996), Commissioner Karel van Miert’s effective intervention in the Boeing–McDonnell Douglas merger, and innumerable other cases since.

Fifth, although Europe’s policy on ex-Yugoslavia came close to being a policy failure, it has, nevertheless, also been a remarkable sign of international ‘actorness’ on the part of the EU, with the Union functioning for three years as the primary external crisis manager (Owen 1995; Jørgensen 1996). This involvement also tells us something about the aspirations of policy-makers concerning the international role of the European Union.

Sixth, Europe has been involved in numerous other geographical and thematic areas, ranging from the EuroMed initiative (Gomez 1998), Central America (Smith 1995) and China (Yahuda 1998), to the well-established so-called group-to-group dialogues with groupings such as the Association of Southeast Asian Nations (ASEAN), the Gulf Cooperation Council and Mercosur (Edwards and Regelsberger 1990); to negotiations on the Non-Proliferation Treaty and policy-making within the UN General Assembly (Brückner 1990), the Organisation for Security and Cooperation in Europe (OSCE) and other international organisations.

This brief overview is merely meant to serve as an appropriate point of departure leading to the following three observations. First, while the overview is necessarily brief it indicates that it is a worthwhile activity to conduct research on EU foreign policy. Furthermore, the processes producing the foreign policy in question are markedly different from traditional state foreign policy-making. They can be perceived as a Janus-face – being both arena and actor. Determining which ‘face’ is predominant depends on the issue-area. Sometimes the EU is the policy arena, sometimes it is an actor, and this seems to change over time. Finally, the relationship adds to the complexity of the matter since in foreign policy-making, the relationship between member states and the European Union is primarily non-hierarchical. Even a casual reading of the practice of collective policy-making reveals some illuminating features.

Communication and argumentation are, for example, essential features of the system. The actors exchange views, argue and aim at reaching a consensus. According to the Danish Foreign Ministry, ‘a small Member State has good possibilities of influencing the decision-making procedure with the right argument. That is what experience shows’ (Interview, Danish Foreign Ministry, March 1998). This suggests that research informed by Habermas’ theory of communicative action is possible, worthwhile and holds the potential of powerful explanations (see Risse 1998; Lose 2001; Müller 2001). Obviously, parts of communicative action consist of pointing out reasons for action, something – as Ruggie (1998a) has pointed out – fundamentally different from causes of action. Intersubjective understandings are what diplomats’ calendars and agendas are full of, whether in the context of procedure (as when the Dutch Presidency in September 1991 was politically punished for transcending unwritten rules or when ‘The Member States . . . progressively developed disciplines, in written and unwritten rules and procedures, with a view to improving their cohesion in the UN through the various modes of expression, in particular joint statements, voting, and common explanations of vote’ (Brückner 1990: 177; see also Jørgensen 1997a), or whether in the context of policy-making.

A prominent example of the latter is Jacques Poos’ famous, or infamous, remark in July 1991 that international crisis management vis-à-vis the break-up of Yugoslavia showed that the ‘hour of Europe’ had come. By this remark he simply expressed a widely held intersubjective understanding among European policy-makers and analysts that the absence of the Cold War almost automatically would elevate the EC into a new and significant actor in international politics and, moreover, an actor with considerable diplomatic and political clout. It was a short-lived understanding but at the time it had a significant impact on policy-making.

Contemporary European foreign policy is full of aspirations. Policy memos with reference to a desired stronger presence in areas ranging from the Far East to the Middle East are legion. In general, key policy statements more often than not present the Union as ‘a strong and credible partner on the international arena’. Furthermore, it is of significant interest that one of the very first EPC declarations was entitled ‘Declaration on European Identity’ (1973), and that the issue of a European identity in international relations has been among the permanent features of European foreign policy ever since. In fact, it is difficult to imagine that it could have been otherwise, and competing notions of European identity have been employed, ranging from the notion of a ‘civilian power Europe’ to a ‘superpower Europe’. This feature makes pre-theories like Alex Wendt’s (1994) about collective identity-formation highly relevant and interesting.

The issue of legitimacy is a second permanent feature of foreign policy, a feature that surely exists in different versions. One version concerns the legitimacy of the common enterprise; that is, whether foreign policy constituencies in member states find the common enterprise legitimate or not. Another issue with regard to legitimacy concerns the practice of European foreign policy; that is, whether – to take an example from current international affairs – military operations without explicit UN authorisation are considered legitimate or not.

