EIoP: Text 1999-001: Full text

The Amsterdam Process: A Structurationist Perspective on EU Treaty Reform(*)

Thomas Christiansen and Knud Erik Jørgensen

European Integration online Papers (EIoP) Vol. 3 (1999) N° 1; http://eiop.or.at/eiop/texte/1999-001a.htm

Date of Publication in

: 15.1.1999

I. Introduction

Intergovernmental Conferences are generally seen as key events in the design of the European Union, the moment at which Member State governments re-assert their hold on the institutional dynamics of the EU(1). In this paper we challenge this traditional intergovernmentalist view of treaty reform. Rather than a succession of relatively rare ‘events’, treaty reform should instead be regarded as a continuous process.(2) In so far as IGCs actually perform a function for national governments, this is rather ceremonial (states celebrating their status as states) and disciplinary (states exerting a slightly greater than usual degree of control over the agenda and time-scale of reform) than actually decisional. In the absence of a complete and final Kompetenz-Kompetenz in the EU, IGCs constitute the attempt by governments to assert their control of the expanding portfolio of EU competences. But the wave of IGCs during the 1990s, we believe, demonstrates conclusively the opposite: governments have lost command and control not just over every-day EU business, but even over something as ‘intergovernmental’ as treaty reform.

Rather than merely adding to the growing list of substantial critiques of intergovernmentalism, this paper is directed at developing a novel, structurationist perspective on EU Treaty reform.(3) Our approach is based mainly on the recognition that the crucial object of analysis cannot be Member State interests, but the process in which these are constructed. Structuration theory, an approach developed by Anthony Giddens (1984) and increasingly debated by a number of scholars in International Relations – though still rare in European Studies – offers a meta-theoretical framework for such an exercise, since it explicitly demands the integration of agency and of structural explanations into a holistic perspective.(4) Political actors and social structures are conceived as co-constituting one another, and it is the focus on the interplay between these two factors which makes process the object of analysis.

The time is certainly ripe for such a perspective on the EU. The last ‘instance’ of treaty revision, the 1996-97 IGC, has emphasized the procedural nature of ‘constitutional reform’: the summit came at the end of an intergovernmental ‘conference’ which had in fact taken more than a year to define the ‘choices’ of Heads of State and Governments. The IGC process involved numerous presidency drafts, and the submissions to the conference from national governments, EU institutions and non-governmental organizations spawned an impressive Commission-run website.(5) But the IGC was itself part of a wider process of pre-reform, which included activities of the Reflection Group as well as a number of formal and informal review exercises at the Member State and European level. In any case, the brief for treaty reform was already enshrined in the Maastricht Treaty, and the ultimate question for national delegations was never the ‘if’, only the ‘how’ of treaty reform.

Neither did the reform process end in Amsterdam. By contrast, a number of crucial "non-decisions", mainly about the re-alignment of the Council voting mechanism, mean that further high-level meetings will be necessary, while other important decisions about the institutional shape have been left to uncertain automatisms which are linked to the entry of new Member States. There is, as yet, no clear picture about the final shape of the reform effort. And just as in Maastricht, also Amsterdam already includes the agreement on the next phase of treaty reform, which is, in any case, necessitated by the prospect of enlargement, now more imminent after the publication of Agenda 2000 and decisions taken on enlargement at the 1997 European Council meeting in Luxembourg.

In this paper we develop a process-based understanding of treaty-reform. As part of this process, IGCs deserve particular attention. But whereas IGCs are traditionally seen as summits, both in political and in analytical terms, we seek to re-direct analytical attention away from the political limelight and into the ‘valleys’ – the periods between the IGC-summits during which in fact the more momentous developments of European integration have occurred. If this paper succeeds in its aims, it will not only dissolve one of the cornerstones of intergovernmentalism – the over-arching significance of IGCs – but also provide an outline of an alternative conception of treaty reform in the European Union. Based on arguments about the nature of treaty reform as constitutionalization and on the essentially procedural nature of this aspect of European integration, future research should then be able to identify more authoritatively where, and how, and by whom, fundamental decisions about treaty reform have been taken.

II. A Structurationist Perspective on European Integration

‘Amsterdam’, in our conception, was in fact a summit: but not so much in the conventional sense of the all-important, "make-or-break" meeting of senior statesmen, but rather as the high-point of a sloping curve that describes the ups and downs in the lengthy process of treaty review, reform and revision. Similar things can be said about Maastricht, which was preceded by 18 months of negotiation and followed by an even longer phase of ratification. Maastricht and Amsterdam, and the IGCs that preceded and succeeded these European Council meetings, cannot be seen as isolated instances of reform. Member States did not come to the conference table each with their set of ‘national interests’, and the ‘final outcome’ was not the compromise or lowest common denominator between these. Throughout, the reform process has been structured massively, through the pre-defined demands on the IGC, the convergence of beliefs about its outcome, the institutionalized allocation of interests built into the EU system and the path-dependency of past choices.

But even beyond this limited time-scale it is possible – and analytically desirable – to view treaty reform not as a succession of specific instances but as part of a larger ‘meta-process’ of constitution-building and polity-formation in the European Union. Just as the Single European Act and the ‘1992’ programme were structured by previous developments in the EU system – the Court’s Cassis de Dijon judgement and the Commission’s White Paper on the Single Market – monetary union agreed at Maastricht was based on the foundation of the European Monetary System and the flexibility clause agreed at Amsterdam was preempted by the opt-outs and the experience of variable geometry, which had already become a hall-mark of the EU system in the 1990s. Treating each IGC in isolation means that this continuous process is cut up and thus a misleading image of significant ‘events’ dominated by actors is being created. If our perspective is a convincing one, then it follows that the significant object of analysis should not be the IGC itself, and even less so the ‘Amsterdam event’, but the ‘Amsterdam process’.

Questioning the ability of actors to ‘decide’ the course of historic processes such as European integration always carries with it the danger of determinism on the other end of the spectrum, i.e. handing over the explanatory power altogether to structural factors. While currently not much in academic fashion, this has nevertheless been a frequent tendency with regard to the integration process. Structural determinants have been with the European Community from its inception, whether it involved the continent’s need to reconstruct, or the European response to the emerging bipolar world of the 1950s, or the western bulwark against Communist expansion (the much cited ‘external federator’ theme has always been around), or the ‘natural’ outcome of the communication and technology revolutions in the 1980s, or the bare necessity as the third leg of the global economic triad in the 1990s.(6)

No one would deny that all of these factors have played a role in the evolution of the European Union, and we will in turn deal with the more important ones. The point to be made here, though, is that none of them can be accepted as the ‘Big Factor’, the sole cause which has propelled the integration process forward. Just as we are suspicious of the actor-centered voluntarism of intergovernmentalism or neo-functionalism, we should be wary of structural determinism. Even a cursory look at the history of the European Union demonstrates that individuals, particular actors occupying a particular place at a particular time, had a massive influence in the course of events.

The study of the Amsterdam process, in a structurationist perspective, requires proper recognition of both the number of actors involved in the process and the routines and practices structuring treaty reform. Due to the actor-centered analysis of much of political science literature on EU decision-making, we know quite a lot already about the actors involved in the process. Precisely because the common approach to IGCs still appears to be centered around the abstract positions of Member States rather than a concern with the negotiation among a group of individuals, much can be gained from introducing insights from policy-analysis, diplomatic studies and sociological symbolic inter-actionism about the socialization effects of continuous, face-to-face contact. By actors, then, we mean individual civil servants, Commission officials, MEPs, national ministers and Prime Ministers rather than personified states.