A third permanent feature concerns interests. Even if policy-makers have been reluctant in making explicit principled statements about common European interests (see Burghardt 1997: 330), such interests can presumably be deducted from foreign policy practice. When policy-makers seek to exert a pressure for change in global politics they in part do so by thinking on the basis of world views. A Dutch analyst neatly describes one such world-view when he describes the EU as ‘a civilian power in an uncivil world’ (Pijpers 1990: 171). Other views are associated with the image of foreign policy as a civilising mission. Still other views are associated with conceptions of rights or responsibilities. It is thus significant that the German Presidency in 1988 summarised EC Ostpolitik during the autumn of 1988 with the words, ‘We want to address central questions of the Twelve’s regional responsibility for peace. Let us not underestimate the new Europe, with its inner dynamism and its invigorating power’ (European Foreign Policy Bulletin online, Document 88/168) (author’s italics).

Scholars familiar with social constructivist perspectives will recognise from the above outline many of the features well known from the rich toolbox of that approach. In other words, Europe’s foreign policy appears to be an ideal case for showing the potential and limits of social constructivism. The claim here is not that social constructivism, at all times and in all cases, is the ‘better’ option and thus by necessity more powerful than its competitors in offering an understanding of EU foreign policy, but rather that it seems worthwhile to investigate the scope of its potential in the selected case. It should be emphasised that the case is not a common European foreign policy per se but rather the more narrowly conceived CFSP, meaning that the communautarian aspects of foreign policy are omitted. The following section aims at exploring possibilities of theorising the CFSP ‘the constructivist way’. First, I describe how the balance between deductive and inductive theorising is quite asymmetrical. Then I use nine rules for creative theorising – developed by Rosenau and Durfee (1995) – as a point of departure and combine their rules with social constructivist ways of framing research questions and designs. For each rule there is consideration of how theories would have been, had Rosenau and Durfee’s prescriptions been informed by a social constructivist perspective.

Theorising CFSP

Most theory-informed research on the CFSP employs the deductive method and a large number of theories or frameworks of analysis have been applied in research on the CFSP. By now, more or less every imaginable theory has been applied: realism, neo-liberal institutionalism, regime theory, world systems theory, and so on (Pijpers 1990; George 1991; Lucarelli 1998). Unfortunately, these applications have been conducted in a ‘single shot’ fashion and more often than not they focus on case studies of selected policies or institutional aspects. We are thus far from having reached a critical mass of applications and thus are unable to fully assess the potentials of each theory.

Inductive theorising can imply that we begin from scratch; that is, we act as if we were naive empiricists simply looking at data in front of us. Or we can move on by means of selecting aspects of various theories and then combining them in a new framework. Surprisingly, perhaps, very few theoretical studies of the CFSP have been conducted in such an inductive exploratory fashion. One reason for this absence might be that, among foreign policy analysts, the CFSP is widely considered an appendix to national foreign policy and why waste time on theorising an appendix? A second reason is probably that the CFSP is a topic beyond the attention of scholars with an interest in international theory. After all, we are dealing with a regional not a global phenomenon and even in the regional European context, the CFSP is only one aspect. A third reason may be that the CFSP has been primarily analysed by European scholars and, for some reason, they generally theorise less than their North American colleagues. When Europeans employ theories, they primarily do so by means of the deductive method, meaning that they contribute to the art of testing theories developed elsewhere and sometimes reflecting other experiences and often serving other purposes. These reasons may explain why an inductive, exploratory approach has been somewhat neglected.

For the purpose of theorising the CFSP I have, as mentioned above, selected two manuals in theorising (Rosenau and Durfee 1995; van Evera 1997). I use Rosenau and Durfee’s rules for creative theorising and, when appropriate, add some ingredients from van Evera’s recipe for theorising. As I proceed I include theoretical reflections from constructivist perspectives and consider which issues could be added to the current research agenda.

Rule 1. Avoid treating the task as that of formulating an appropriate definition of theory (Rosenau and Durfee 1995: 129–30). Certainly, they have a good point here. Many attempts at theorising have stalled at the very beginning because they turn into a definitional exercise, in that the act of definition becomes a hindrance to developing abstract knowledge about world politics. Rosenau and Durfee acknowledge that it is preferable to have a precise conception of theory rather than a vague one and they emphasise features of the demanding art of theorising, an art they associate with ‘a set of predispositions, a clutch of habits, a way of thinking, a mental lifestyle, a cast of mind, a mental set’ (pp. 178, 180). While I do not know how many scholars can live up to these requirements, I do think it is a well-intended rule, and to some degree also a helpful one. It is, nonetheless, important to note that they consistently refer to theory in the singular, as if the category ‘theory’ does not include different types of theory. Such a reduction is most unsatisfactory and, as it turns out, Rosenau and Durfee are soon forced to point out that the rules they formulate only apply to what they call ‘empirical theory’. While I cannot but disagree concerning the applicability of their rules, their delineation leads to the second rule.

Rule 2. One has to be clear as to whether one aspires to empirical theory or value theory (Rosenau and Durfee 1995: 120–2). Their reasoning is built on the classic distinction between ‘is’, leading to empirical theory, and ‘ought’, leading to value theory. As they explain, this distinction underlies entirely different modes of reasoning, a different rhetoric and different types of evidence, leading them to a significant second delineation: ‘Aware that our own expertise, such as it may be, lies in the realm of empirical theory ensuring discussion makes no pretence of being relevant to thinking theoretically in a moral context’ (p. 181). Van Evera (1997) has chosen a much similar delineation, as he informs his readers that rational choice, critical, postmodern and constructivist approaches have been left out of his guide. We are thus dealing with two manuals, both of which theorise on the basis of narrowly defined conceptions of what it implies to theorise.

One path to follow from this observation could be to make a plea for more value theory in our theorising about the CFSP, or more empirical theory. However, the exact balance between the two types of theorising in studies of the CFSP is irrelevant for my present purposes. I want to follow a different path which, as its point of departure, realises that the very distinction is an immensely important part of Rosenau and Durfee’s conception of what constitutes theory and derived from that what theorising is all about. In line with arguments within political theory as presented by Ball (1995), I do not share the view that the distinction between ‘is’ and ‘ought’ and, hence, between empirical theory and value theory, is valid. Furthermore, I want to include more in my theoretical exploration of the CFSP than what Rosenau and Durfee call empirical theory. However, I will nonetheless henceforth use their rules as reference points and point out how an incorporation of other types of theory makes a difference, leading to a different research agenda, and hopefully to a better understanding of common European foreign policy.

What other types of theory are available? According to Chris Brown, theory at its simplest is reflective thought: ‘We engage in theorising when we think in depth and abstractly about something’ (1997: 13). From this rather broad conception, he proceeds by emphasising that it is helpful to make a distinction between at least three different types of theory – each associated with a class of questions. The first class of question is: Why do things happen? – leading to explanatory theory. An example from the field of CFSP could be why the European Union did not intervene in Albania in the spring of 1997. A second class of question is – What should we do? – where action can be either instrumental action or morally right action, leading to normative or prescriptive theory. Examples from the field of CFSP could be what ‘we’ should do concerning ‘failed’ states in Africa, for instance the falling apart of Kenya? When should ‘we’ intervene in international or domestic crises? The third class of question is oriented towards meaning – What is the meaning of this or that? – leading to interpretive theory. Examples from this area include the following: What does ‘common’ in the common foreign and security policy mean? Equally important, what does ‘foreign policy’ mean, given that we are dealing with an actor like the European Union, characterised by its specific type of polity? Most often such questions are not even raised; the meaning is simply taken for granted: foreign policy is foreign policy is foreign policy.

No matter which type of theory we have in mind, a theory consists of a number of assumptions, a set of definitions and concepts, and an explicitly described area of applicability. A theory includes assumptions about actors, structures, institutions, processes and the relations between these four components. Theories are always embedded in a certain metatheory or a certain philosophy of the social sciences. In this light, Wendt (1987, 1991) argues that a given metatheory privileges certain theoretical constructions and makes others less likely.