More important still, in the context of a rigorously structurationist account of EU treaty reform, is the identification of elements of structure. This includes a specific acquis conferencielle, our term for the formal arrangements and institutional set-up of IGCs, which have emerged over the past decade and to which IGC participants now have to conform. It also includes the path-dependency of institutional development in the EU: IGCs do not start from a tabula rasa, on which national governments put their shopping lists, but from an already existing mosaic to which some parts may be added and others re-arranged, but which essentially is already in existence, which means that a whole range of potential shopping-list items are considered so inappropriate that they never make it to the IGC. There is also the structuring force of the particular IGC discourse that precedes and pervades the negotiations. Many actors actively contribute to this discourse, but in due course every participant – including those from national governments – eventually becomes a respondent to previously made statements by others.

Having identified these elements, our analysis demonstrates how it is this interplay between agency and structure, rather than one or the other, which ultimately determines the nature and trajectory of the process. This, we argue, sheds new light on the understanding of EU treaty reform and the wider process of integration. In terms of the theoretical debate, it will provide a constructive critique of inherently actor-centered approaches like intergovernmentalism. At the same time it helps produce an account of integration as a constitutive process that is more rigorous than neo-functionalism, the more process-oriented integration theory.

Notwithstanding the acknowledged difficulties of its application (Stones, 1991; Layder, 1987; Gregson, 1987; 1989), we believe that a structurationist approach has significant value – not as a grand theory but as an analytical perspective. It is of value for this kind of research because it proposes a reconciliation of agency and structure through a focus on process – the process of structuration. Since process is precisely what the study of European integration should be about, a structurationist perspective is an obvious one to take when studying treaty reform.

There is no space here for an exhaustive elaboration of the theory of structuration. It is, after all, less of a theory than a set of abstract propositions about the reproduction of social life. For the purposes of this paper, it needs to be adapted for empirical application. What follows is just a brief introduction to the essential notions of structuration theory that are of relevance for our analysis.

The central concern for structuration theory is the resolution of the structure/agency problem in social science. A key concept in this endeavour is the concept of a social praxis, which is defined as

the constitution of social life, i.e. the manner in which all aspects, elements and dimensions of social life, from instances of conduct in themselves to the most complicated and extensive types of collectivities are generated in and through the performance of social conduct, the consequences which ensue and the social relations which are thereby established and maintained... This view of praxis is equally relevant to the constitution of action and the constitution of collectivities (Cohen, 1989: 12).

This focus on social praxis involves further theorizing on actors and structures. In structuration theory, actors are being ‘de-centered’ ‘in favour of a concern for the nature and consequences of the activities in which [they] engage during their participation in day-to-day life’ (Cohen, 1989: 12). Of central significant is the concept of a ‘duality of structure’. Structures are understood as both medium and outcome of the reproduction of practices. Structure enters simultaneously into the constitution of social practices and ‘exists’ in the generating moments of this constitution (Giddens, 1979: 5).

Through the concept of the duality of structure, structuration theory develops an account of the institutionalization of social life – an account in which institutions are understood as established social practices rather than as mere organizations. The structurationist points of departure are the rules, norms and patterns of behaviour that govern social interaction. These are structures, which are, on the one hand, subject to change if and when the practice of actors changes, but on the other hand structure political life as actors re-produce them in their everyday actions. A brief glance at the way in which this abstract understanding can be applied to EU treaty reform ought to demonstrate its utility.

As we noted, most research into treaty reform is of an actor-centered nature. Frequently, though, agency is equated with the actions of Member States, which are assumed to act collectively in the ‘national interest’. But it doesn't take much deconstruction to recognize that ‘Member States’ as such have no capacity to act. As we seek to show in this paper, if the ‘Member State’ is ‘unpacked’, the image of hierarchy may be replaced with a more accurate picture of individual action based on empirical research. This includes the recognition of internal divisions, for example when negotiators in Brussels act in response to perceived opportunities and constraints at the conference table. It also opens the possibility of transnational loyalties, such as the potential formation of policy-networks or even epistemic communities among the national experts who conduct these ‘intergovernmental’ negotiations collectively and over long periods of time. Seen in this light, the single-actor-capacity often attributed to Member States in the course of treaty reform is quickly exposed as a myth.

In cases where the intergovernmentalist approach does recognize that agency is located at the level of the individual, these are usually identified as the senior officials in national administrations, permanent representatives, ministers and prime ministers who are directly involved in the negotiations. But whereas traditional research proceeds from there, assuming that it is agency within the Member States which drives treaty revision, the structurationist perspective proposed here helps to draw a wider circle of actors into the picture. If the long-term process of treaty reform is the starting point for analysis, then Commission officials and Council Secretariat staff (who prepare the meetings and thereby help to set the agenda), activists in non-governmental organizations (whose demands might find their way into IGC deliberations) and the judges of the ECJ (whose interpretations of primary law will transpose the letters of the treaty into practice) all have a part in treaty reform.

The analysis below provides details on the role these actors play in the treaty reform process. This demonstrates, on the one hand, that agency matters, and that it is taken into account in a structurationist approach advanced here. But it also shows that removing the ontological primacy, which is traditionally bestowed upon Member State representatives in the analysis of treaty reform, helps to remove some of the false assumptions about this aspect of European integration. ‘Proving’ that Member States are in control of ‘intergovernmental bargaining’ by starting with the input from Member States is a tautology which ultimately obscures much of what is analytically relevant.

Beyond the inclusion of a wider set of actors, a structurationist perspective also adds to the analysis by embedding them within social structures. Based on the recognition that no social action occurs within a social vacuum, structuration theory situates the role of agency within the social structures surrounding it. It does so not in any deterministic way, as agency is potentially able to change any aspect of these structures. But the extent to which much agency reproduces rather than changes political structures, the nature and the effect of these structures ought not to be ignored. Structuration theory here draws attention to the ‘unconscious motivation’ of actors – the recognition that much of what actors do is repetitive and regular practice based on internalized customs and norms rather than due to any preconceived, rational intentions (Cohen, 1989: 52).

Much of EU treaty reform has become structured by such ‘routine modes of conduct’ internalized by the actors involved, just as much of EU decision-making more generally has become increasingly regularized. In analogy to the acquis communautaire, we have termed this pattern of accepted behaviour the acquis conferencielle. This includes, for example, the practical arrangements of conference room negotiations as well as the procedures for the politically very sensitive drafting of revisions to the treaty. With respect to the latter it is a matter of custom now that progressive drafts are prepared by the Presidency with support from Council Secretariat and Commission.(7) This arrangement is not an ‘iron law’, and – as the example of the Dankert draft of the Maastricht Treaty demonstrated – agency might well come into play and lead to unforeseen – and perhaps unintended – consequences.(8) But such examples are exceptions to the rule, which is one of the many rules of treaty reform that actors normally do not question. In this perspective, it is analytically significant to recognize that the acceptance of such ‘rules’ is not based on intentional action but on unconscious modes of conduct – as identified by structuration theory.

Thus, structuration theory does not deny the freedom of agents to intervene and to change the course of events. Instead, it develops an account of social life in which agents interact continuously with rules, customs and other social institutions. The social practices, which result from this interaction, constitute political and social structures. Structuration theory therefore does not point to any larger structure controlling human behaviour, but simply recognizes "institutionalized routines as constitutive of the daily transaction of events" (Cohen, 1989: 39). In advancing such a structurationist account of treaty reform, we therefore do not point to any externally determining forces compelling Member States in any way, but to the social practices and routines that their representatives and other actors have developed and institutionalized in the process of treaty reform. Recognising these elements of structure, as well as a more inclusive set of actors, is the additional value that structuration theory brings to the analysis of EU treaty reform.