In contrast to theorising within a positivist or behavioural framework, there are unfortunately no manuals about theorising within the constructivist perspective. The absence of such a manual makes the task more difficult but not impossible. To begin with, we can note that constructivism is not itself a substantive theory of international relations. Rather, at the highest level of abstraction, it is a philosophy of science category that has a number of consequences for theorising international relations (see further in Jørgensen 2001). If constructivism is not a theory, how can we describe possible links between constructivism and first order theory? Are there any constructivist theories available? Once again Wendt (1994) has something to offer. He explains that his constructivist (pre)theory of collective identity formation makes the following core claims: ‘states are the principal units of analysis for international political theory; the key structures in the state system are intersubjective rather than material; state identities and interests are in an important part constructed by these social structures’ (Wendt 1994: 385).

Useful or not, it is only one example of constructivist theory. Let us therefore examine an alternative proposal. According to Checkel (1998), one challenge lying ahead for constructivists is to theorise. In my view, Checkel is right in pointing out that theory-building is among the key challenges for constructivists. In more specific terms, he suggests that constructivists should ‘specify more clearly the actors – structures and agents – and the causal mechanisms bringing about change, the scope conditions under which they operate and how they vary cross nationally’ (1998: 348). The language is very eloquent, facilitating dialogue about theory-building and, furthermore, the approach has the potential for bringing about contending theories. However, I have certain reservations concerning his explication of what it means to theorise from a constructivist perspective.

Checkel proposes that theories should be able to explain change, but why privilege factors bringing about change? While I think Checkel has a point and is right in believing that structural leanings among constructivists can be explained by reference to their opposition to methodological individualism, I find it equally important that many social constructions actually are characterised by stability and therefore are relatively enduring or static. In principle that could be a second reason for constructivists to favour elements of structure in their studies. Furthermore, why should we explain why and how constructions vary across countries? It puzzles me that social constructions are assumed to be neatly country-specific; obviously some are but that is far from being the complete picture. One example where it is possible to conduct comparative studies, across countries, is Anderson’s (1983) approach to research on processes of nation-building.

I believe it is equally self-evident that a whole range of social constructions are anything but country-specific, and then it does not make sense to ask theorists to use a design that makes comparisons across countries possible. Finally, it is puzzling that Checkel finds the need for theory especially evident at the domestic level. He explains that constructivists theoretically operate most strongly at the systemic level because they draw on what he calls systemic sociological literature. Here he seems to assume a certain congruity between the systemic level of international relations and systemic sociological theory. This assumption is quite misleading as systemic sociological literature can equally be applied at the domestic level and, besides, we also have other currents of constructivism. Why favour one over the other? In conclusion on this point, Checkel seems to believe that the choice of a constructivist theoretical framework has few consequences for key questions such as, What is theory? What can theory do for us? and, consequently, How can we engage in theory-building? In summary, instead of ending up opting for either empirical or normative theory, we can draw on Brown’s three types of theory. Furthermore, Checkel has sharpened important tools for constructivist theorising.

Rule 3. One must be able to assume that human affairs are founded on an underlying order (Rosenau and Durfee 1995: 182–3). They explain that this rule implies that theorists should assume that everything is potentially explicable; that is, presume that nothing happens by chance, capriciously, at random; that for every effect there must be a cause. Theory is based on the law of probability, hence it purports to account for key tendencies. The rule can be read in both an open and a critical strict fashion. An open interpretation would emphasise the law of probability; – that is, the view that not every single incident needs to be theoretically explicable and, furthermore, that to theorise is to be concerned with central tendencies. The element of ‘capriccio’ is to accept that things can actually happen by chance and will thus be theoretically inexplicable. Alternatively, we can imagine that certain apparently inexplicable developments may be explicable with reference to a competing underlying order – an order analysts hitherto have been unaware of. A strict interpretation of the rule will read it as yet another aspect of so-called ‘normal’ science. Van Evera has a clear preference for studies of cause/effect relations, which means that his book contains many useful hints concerning this particular type of analysis: causal analysis. However, I want to engage in something different, suggesting that other types of theory are possible and ought to be explored.

In contrast to van Evera, Ruggie points out the possibility for noncausal explanation, suggesting that ideational phenomena such as aspirations, legitimacy and rights ‘fall into the category of reasons for action, which are not the same as causes of action’ (Ruggie 1998a: 22, original emphasis). Wendt describes the contrast between causal and constitutive explanation in the following terms:

Clinton (1993) asks the interesting question, ‘International Obligations: To Whom Are They Owed?’ and answers that obligations can have at least three sources: (1) They can be owed to others, (2) to a set of rules, or (3) ‘first and foremost’ to oneself. In my view, it is clear that both the ‘open’ interpretation and Ruggie, Wendt and Clinton’s way of reasoning are of significant relevance to studies of the CFSP. In order to point out how, I will present three examples that lead to different types of theorising.

In the case of the CFSP, we can actually identify an underlying order, in this case a set of formal and informal rules and norms; an ever expanding set of treaty provisions; reproduced policy practices; an ever expanding scope of policy areas, engagements, commitments; ever higher aspirations; a policy area which because of subsequent enlargements of the Union, increasingly deals with relations between Europe and the rest of the world. Yet, we also have a dimension of order that we can characterise by apparently unavoidable instances of inaction, lack of consensus, lack of impact on external actors, events and developments. How can we explain this dialectic between unity and diversity when dealing with common European foreign policies? Can we perhaps say that the CFSP has had a significant impact on the ontological dimension – ‘who are we’? If so, this would lead to the conclusion that collective intentionality is far more predominant today compared with situations ten, twenty or thirty years ago.

The second example begins with the fact that, according to a narrow formal conception, the CFSP actually is an archetypical example of intergovernmental cooperation. It functions at different levels, from junior officials to senior ministers and diplomats; it makes use of a range of foreign policy instruments and has certain outputs in the form of declarations, demarches, statements and military deployments. The CFSP is nothing less than shorthand for the modern European foreign policy system; thus we can here potentially establish links with research conducted within the tradition of foreign policy analysis (FPA) (Smith and Carlsnaes 1994; White 1998). In this regard, a mathematical formula is helpful:

The original ‘Six’ made a system of fifteen bilateral relationships. ‘The Nine’ constituted a system of thirty-six bilateral relations and ‘The Twelve’ had sixty bilateral relations. The current system, ‘The Fifteen’, makes 105 relationships possible.¹ A twenty-state Union makes a system of 190 bilateral relations. Needless to say, the cultivation of such an impressive number of relations would be very costly in terms of time and manpower resources. So, clearly, the CFSP is a subsystem of some sort but is it more than that?

Example number three. In The Expansion of International Society (1984), Hedley Bull emphasises in the introductory chapter that the expansion and development of Europe’s international society was a co-constitutive process, meaning that Europe did not first develop international society and then embark on a huge system-export operation. Most of The Expansion of International Society is devoted to studies of how non-European states joined international society. The book is thus a typical product of the English School and, as such, it has a pronounced blind spot on European integration. Apparently, the honourable members of the School suffered from a well-developed phobia concerning European integration. In a sense, studies of the CFSP describe what happened to Europe’s international society after its expansion. Given Bull’s acknowledgement of co-constitutive processes, I assume that the ever expanding society is of importance for the original (European) international society, also after the Second World War, during decolonisation and during the decline of the Western European great powers. In other words, though I find the themes analysed in The Expansion of International Society interesting, I find it at least as equally interesting to analyse what has happened to international society in Europe during the last three to four decades. It seems to me that CFSP practices constitute a significant part of contemporary international society.

Rule 4. One must be predisposed to ask about every event, every situation or every observed phenomenon. Of what is that an instance? (Rosenau and Durfee 1995: 183–5). They encourage us to search for patterns, generalisations and abstractions, suggesting that we avoid treating phenomena as unique and propose that we cultivate our impulses to always search for more general theoretical insights. This rule is absolutely splendid and hence a very fruitful component of any research strategy. Of course, once again, the nomothetical ideal lurks in the background and some of the explanations are more purist nomothetical than I find necessary. The problem is, in Denzin and Lincoln’s words, that ‘too many local case based meanings are excluded by the generalising nomothetic positivist strategy. At the same time, nomothetic approaches fail to address satisfactorily the theory and value-added nature of facts, the interpretative nature of enquiry and the fact that the same set of facts can support more than one theory’ (1994). Constructivism focuses on local (emic) rather than on general (etic) phenomena; on middle range rather than grand theory, implying that it is a great paradox that we have so few constructivist middle-range theories. Middle-range theorising should be to constructivists what Old Trafford is to Manchester United. According to Ruggie (1998a: 34) ‘concepts are intended to tap into and help interpret the meaning and significance that actors ascribe to collective situations in which they find themselves.’ The huge gap in the literature is ready to be filled, and I think it is a fairly reasonable prediction that this will be one of the most busy construction sites in the near future.