III. Treaty Reform and Constitutional Reform in the EU

Before applying the structuration approach to what we view as the ‘Amsterdam Process’, we need to identify in more detail what we regard as the elements of treaty reform in the European Union, both in principle and in history. From the above follows that a structurationist perspective is sceptical of both voluntarism and determinism. Indeed, the approach is sceptical of any orthodox ism, whether intergovernmentalism, structuralism, or methodological individualism. However, it is not only our meta-theoretical perspective that differs from traditional perspectives – so does our understanding of EU treaty reform.

The conventional understanding is that treaty reform is the result of changes in the treaty as a result of an intergovernmental conference. As such, it is most commonly regarded as an event – the symbolic reference point of this conception is, of course, the signing of a new or revised treaty during a European Council meeting. This, after all, was how the Maastricht and Amsterdam treaties received their popular names. This approach has the obvious advantage of delimiting the inquiry to an often short span of time and to formally organized venues. The intergovernmental nature of IGCs appears to determine that prime decision-makers are governments. Other actors may try to influence the agenda of IGCs, or they may try to influence the decision-making process at IGCs, yet governments retain the power derived from being formal decision-makers. Such an understanding of treaty reform makes various intergovernmental approaches the obvious choice, reducing research design to a choice between various versions of intergovernmentalism, including neo-realism (Grieco, 1991), pragmatic intergovernmentalism (Keohane and Hoffmann, 1990), liberal intergovernmentalism (Moravcsik, 1994), rational choice (Garret, 1992), and game theory (Tsebelis, 1994). These versions have all been ‘proven’ in analytical battle, and the majority has been developed to a sophisticated level of inquiry.

But in our view, this is a limited and limiting perspective, which distorts the procedural nature of treaty reform. A broader conception would also include the interpretation or explication of existing treaties, agenda-setting for the IGC and the post-IGC implementation of any decisions taken. To substantiate this argument we draw attention to the fact that governments are clearly not the prime authoritative interpreters of EU primary law. This does not mean that any treaty interpretation goes. Instead, certain (re)interpretations become, by various means and for various reasons, authoritative and therefore trend-setting. We also include more than political decision-making because processes of administrative implementation are widely regarded in administrative science to influence policy-outcomes significantly. The literature on multi-level governance, for example, focuses predominantly on policy-making. Yet, the governance problematique also includes aspects of constitutional politics, organizational design and judicial politics. Our point of departure, then, is the observation that the conventional approach to the study of EU treaty reform has, by and large, failed to connect the analysis to the continuous process of EU constitutionalization.

Let us begin in the known world and on solid ground. Perhaps surprisingly, Kenneth Waltz (1979) can be of great help for our exploration. Whereas he does not include many sentences on the EU he does emphasize that in contrast to the international anarchical system, domestic systems are centralized and hierarchical. How can Waltz’ distinction be helpful for our exploration? For one, it points to the close but largely unexplored connection between realism and law. It is very much the pure, positive law tradition that underpins realists’ conception of the international ‘anarchical’ system, indeed the key distinction between hierarchy and anarchy. Next, and here we follow Stone’s observation that Waltz’ description of domestic systems "very much [is] the language of constitutional law" (1994: 449). Waltz does not raise the issue of the relevance his description has for our understanding of the EU so here we arrive at the point where we have to depart from Waltz.

But before going further, it is necessary to look at the intergovernmentalist approach to treaty reform in more detail. Below is an argument which is exemplary of this approach and which therefore deserves to be quoted at some length:

The primacy of the Member States is reinforced by the fact that, unlike ‘normal’ states, the EU is founded not on a constitutional but on a treaty base. In the first place this leads to a balance of power between the centre and the parts (Member States) very different from that which exists in federal systems, in that only the Member States are signatories of the amended treaty. Second, the treaty basis allows Member States to isolate some areas of policy from the ambit of the EU, limiting the power of the supranational institutions over them, as they did with the creation of the pillared structure in the treaty on European Union. Treaty evolution can occur without reference to previous agreements. While constitutional amendments require attention to be paid to previous developments, new developments can be quite separate, as with the signing of a treaty on Political Union quite separate from the Treaty of Rome. This clearly endows Member States with a large degree of control over institution-building (Hurrell and Menon, 1996: 391).

This is a bold statement that sounds convincing in its entirety. But by examining each aspect of this statement we show that it is precisely this understanding of treaty reform which prevents an alternative view – one in which Member States have very limited control over institution-building.

In assessing "[t]he primacy of the Member States", Hurrell and Menon work with a unitary actor conception of states. In talking about Member States they do not distinguish between national governments (who are the formal players in IGCs) and other state- and non-state actors who intervene in the process of constitutionalization outside IGCs. These actors are parliaments, national courts, national central banks as well as domestic interest groups, all of which do have positions on the constitutional issues raised by treaty reform. In Moravcsik’s work on the ‘strengthening of the state’ in the process of integration (1994), an awareness of the necessary internal differentiation of Member States was added only with hindsight. And yet, as the first Danish Maastricht referendum and the Maastricht judgement of the German Constitutional Court demonstrated, such domestic forces have a potentially rather powerful influence on the negotiation and re-interpretation of any IGC ‘decision’. Lacking a way of distinguishing between governments and the much wider significance of ‘Member States’, Hurrell and Menon cannot measure the relative influence of these institutional actors on the process, or, for that matter, the impact of the process on these actors.

Hurrell and Menon argue that "the EU is founded not on a constitutional but on a treaty base" and that "Treaty evolution can occur without reference to previous agreements [and that, w]hile constitutional amendments require attention to be paid to previous developments". But, whereas the EC originally was indeed founded on a treaty base, it is now based on a constitutional legal order, or, as the ECJ termed it, on a "constitutional charter". If they are denying the existence and significance of this process of constitutionalization – which has created a so-called hard core of constitutional principles which cannot be changed, not even by the Member States collectively – then Hurrell and Menon might just as well deny the relevance of the ECJ.

Hurrell and Menon are looking for "a balance of power between the centre and the parts (Member States)". But this focus may well be the reason that they are unable to identify any significant changes. What is required in the analysis of treaty reform is not so much the balance of power between to levels, but their interaction in the creation of new structures. In this regard, it is more useful to employ a concept which views power in structural rather than in purely relational terms. What matters in a system as densely structured as the EU is the evolution of structural forms of power, and analytical concepts used to study this phenomenon ought to be able to capture that.

Hurrell and Menon claim that the foundations of the EU are "very different from that which exists in federal systems, in that only the members states are signatories of the amended treaty" – in our view, a very formalistic view of things. Certainly, Member States are the sole signatories of the amended treaty, but the EU legal order is in fact very similar to the one which exists in federal systems (Weiler, 1991). Perhaps governments intended the EU legal order to be different from federal systems (they have, after all, eschewed the dreaded F-word so far), but crucially it has become constitutionalized even though governments are the ‘sole signatories of the amended treaty’.

According to Hurrell and Menon, "the treaty basis allows Member States to isolate some areas of policy from the ambit of the EU, limiting the power of the supranational institutions over them". Again a very formalistic assumption about the keeping of watertight compartments. The example of the British Social Chapter opt-out, with its first directive on European Work Councils seeping into the UK via the voluntary application of its provisions by industry, is a practical example of this dynamic. Governments can try to isolate some areas of policy but they are hardly ever successful in such attempts. Weiler’s (1993) critical evaluation of the pillar structure developed at Maastricht, which was to fulfil precisely this function, convincingly demonstrates the ‘leaks’ between the divisions national governments thought up as a response to the gradual loss of control over the process.