Rule 5. One must be ready to appreciate and accept the need to sacrifice detailed observation for broad observations (Rosenau and Durfee 1995: 185). This rule is explicated by Rosenau and Durfee with the following words: ‘Theory involves generalizing rather than particularizing and requires relinquishing, subordinating, and/or not demonstrating much of one’s impulse to expound everything one knows’ (p. 185) as such, the rule is interesting and fruitful for my purpose. Presumably, failures in observing this rule led Keohane and Hoffmann to their assessment of European research as being ‘longer on detailed description than analysis’ (1990: 276).

The relevance of the rule for studies of the CFSP is considerable. Not uncommonly they are rich in detailed observation and somehow lack broad, generalised observation. Two things are relevant in this context. First, if we consider the passage suggested by Rosenau and Durfee, it does not necessarily imply that we enter analytical modes that depend on large n data-sets that may lead to correlations and identification of possible cause–effect relationships. In van Evera’s discussion of large n data-sets, he points out some of the weaknesses: either that we have to assemble data (may be difficult or costly) or, if we rely on existing data-sets, they may prove inadequate for our research question. The obvious example within studies of the CFSP is the use of statistics on European governments’ voting behaviour in the UN. The good news is that such statistical data are available. The bad news is that these statistical data cannot serve as exclusive data if we want to understand voting in the UN. The most illustrative and interesting things about voting in the UN are to be found in the explanations for voting; that is, in reasons for voting rather than in the votes themselves. Second, it has been an implicit or explicit axiom among CFSP analysts that it is best seen as a unique phenomenon. Consequently, it has been regarded as futile to examine similarities and differences between, say, CFSP, WEU, OSCE and NATO. The CFSP is certainly a unique phenomenon but one could and should ask whether it is also an example, among others, of multilateral intergovernmental forms of cooperation. In other words, how different is intergovernmental cooperation in the CFSP from intergovernmental cooperation within NATO? Such a comparison, to my knowledge, has not been attempted by anyone, meaning that we are unable to specify differences between the two institutions and, in turn, to specify how different or similar processes of collective identity formation work in the two institutions.

The absence of such comparative studies is unfortunate because we cannot reach general conclusions about intergovernmental cooperation or something about the specifics of each of the institutions mentioned. Concerning the specifics of the CFSP, the body of knowledge that is constituted by actors’ self-reflectivity has seldom been analysed systematically (for exceptions, see Andersen 1998; Glarbo 1999). In other words, we arrive at the classic triangle including author, reader and text; and thereby arrive at equally classic issues such as author intention, original intent and the role of the reader (see Collini 1992; Ball 1995).

Rule 6. One must be tolerant of ambiguity, concerned about probabilities and distrustful of absolutes (Rosenau and Durfee 1995: 186). They emphasise that ‘To be concerned about central tendencies, one needs to be accepting of exceptions, deviations, anomalies, and other phenomena that, taken by themselves, run counter to the anticipated or prevailing pattern’ (ibid.). Furthermore, they emphasise that theories have to live with or work with uncertainty. This rule is as if tailored for CFSP studies. Doing research on the CFSP requires a high degree of uncertainty. We should therefore not be too concerned about fluctuations between substance and empty rhetoric in CFSP policies, about the fact that probabilities live a prominent life in the world of CFSP policy-makers or that uncertainty continues to reign concerning the significance and influence of European foreign relations. Instead, we should accept that knowledge about the CFSP is produced as a product of continuous dialogue between the CFSP and its critics and the critics and their CFSP. Exceptions have long played a prominent role in research on the CFSP, primarily because analysts for various reasons have attempted descriptions that take realism in art as inspiration; that is, they feel that every single detail ought to be present.

Rule 7. One must be playful about international phenomena (Rosenau and Durfee 1995: 187–8). They point out that the core of theorising is creative imagination. Only deep penetration into ‘a problem discerning relationships that are not self-evident and might even be opposite to what seems to be virtually apparent can produce incisive and creative theory. Good theory ought never be embarrassed by surprises, by unanticipated events’ (p. 187). They also suggest that we play the game ‘as if’; that is, engage in counter-factual analysis. An obvious example of being playful about the EPC/CFSP is to ask what would have happened had the EPC/CFSP not existed; that is, whether European foreign policies, state identities, state relations, Europe’s influence on non-European affairs and so on would have been any different had the EPC/CFSP not existed during the last thirty years. If this is not the case it would obviously be a waste of time to do research on the CFSP and even more futile to theorise about it. It is, therefore, a key assumption that the CFSP actually makes a difference, meaning that the task is to specify the differences it makes.

A second example is even more playful. Pierre Hassner provides the clue of this example raising, back in 1987, a very provocative question. Even if it is a lengthy quote, in the present context it makes sense. Hassner asks: ‘Does any Western power have an East European policy, or are all Ost Europa Politiken both more and less than policies towards Eastern Europe in the sense that they usually involve more than Eastern Europe but amount to less than policies?’ (p. 189). The question is relevant for all Western states but, according to Hassner, for none more than France. He adds,

And then he ends, ‘from de Gaulle to Francois Mitterand French leaders were cast more than ever as specialists in vision rather than policy, in words rather than deeds’ (ibid.). Many are inclined to take it for granted that France has a policy towards Eastern Europe but it appears not to be as simple as that. Maybe we should be aware of more subtle levels of policies or visions.

The third example is based on pure behavioural observation. The EPC/CFSP experience covers a period of thirty years. Junior diplomats, senior foreign ministers, European commissioners, and heads of state and government are all engaged in the EU’s external relations. A lot of work, time, travelling, meetings, declarations, consultation take place, implying that a significant number of diplomats devote a considerable part of their time to CFSP issues and considerable financial resources are spent on the CFSP. We can add that the EU has more diplomatic representations around the world than most middle or great powers have embassies. While we professionally know that politics and diplomacy can function rather well, even at a long distance from a common-sense reality, it, at least to me, is somewhat difficult to imagine that all these people, engaged in all this activity, would do so if it did not serve a purpose or if it did not make a difference. I therefore take practices as a hint that we are dealing with an important aspect of modern European diplomacy. In line with this, one could ask whether an analysis of CFSP’s output really is futile. By ‘output’ I am not only thinking of the about 5000 documents adopted during the last decade² but also of hypothetical output in the form of changed foreign ministry organisation, state identities and interests, and diplomatic practices.

The tentative findings gained from the three examples suggest that the CFSP actually makes a difference and thus seems to be an important aspect of modern European diplomacy, but also that certain key features of states’ dealings with foreign affairs seem to have changed during thirty years of increasing interaction.

Rule 8. One must be genuinely puzzled by international phenomena (Rosenau and Durfee 1995: 188–9). According to Rosenau and Durfee, we should be ‘as concerned about asking the right questions . . . as one is about finding right answers . . . . Genuine puzzles are not idle, ill-framed, or impetuous speculations. They encompass specified dependent variables for which adequate explanations are lacking’ (p. 188).

Bridge-building between different perspectives seems possible here, particularly because constructivists are inclined not to take things or concepts for granted. It thus seems to me that constructivists are well armed to meet this challenge of asking the right questions. Reconceptualisation is an important part of constructivists’ research agenda, an agenda that can be motivated both by normative and analytical concerns. Following Berger’s (1986) advice, constructivists do not do research on a phenomenon’s function before they have done research on its meaning. In Berger’s words, we should not put the cart before the horse. In other words, this is where meanings, understandings and ideational factors come in.

Let us examine three examples. The first example concerns Germany’s role in world politics. German debates during the 1990s on participation in peace support operations (PSOs) and permanent membership of the UN’s Security Council demonstrate the importance of state identities. Obviously, such debates can be reduced to debates about means and ends but they can also be analysed from the perspective of identity politics, raising questions such as, Is Germany a great power? Ought Germany be represented in the Security Council? Is Germany a civilian power? Is an engagement in PSOs compatible with perceptions of Germany’s role in world politics? (see further, Zehfuss 1998).