This leaves, in our view, very little substance to the "large degree of control over institution-building". For a start, the statement begs the question of what is understood by "control". Clearly it can only be seen in terms of ‘more’ or ‘less’ rather than absolute control. In that sense it is certainly possible to look at institution-building during the first 45 years of European integration and come to the conclusion that governments did not have a "large degree of control". From the outset, a secular process of creeping expansion of competences and powers of the supranational institutions achieved precisely what key Member States such as France (under de Gaulle) and Britain (under Thatcher and Major) vowed to prevent. Occasional ‘crises’ such as France’s ‘empty chair policy’ 1966-7 only serve to demonstrate the way in which governmental attempts to impose control ultimately failed. Governments may have sought to re-assert their control in this respect during the 1990s, but even formally the 1996-97 IGC was not a ‘pure’ intergovernmental affair since the ‘right’ to participate was granted to the European Commission and to European Parliament representatives.

Let us come back to what Hurrell and Menon know "as a fact": that the base of the EU is not constitutional. We have reasons to doubt this "fact". Indeed our inquiry leads inevitably to the conclusion that EU treaty reform is a constitutional process. In order to persuade sceptics and to make our conclusion plausible, we take a broad view of constitutional reform, one that encompasses constitutionalism (Weiler, 1997), constitutional law (Pescatore, 1981; Gialdino, 1995; Temple Lang, 1998), constitutional politics (Shapiro and Stone, 1994; Alter, 1996; Bellamy et al., 1996) and constitutional interpretation (Graber, 1994; Ball, 1995: 250-72).

What is the exact relationship between treaty reform and constitutionalization? Our argument rests on the recognition that there is a strong connection between the two processes, that, in fact, they are best viewed as a single meta-process of constitutional reform. Our main reason for arguing this is that treaties are far from being just the ‘frozen’ outcome of formal IGCs. Instead, treaties are undergoing continuous change (reform) also, and probably primarily between IGCs. Treaties are being interpreted, the meaning of provisions are being contested, discussed, or re-constituted. Sometimes certain provisions are simply ‘framework agreements’. This means that they are, in themselves, ‘empty’, and need to be given ‘substance’ in order to acquire a life of their own in order to be more than ‘dead letters’ in a treaty text.

Generally, treaties are given life through repetitive, constitutive and sometimes controversial processes of communicative action. Seen in this light, it is falsely formalistic to argue that the constitutionalization of the Community legal order does not constitute an important aspect of treaty reform. Whereas many legal scholars have known this for a long time, IR scholars seem to have something like an obsession with a certain version of (legal) formalism. Indeed, it is amazing to observe how neatly boundaries between interpretative communities follow boundaries between legal and political practice. IR-scholars in particular seem to have serious difficulties in recognising the existence of a European constitution. Instead, as we just saw, they know for "a fact" that there is no such thing.

How, then, can a link between treaty reform and constitutionalism be established? It may be necessary to emphasize that our structurationist point of departure implies that we do not a priori assume which agents and structures are prime players for the constitutional process. Instead, we seek to identify a cluster of agents which all have an impact on the constitutional process. In doing this, we depart from every possible unitary actor perspective on states, and we allow for unintended consequences of strategic interaction. Following Stone and Sandholtz (1997), we treat parliaments, courts, constitutional discourses and legal systems as institutions. Furthermore, we treat law professors, lawyers and judges as much as agents as those who are traditionally regarded as policy- and decision-makers.

Nevertheless, constitutional reform looks like a co-constitutive process that requires the involvement of both legal and political actors. So far the involvement has been markedly asymmetrical. It has predominantly been actors in the legal sphere which have reasoned in terms of constitutionalization. Political actors have been much more hesitant if not hostile to the idea, meaning that a European constitution resonates more in some quarters than in others.(9)

Critics might argue that our dual – political and legal – conception of treaty reform and constitutionalization is just a mirror image of the good old split between legal and political science perspectives on social affairs; that we should know better, because legal perspectives are just that, and therefore something which "enlightened" IR-scholars have left behind in the process of becoming liberated from legalism; that E.H.Carr already noted the impotence of legalism; that the soul of the IR-discipline is to have moved beyond legal scholarship; and so forth.

We are aware of such arguments, but in our view they lack validity for a number of reasons. First, because we do not defend legalism, and we do not have a weak spot for positive law. Second, because it seems to us that a large part of legal scholarship is more sophisticated than the average IR-scholar is ready to acknowledge. Third, because certain legal theories, and not the most modern, live their own quiet and secret life in various IR theories, from Hedley Bull to Kenneth Waltz. Fourth, just as a ‘law in context’ school has emerged (cf. its celebration by Weiler, 1997), we make a plea for an ‘IR in context’ perspective. In this respect we consider structuration theory to be a helpful aid in the creation of such a perspective (Christiansen, 1998). Finally we note that at least some strands of legal scholarship – those concerned with the construction of a system of written and unwritten rules and norms in the international realm are currently being reconsidered in the context of IR theorising (Onuf, 1989; Kratochvil, 1989).

IV. An Alternative View of ‘Intergovernmental’ Conferences: The Image of Summits and Valleys

According to this image IGCs are seen as important events in the history of EU treaty reform. A simple count provides us with the information that five major IGCs have taken place since the first in 1950-51 and the one concluded in 1996-97.(10) These five IGCs are the ‘mountain summits’. It should be added that we do not have problems with accounts that call the outcome of IGCs super-systemic decisions, to use Peterson’s (1995) vocabulary. It is very appropriate, and, generally, we do not dispute the traditional conception of IGCs. Our argument is rather that the isolation of IGCs creates an unnecessarily narrow conception of the role of IGCs, and by extension that a broadening of the perspective gives us a better understanding of the nature and function of EU treaty reform. How can such broadening look like? It seems to us that IGCs, i.e. the ‘summits’, should be analysed in connection with the ‘valleys’ between them. This means that we arrive at an image that includes four phases, i.e. the intermediate phases that are located between the five IGCs since 1950:

Figure 1

While we admit that IGCs are important events, we argue that IGCs can only be analysed in connection with their "base". In other words, events research and process research are complementary research designs. They are not competing in an analytical zero-sum game. Which type of connection between IGCs and base do we have in mind? Which ‘base’ are we talking about?

Social and material structures constitute part of the base, e.g. the acquis conferencielle, i.e. the revision procedure laid down by the treaties; the acquis communautaire is a social structure, or rather a cluster of social structures, of immense importance for European governance. Gialdino (1995) claims it useful to introduce a distinction between different types of acquis: i) the accession acquis; ii) the institutional acquis; iii) the Lomé acquis; iv) the EEA acquis; and the ‘untouchable hard core’ acquis. But the term has also been introduced in policy-specific areas. Thus, Wiener, talks about the acquis concerning citizenship, and others about the acquis politique (.e. the acquis in foreign policy) or the Schengen acquis.

A closer look at the process of constitutionalization which has occurred in the valleys – hidden from the view of the summit-watchers – indicates that governments have been unable to control treaty reform. Thus, what we are talking about is controlof IGCs as well as the control of institution-building and constitutional reform that happens between IGCs. In fact, this has been the case since the very inception of the Community. But already Monnet suggested that was difficult to maintain that governments were in full control of even the very first European integration IGC. Some significant ideas were introduced by actors outside governments, and other ideas and organizational designs had been around since the 1930s (Griffiths, 1997). Socialization and adaptation occurred throughout the conference.

Important things happened after the first IGC, in what turned out to be the first intermediate phase. As a result, the basic political-organizational superstructure of what became an institutionalized Europe was created, not during specific instances of treaty reform, but as part of a process during the intermediate phases between IGCs (Griffiths and Milward, 1986). This is an argument that we make in more detail, in the next section, for the most recent instance of treaty reform. But in order to show that this is a perception that holds true more generally for the European Union, we briefly look at processes of constitutional change in the past.

The vertical distribution of power between Member States and Union institutions – an issue of federal constitutionalization if there ever was one – was not settled at any one IGC. While the Paris and Rome Treaties ‘specified’ certain provisions regarding the relative weight of individual Member States and the Community, the arrangement which dominated this relationship for some 20 years resulted from an extra-treaty agreement among governments, the so-called Luxembourg Compromise. Every student of European integration will be aware of this feature which was meant to safeguard the national veto in the Council in the face of contrasting provisions in the treaty. Even though it was never espoused by the ECJ or became part of written law, it certainly was a constitutional settlement of fundamental significance.

Later IGCs responded to the inefficiencies created by the Luxembourg Compromise by specifying the areas in which qualified majority vote would become the norm. The Single European Act, in particular, must be regarded as such. The very nature of the ‘bargain’ struck in the IGC preceding the SEA – the linkage between market liberalization and institutional reform – cannot be understood if it is not seen against the background of 20 years of national veto in the Council.

In those decades, other devices had been added to the institutional structure of the Community, but had not arisen from the treaty. An obvious one to mention is the institution of the European Council. It is impossible to imagine the EU today without the European Council, which is so closely tied to the presidency of the Council of Ministers. But it is no secret that the European Council was not the result of a treaty change, but an informal and ad hoc response for the need for bargaining at the highest level. This, in turn, became necessary in the mid-1970s, because the agenda of the Community had widened to an extent that required the direct participation of the holders of highest national political office in EC decision-making.

The Luxembourg compromise, the European Council and the greater weight given to the Presidency had paradoxical effects on the vertical distribution of power in the Community. On the one hand, these devices were explicitly designed to strengthen the intergovernmental aspect of the Community and thus to preserve every Member State’s ability to set and control the agenda. On the other hand, this had the countervailing effect on the national level. Foreign Ministries, as managers of the Presidency programme, and Prime Ministers, as participants in the European Council, became increasingly exposed to, and eventually socialized into, the ‘Community method’ and the requirements of searching for and achieving consensus. Just as the EC became ‘re-nationalized’, national governments became increasingly ‘Europeanized’. In the process, they added weight and significance – and thus power – to European level decision-making. By holding regular ‘summits’ and institutionalising these, EC decision-making became, almost by default, an issue of high politics, even though, in substance, it might not have been regarded as such.

This is the EU equivalent of what in federal theory has been called the ‘Madisonian paradox’ (Dehousse, 1989) – the search of constituent units for greater influence in the process of central decision-making leading inevitably to central decision-making becoming relatively more significant, while sub-central autonomy inevitably suffers. Thus, at the beginning of the 1980s, the Community had gone through fundamental transformations, which saw both national and European institutions in substantially altered roles from the ones envisaged in the treaty – even though no IGC had been held and, on paper, nothing had changed.

This is no isolated instance of treaty reform outside the IGC – it is the normal way in which the European institutions have always been reformed. There is no space here to conduct a comprehensive survey, but a brief glance at the way in which two key policies – CFSP and EMU – became part of the EU’s portfolio will serve to confirm the argument. The development of a Common Foreign and Security Policy, now an integral part of the structure of the EU, began in the 1970s as ‘European Political Cooperation’ (EPC). For some 15 years EPC remained entirely outside the treaty and was therefore an informal process of co-ordination and cooperation among Member State foreign ministries. This gradual establishment of CFSP can hardly be understood as the outcome of so-called national interests of Member States. It makes more sense to regard the creation of EPC and its development into CFSP as the result of the sectoral interests among foreign ministry officials and ministers who had, over previous decades, lost considerable control over the integration process as sectoral ministries took over the expanding agenda of Community business. Later IGCs merely formalised the practices of European foreign policy-making that had already become established.

The creation of the single currency also demonstrates a pattern of gradual evolution. Key decisions which were to provide the basis for later ratification through the Maastricht Treaty where made in the course of setting up the European Monetary System (EMS) and the Delors Committee. The EMS was set up in the late 1970s and constituted the backbone of European economic cooperation for 15 years, thus laying the foundations of economic convergence and policy-learning in the field of monetary policy. On this basis, the Delors Committee, composed of central bank presidents and chaired by the Commission president, developed the blueprint of a single currency as it was later set out in the Maastricht Treaty. To be sure, issues of high political tension – such as the British opt-out – still needed to be resolved in the course of subsequent summits and IGCs, but we can say with hindsight that the key decisions regarding monetary union – the choice in favour of a single currency and the agreement about a ‘stability-oriented’ monetary policy – had been made before the IGC. In addition, a number of crucial issues – the precise degree of independence of the European Central Bank, the relationship between the ‘ins’ and the ‘outs’, the policing of member States’ deficit spending – all remained to be resolved after Maastricht. Clearly, the IGC was just one stepping stone on the way towards monetary union, and it does not seem to have been the most important one.

In conclusion, even this brief overview demonstrates that the traditional image of norm and exception – of over-powering mountain summits and insignificant intermediate valleys – needs to be reversed. The image of mountain summits is misleading since the crucial choices about the nature of treaty reform are not made at momentous IGCs, but in fact result from long-term processes. Our critique of the conventional wisdom that governments take key decisions about the nature of the integration process at IGCs rests on three conclusions.

First, Member State representatives spend as much – perhaps more – time shaping constitutional decisions outside as they do during IGCs. Corbett (1992: 271-2) notes that concerning negotiations about the future of the EC/EU, "Member States have been engaged in such a process in more years than not", which leads to Corbett’s highly interesting observation that "negotiations about the future direction of the EC is, in fact, the norm rather than the exception" (emphasis added).

Second, given that the EU’s constitutional choices are shaped outside the insulated meeting rooms of the IGC, a plurality of actors actively participate in the process. Much like the more mundane process of policy-making, treaty reform has spawned a network of actors, a body of highly ‘technical’ knowledge and purpose-made rules of procedure which include non-state actors – Commission officials, private interests, MEPs – as well as Member States’ representatives. Even the very question of who represents the Member States is being redefined in these circumstances, as sectoral ministers, constitutional courts, central bankers, local and regional governments become actively involved in the process. In short, admitting that treaty reform is not an event (with clearly defined membership) but a process (with fairly open access to a wide range of interests) is to admit that Member States loose control over the decision-making process.

Finally, we have seen that treaty reform is a heavily structured process. Actors – Member State representatives as well as non-state actors – are not free to make any agreement they see fit, but choose certain avenues among the limited options available. We can identify four such elements of structure. First, there is the pre-existing institutional order that can at best be tempered with. More usually, though, actors have simply accepted the emergence of a certain institutional development and used treaty reform to codify and add to the existing structure. Second, treaty reform has had to accept the Community legal order and the vast and constantly expanding acquis. Given that legal order and the significance of the judicial review exercised by the ECJ, agency has been tightly circumscribed. Thirdly, there are discursive structures that to a significant extent predefine what IGCs are about and can achieve. This includes both the language used in the preparation and justification of treaty reform and the public debate surrounding the integration process at large. Even though IGC proceedings are of limited transparency, a certain range of options is defined by the way in which treaty reform has been framed in the public domain. Finally, the nature of the economic order in Europe is a canvass on which treaty reform has to be painted. Economic structures include the changing nature of domestic political economies but more significantly the process of globalization that has driven European integration from the outset.