The second example concerns contemporary conceptions of states. Current hegemonic ideas about what constitutes a modern European state mean that the European Union has become a centre of gravity, that the Union has established itself as the prime institution representing what is widely considered modern Europe. Previous alternatives such as the Council for Mutual Economic Assistance (CMEA or Comecon) and the Nordic model have all vanished, or declined into romantic dreams about the past. Ideas about modernity combined with the so-called ‘magnet effect’ mean that non-member states know in which direction they have to move if they want to become members and thereby be recognised as modern, as non-marginalised, as premier league players. The Austrian slogan of the late 1980s – ‘Wir sind Europareif’ – was among the first expressions of this new balance of ideational power.

The third example is a bit more abstract and concerns the relevance of Habermas’ theory of communicative action. Risse (1998) points out that

These observations are very relevant for our understanding of the CFSP. Why? Because contrary to NATO, the CFSP does not include a hegemonic enforcer.³ Risse (1998) expects pockets of discursive and argumentative processes in world politics: ‘I hypothesise that international institutions can serve as a substitute for the absence of a common life world . . . they provide a normative framework structuring interaction in a given issue-area. They often serve as fora or arenas in which international policy deliberation can take place’.

In contrast, it is significant that van Evera (1997: 26) reduces the importance of actors’ views and perceptions: ‘the insights of the actors or observers who experienced the event one seeks to explain can be mined for hypotheses’ because they ‘often observe important unrecorded data that is unavailable to later investigators.’ Again we see limited ambitions: the aim is merely to explain an event – and insiders’ insights are reduced to a means for hypothesising. In contrast, in constructivist approaches ‘the inquiry aims . . . are oriented to the production of reconstructed understandings, wherein the traditional positivist criteria of internal and external validity are replaced by the terms trustworthiness and authenticity’ (Denzin and Lincoln 1994). Because practitioners have been associated with research on the EPC/CFSP, no matter which strategy one chooses, there is a rich literature to ‘mine for hypotheses’ or to use for the reconstruction of understandings (see Andersen 1998; Glarbo 1999).

According to Ruggie, constructivist theories represent a ‘commitment to the study of the world from the point of view of the interacting individuals’ (1998a: 85). In the case of the CFSP, it follows that three groups of people are of interest: diplomats, politicians and observers (journalists, commentators and academics). Diplomats and politicians produce texts (statements, declarations, speeches) and observers interpret these texts in their way – which is sometimes far from the intention of the text producers. Nevertheless, they (we) contribute in the exercise that gives meaning to discursive practices. Ruggie (1998a) emphasises that ideational factors shaping actors’ outlooks and behaviour include culture, ideology and aspirations.

Rule 9. One must be constantly ready to be proven wrong (Rosenau and Durfee 1995: 189–90).

Conclusions

Instead of hard-rock conclusions, I reach a number of hypotheses. Could it be that the CFSP actually is ‘better’ than the rumours about it? Among the reasons for not knowing could be that we have analysed the CFSP in the wrong way and for the wrong things. A second reason for the CFSP’s bad publicity could be that many analysts are over-ambitious on behalf of the object they study. The third reason could be that drama, failure and disagreement are what make good newspaper stories that subsequently are read and distilled by foreign policy analysts, specialising in the CFSP. A fourth reason could be that analysts do not tend to care about actors’ self-reflections about the CFSP’s weaknesses and strengths. The fifth and final reason could be that analysts tend to think that knowledge acquired during the Cold War still applies. Could it be that the world has changed and the CFSP has changed with it? Let me summarise my view in a rather unusual way.

I think the relationship between the CFSP and national foreign policies can be described in a way very similar to how the Danish painter, Asger Jorn, painted The Disquieting Duckling. In the background we see the well-known world, the world people feel comfortable with, well described and viewers are not uneasy about it because it gives them comfort to think in a rather kitschy way. Then enters the disquieting duckling, an instance of The Shock of the New (Hughes 1991), something we are not used to, not familiar with and not comfortable with. We are uncertain about precisely what it means and what it implies. After some thirty years, we still do not know precisely how to make sense of the CFSP. So we choose between neglecting it, meaning that we study national foreign policies of European states as if the CFSP were of no significance whatsoever – or we analyse the CFSP as if it were the embryonic foreign policy of a state in the making, sort of Italian foreign policy in the nineteenth decades of the last century (on this, see Chabod 1996). Maybe, if we are interested in the significance and impact of the CFSP, we have been looking for the wrong things and used the wrong instruments.