It is against the insights from the brief historical overview undertaken here that we now look at the Amsterdam Treaty, or rather the Amsterdam processes. We have found that, in the past, treaty reform

has been a process of long-term constitutional change which has mainly taken place in the periods between IGCs
has been driven by a plurality of actors which include Member State representatives as well as non-state actors
has been governed by legal, institutional, discursive and economic structures limiting the available choices.

We would expect that we find the same to be true for the case of Amsterdam, and that from that basis we can make more general conclusions about the nature of the process of EU treaty reform.

V. The Amsterdam Process

When was Amsterdam? In the early summer of 1997? Even the summit-watchers would agree that one needs to start a little earlier and consider the development of the actual Intergovernmental Conference. One issue of potentially great significance – the issue of flexible integration was – according to participants settled in 7 minutes at the end of the summit, yet the IGC had worked on the relevant passages for 15 months. With such examples it is not difficult to assume that most decisions were made at the IGC, not the actual Amsterdam European Council. But the IGC hardly started from a clean table. Governments who were keen to circumscribe the agenda of the IGC in a reasoned manner had previously instituted a so-called ‘Reflection Group’ whose report reflected, in effect, the negotiations of the subsequent IGC. But even the reflections of this Group were, by and large, predefined by an already existing agenda and the fact that the convening of a new IGC to deal with any unfinished business had been written into the Maastricht Treaty. The ‘result’ of the Amsterdam summit – the Amsterdam Treaty – was therefore not so much a reflection of decisions taken by politicians at the summit, but rather of the way in which the evolving agenda of the EU had been processed by EU and national officials in the preceding years.

But let us nevertheless briefly, just for the sake of argument, perceive Amsterdam as an event and review the ‘decisions’ that were taken there. Following after the argument made above this is obviously a hypothetical exercise, but one that we will execute in all honesty. Thus we can identify three major types of decisions taken at Amsterdam. These areas are, firstly, institutional reforms; secondly, the introduction of ‘flexible integration’; and thirdly, adjustments made to the contents of the three pillars. A brief look at each of these will serve to identify the degree to which the above conclusions about treaty reform are confirmed in this case.

Institutional reform consisted of changes for each of the three legislating institutions. The EP was given extended powers in a range of areas. Was this a decision made at Amsterdam? Or was it rather the logical result of a process that began with the introduction of direct elections in 1979 – yet another instance of a fundamental change to the European constitution decided outside any IGC – and carried on through similar instances of gradual upgrading of parliamentary powers at the previous two IGCs. An expanding number of inter-institutional agreements has already tied the EP into the everyday workings of the Union. The continuing discourse about the democratization of the EU and its closeness to the citizens did its part to ensure that there was never any doubt that an improvement of the EP’s institutional position would be on the agenda for, as well as in the Treaty of, Amsterdam. The final shape of the reform will only become clear, as it did with the changes effected at Maastricht, once the EP gets down to testing the limits of what it can, and cannot, do with its new powers.

The key ‘decision’ about the Commission – to limit its size in the face of enlargement – was even more obviously a non-decision. It has been linked explicitly to the number of new Member States joining the Union at the next round of enlargement. By indicating the number of Member States (20) at which each state will get to nominate only one Commissioner, the IGCs has effectively left the final decision to wider changes dominated by other issues (since it is inconceivable that one of the larger states will veto the 20+ enlargement to protect its second Commissioner post). The way in which control over this institutional change is slipping out of the hands of the Member States is evidenced by the progress on enlargement.

The Commission – which had itself proposed the reduction of Commissioners to one per Member State – proposed in its Agenda 2000 to start enlargement negotiations with six applicant countries. This would bring the total membership of the Union to 21 – something which may not be a coincidence. Governments are neither formally nor informally in control of enlargement negotiations. The ‘final decision’ on enlargement requires parliamentary approval, and preparations for, and negotiation of, membership requires such an amount of technical information that the Commission, more than any other actor, is in the driving seat here.

The decision on voting in the Council of Ministers, which – like the size of the Commission – was seen as an essential reform in view of enlargement, was even more comprehensively ‘not decided’ at Amsterdam but deferred to a later date. It was mentioned in the Treaty, and a future settlement was envisaged, so that no further ceremonially "full-blown" IGC will be needed to settle this issue – a treaty change agreed by a minor non-ceremonial IGC on the fringes of the Council or European Council might do the job. This is perhaps the clearest example to date that ceremonial IGCs fail to provide the forum in which constitutional issues of this complexity could be settled.

A second major change was the introduction of ‘flexible integration’ (Edwards and Philippart, 1997). The provision is designed to permit groups of Member States to go ahead with further integration in the absence of general agreement to develop the EU further. Towards this aim they may, under certain circumstances, use the institutions of the EU. In many ways this is the reverse of opt-outs that had been one of the features at Maastricht. At Maastricht (and subsequently at Edinburgh), the majority of Member States keen to move integration forward and frustrated with the obstinacy of a small minority, had to accept a departure from the previous principle of universal application of EU rules and decisions. In important areas such as social policy or monetary union, agreement on further integration was only possible after Member States received the right to ‘opt-out’. At Amsterdam, the tables were turned and the pro-integrationist Member States adopted the possibility of selected ‘opt-in’ in the form of flexible integration – a development initially opposed by those Member States left behind. But the practice – a Europe characterized by ‘multiple speeds’ or of ‘variable geometry’ – had been a reality established long time before and remained essentially the same.

Flexible integration implies the unintended consequence of at least a number of Member States (non-opt-ins) loosing actual control over, and influence on, the integration process. The governments in question were naturally keen to prevent flexible integration and the resultant loss of control over the future development of the integration process, but – as the principle of differentiated integration had been established earlier – this proved impossible. Institutionalized flexibility, then, can be seen as a substitute for formal treaty reform.

Finally, an important feature of Amsterdam had been the move of substantial parts from pillar three to pillar one, and the incorporation of the Schengen agreement and its acquis into the EU structure. As with EPC in the SEA and the recognition of the WEU in Maastricht this was essentially the codification of an already existing aspect of integration among a group of governments. Like the EMS and the WEU, Schengen had been developed outside the Community framework, by some governments, but was, at the same time, closely linked to the mainstream integration process. Bringing it into the treaty – while at the same time confirming the opt-out for those governments (UK and Ireland) that wished to remain outside it – was little more than the formal culmination of a process that had begun in 1985.(11) This is not to deny that important decisions on the details of this move – for example the incorporation of the Schengen acquis into the acquis communautaire – were not predestined, but were actually taken by governments in the course of negotiations. In this case, as in general, our argument is not that the shape of ‘intergovernmental’ agreements is wholly determined by structural factors, but merely that it is the interplay between governmental (as well as supranational) agency and elements of structure which explains the nature of treaty reform.

We can see now that the Amsterdam Treaty did not constitute the ‘moment’ at which governmental representatives took the decisive steps towards further integration. At most, they bargained over the way in which existing trends in the integration process – the increasing significance of the European Parliament, the achievement of greater efficiency in the Commission and the active use of differentiated integration – were formally incorporated into the treaty.

VI. Conclusions

This paper is about the development of an alternative research strategy on treaty reform. Towards that aim we have introduced a theoretical perspective that enables rather than constrains novel interpretations. A structurationist perspective along the lines presented here does not prejudice one set of actors above any others. A focus on the procedural nature of treaty reform has brought into view a much wider range of actors than is traditionally studied in state-centric analyses. The significance of their role in the treaty reform process is a matter of empirical research – but we believe this paper has shown that missing out on such an avenue for empirical research is a blind spot of intergovernmental approaches.

An emphasis on ‘process’ puts the negotiation of treaty changes into the wider perspective of constitutional reform. And while traditional analyses regard IGCs as the ‘summits’ of treaty reform, a wider focus reveals that important developments actually occur in the ‘valleys’ of constitutional reform. On occasion, IGCs make important changes to the treaty structure of the Union, but more often they merely codify changes which have already occurred. The constitutionalization of the Union – the most significant aspect of treaty reform – has taken place away from the ‘intergovernmental’ negotiating table, in the depths of the ‘valleys’ between IGCs. The same is true for the other key institutional features of the Union: the institutionalisation of regular meetings of heads of state and government, the development of mechanisms of foreign policy co-ordination, the growing significance of the Presidency, the empowerment of the European Parliament and the introduction of differentiated integration through selected ‘opt-outs’ and ‘opt-ins’.

Crucially, a structurationist approach allows us see the link between these ‘summits’ and ‘valleys’ as a continuous and open-ended process of treaty reform, structured not by the interests and resources of Member States, but by the social institutions that are established and reproduced in the course of reform. This does include Member State agency, and the arguments presented here should not be misunderstood as an attempt to replace the representation of national interests with structural determinancy. In the perspective adopted here, agency, including that from national governments, continues to matter and to make a difference. Indeed, this sensitivity to patterns of agency and their ability to change the course of events is what distinguishes structuration theory from straightforward structural approaches. But, compared to traditional studies of treaty change, our perspective leads us to see governmental agency in the context of a structural environment – the trajectory of past decisions, the multilateral generation of reform agendas, the institutionalised patterns of negotiation and decision-making, the constitutionalisation of the EU order – which severely compromises the ability of national governments to negotiate on the basis of their ‘national interests’.

After the Amsterdam Treaty, the European Policy Center has concluded that

the existing process of intergovernmental conferences to amend and adapt the Union’s Treaties has reached the end of its useful life. The EU cannot afford the repetition of a protracted process of intergovernmental negotiation followed by the anti-climax of negative political conclusions drawn at the end of the day (1998).

But what if an IGC to end all IGCs is convened? It would obviously be a zenith of European constitutionalism, but it would equally be the end-product of a very long constitutional process. This is a hypothetical question: there are, presently, no signs of any such event – it simply is not in the nature of the evolutionary process of European integration.

Instead, Amsterdam has confirmed what turn out to be general principles of EU treaty reform: IGCs are not the key events in treaty reform as they are often hailed; they are not isolated events but embedded in a wider process; and national governments are not in control of this process. The more important phases of treaty reform, conceived as constitutional reform, occur during the intervening phases between IGCs – it is in those periods that the crucial decisions about the shape of a European constitution are taken. This includes not only agenda-setting and implementation – which, as we know from the study of public policy-making, are as important as the actual process of decision-making – but also the very taking of decisions and creation of institutional structures.

If anything, Amsterdam, with its high incidence of ‘non-decisions’, has further formalised the way in which governments have accepted their inability to control the process. In a number of key instances the actual ‘decision’ about institutional reform has been left explicitly to small ‘technical’ IGCs or to non-IGC venues and times. Flexible integration is perhaps the best example to date to demonstrate that constitutionalization will not be settled for a long time – if at all it is possible to imagine a final solution in constitutional politics. In permitting a small group of governments to press ahead, using the Union’s institutional structure and thus developing it further, it has been made abundantly clear that individual states have lost control over the direction and final destination of the process.

In leaving, now formally, constitutional reform to the intermediate phases between IGCs (and reducing IGCs to mere ratification and legitimation exercises), national governments have also accepted that other actors will come in and influence the process (Christiansen and Jørgensen, 1998). Perhaps more interestingly yet, they have also accepted the significance of structural changes which cannot be anticipated. The Amsterdam process is a prime example of how actors bowing to the limitations which legal, discursive and economic structures place on their potential for action.

Finally, on a sobering note, one may ask whether the Amsterdam Treaty even serves as an attempt to addressing the constitutional issues that were called for. It was an IGC that was explicitly meant to prepare the Union for the process of enlargement. But not much research is required to note that enlargement – just like monetary union – will introduce a lot of further uncertainty into European constitutionalization. Both are designed as long-term processes with a significant number of constitutional issues to be settled in due course. Both processes require responses from the Union about the relationship between its parts and its whole, about its commitment to redistribution, about the accountability of its central institutions and about the way in which ‘history-making’ decisions are being made. With regard to EMU one only needs to point to the potential disputes between the ‘ins’ and the ‘outs’ to recognize that crucial decisions about the nature of the post-EMU EU still need to be taken. The very existence of an EMU Council had not been decided at Maastricht, nor was its potential membership being discussed. The EMU dispute about the degree to which the ‘outs’ may participate in the decision-making among ‘ins’ is only a first taste of the issues which will result from the active use of flexibility. The EMU experience demonstrates that these constitutional choices are not made at IGCs, but in the ‘valleys’ between them.

The decisions relating to enlargement, the conditions of and negotiations for membership as well as the institutional adaptation of the Union were also not taken at Amsterdam. As with EMU, the big choices were presented to the Union after the IGC. The Commission, in publishing its Agenda 2000, set the agenda and has to a significant degree structured later developments. The 1997 Luxembourg summit attempted a partial departure from the proposal of the Commission – papering over the distinction between the first wave of entrants and the rest – but has also found it hard to leave the trajectory of further developments envisaged by the Commission. And the red-green government elected in Germany in 1998 made it its Presidency goal for the first half of 1999 to conclude the reforms set out in Agenda 2000 – an ‘agenda’ proposed by the Commission and generally seen as a precondition for enlargement. In it are contained the fundamental changes to the structures of the Union which one could have expected from an IGC.

If we compare the twin processes of EMU and Agenda 2000/enlargement on the one hand, and the treaty reform undertaken at Amsterdam on the other, we see how in the 1990s even more than previously the fundamental choices about the nature of the integration process are made outside the formal venue of IGCs. Given the significance of both enlargement and monetary union, it is really nothing short of extraordinary that the event ‘Amsterdam’ did not succeed to engage in a serious reform of constitutional issues resulting from these processes. The only way of making sense of this development is the recognition that governments have now accepted formally that treaty reform does not occur during but in-between IGCs. That in itself is a constitutional settlement, one that has now been codified by the 1996-97 IGC.

References

Alter, K. (1996), ‘The European Court’s Political Power’, West European Politics, 19(3): 458-487.

Ball, T. (1995), Reappraising Political Theory. Revisionist Studies in the History of Political Thought, Oxford: Clarendon Press.

Bellamy, R. and Castiglione, D. (eds.), (1996), Constitutionalism in Transformation. European and Theoretical Perspectives, Oxford: Blackwell.

Christiansen, T. and K.E. Jørgensen (1998), ‘Negotiating Treaty Reform in the European Union: The Role of the European Commission’, International Negotiation, 3(3): 1-18.

Christiansen, T., (1998), ‘Bringing Process Back' In: The Longue Durée of European Integration’, Journal of European Integration, 21(1): 99-121.

Cohen, I.J. (1989), Structuration Theory. Anthony Giddens and the Construction of Social Life, Basingstoke: Macmillan.

Corbett, R. (1992), ‘The Intergovernmental Conference on Political Union’, Journal of Common Market Studies, 30(3): 271-298.

Curtin, D. (1993), ‘The constitutional structure of the Union: A Europe of bits and pieces’, Common Market Law Review, 30(1): 17-69.

Dehousse, R., (1989), ‘Il Paradosso di Madison: Reflessioni sul Ruolo delle Camere Alte nei Sistemi Federali’, Le Regioni, 17(5): 1365-1400.

Dehousse, R. and Majone, G., (1994), ‘The Institutional Dynamics of European Integration: From the Single Act to the Maastricht Treaty’, in S. Martin (ed.), The Construction of Europe, Dordrecht: Kluwer Academic Publ., pp. 91-112.

Edwards, G. and E.Philipart, (1997), Order and Flexibility in the European Union, paper prepared for the Centre for International Studies Roundtable, Cambridge.

Edwards, G. and A. Pijpers (eds.)(1997), The Politics of European Treaty Reform: the 1996 intergovernmental conference and beyond, London: Pinter.

European Policy Centre, (1998), Making Sense of the Amsterdam Treaty, Brussels.

Forster, A., (1998), ‘Britain and the Negotiation of the Maastricht Treaty: A Critique of Liberal Intergovernmentalism’, Journal of Common Market Studies, 36(3): 427-468.

Garret, G., (1992), ‘International cooperation and institutional choice: the European Community’s internal market’, International Organization, 46(2): 533-560.

Gialdino, C.C. (1995), ‘Some Reflections on the acquis communautaire’, Common Market Law Review, 32: , pp. 1089-1121.

Giddens, A., (1984), The Constitution of Society, Cambridge: Polity Press.

Graber, M.A. (1994), ‘Why Interpret? Political Justification and American Constitutionalism’, The Review of Politics, 56(4): 416-440.

Gregson, N. (1989), 'On the (ir)relevance of structuration theory', in D.Held and J.B.Thompson (ed.), Social theory of modern societies: Anthony Giddens and his critics (Cambridge: Cambridge UP), pp.235-248.

Gregson, N. (1987), 'Structuration theory: some thoughts on the possibilities for empirical research', Environment and Planning D: Society and Space, 5, 73-91.

Grieco, J., (1991), The Renaissance of the European Community and the Crisis of Realist International Theory, Working Paper No.151 of the Duke University Program in Political Economy, Duke.

Griffiths, R.T. (1997), ‘The Creation of European Supranational Institutions’, in K.E. Jørgensen (ed.), pp. 87-106.

Griffiths, R.T. and Milward, A. (1986), ‘The Beyen Plan and European Political Community’, in W. Maihofer (ed.), Noi si mura, Florence: European University Institute, pp. 596-621.

Hollis, M. and Smith, S. (1991), ‘Beware of Gurus: structure and action in international relations’, Review of International Studies, 17: 393-410.

Hurrell, A. and Menon, A. (1996), ‘Politics like any other? Comparative politics, international relations and the study of the EU’, West European Politics, 19: 386-402.

Jørgensen, K.E. (1997), Reflective Approaches to European Governance, London and New York: Macmillan and St Martin’s Press.

Keohane, R. and Hoffmann, S., (1990), ‘Conclusion: Community Politics and Institutional Change’, in W.Wallace, The Dynamics of European Integration, London: Pinter, pp. 276-300.

Kratochvil, F., (1986), ‘Of Systems, Boundaries and Territoriality: An Inquiry into the Fornmation of the State System’, World Politics, 39( 1):: 27-52.

Layder, D., (1987), 'Key Issues in Structuration Theory: Some Critical Remarks', in J.Wilson (ed.), Current Perspectives in Social Theory (London: Jai Press)., pp.25-46.

Marks et al., G., (1996), ‘European Integration from the 1980s: State-Centric v. Multilevel Governance’, Journal of Common Market Studies, 34( 3): 341-379.

Matlary, J.H. (1993), ‘Beyond Intergovernmentalism: The Quest for a Comprehensive Framework for the Study of Integration’, Cooperation and Conflict, 28(2): 181-210.

Mattli, W. and Slaugther, A.-M.., (1995), ‘Law and politics in the European Union: a reply to Garret’, International Organization, 49(1): 183-190.

Moravcsik, A., (1993), ‘Preferences and Power in the European Community: A Liberal Intergovernmentalist Approach’, Journal of Common Market Studies, 31(4): 473-523.

Moravcsik, A., (1994), Why the European Community Strengthens the State: Domestic Politics and International Cooperation, paper prepared for the Conference of Europeanists, Chicago.

Onuf, N., (1989), World of Our Making: Rules and Rule in Social Theory and International Relations, University of South Carolina Press.

Pescatore, P. (1981), ‘Aspects judiciaires de l’acquis communautaire’, Revue trimestrielle de droit européen, 20: 617-651.

Peterson, J. (1995), ‘Decision-Making in the European Union. Towards a Framework for Analysis’, Journal of European Public Policy, 2(1): 69-93.

Shapiro, M. and Stone, A. (1994), ‘The New Constitutional Politics of Europe’, Comparative Political Studies, 26: 397-420.

Snyder, F., (1990), New Directions in European Community Law, London: Weidenfleld and Nicolson.

Stone, A. and Sandholtz, W., (1997), ‘European integration and supranational governance’, Journal of European Public Policy, 4(3): 1-24.

Stone, A. (1994), ‘What is a supranational constitution? an essay in the international relations theory’, The Review of Politics, 56(3): 441-474.

Stones, R. (1991), 'Strategic Context Analysis: A New Research Strategy for Structuration Theory', Sociology, 25, 673-695.

Temple Lang, J. (1998), ‘The Duties of National Authorities under EC Constitutional Law’, European Law Journal, 23(2): 109-131.

Tsebelis, G. (1994), ‘The Power of the European Parliament as a Conditional Agenda-Setter’, Americanm Political Science Review, 88(1): 128-142.

Wallace, W. (1990), Dynamics of European Integration, London: Pinter.

Waltz, K. (1979), Theory of International Politics, New York: Random House.

Weiler, J.H.H. (1993), ‘Neither Unity nor Three Pillars – The Trinity Structure of the Treaty on European Union’, in J.Monar, W.Ungerer, and Wessels (eds.), The Maastricht Treaty on European Union, Brussels: European Interuniversity Press, pp. 49-62., pp. 49-62.

Weiler, J.H.H. (1991), ‘The Transformation of Europe’, Yale Law Journal, 100: , pp. 2403-2483.

Weiler, J.H.H. (1997), ‘The Reformation of European Constitutionalism’, Journal of Common Market Studies, 35(1): 97-131.

Wendt, A. (1987), ‘The Agent-Structure Problem in International Relations Theory’, International Organization, 41:335-70.

Wendt, A. (1991), ‘Bridging the theory/meta-theory gap in international relations’, Review of International Studies, 17: 383-392.

Wendt, A. and Shapiro, I. (1997), ‘The Misunderstood Promise of Realist Social Theory’, in K.R. Monroe (ed.), Contemporary Empirical Political Theory, Berkeley: University of California Press, pp. 166-187.

Westlake, M. (1998), The European Union’s ‘blind watchmakers’: the process of constitutional change, in M. Westlake (ed.), The European Union beyond Amsterdam: New Concepts of European Integration, London and New York: Routledge.

Wincott, D. (1995), Institutional Interaction and European Integration: Towards an Everyday Critique of Liberal Intergovernmentalism, Journal of Common Market Studies, 33:597-610.

Wind, M. (1997), ‘Rediscovering Institutions: A Reflectivist Critique of Rational Institutionalism’, in K.E. Jørgensen (ed.), pp. 15-35.

Zimmerling, R. (1991), Externe Einflüsse auf die Integration von Staaten – Zur politikwissenschaftlichen Theorie regionaler Zusammenschlüsse, Freiburg/München: Verlag Karl Alber.

Endnotes

(*) A previous version of this paper was presented at a panel at the British International Studies Association's annual conference in December 1997 and at a meeting in Brussels in May 1998. We would like to thank Geoffrey Edwards and Eric Philippart (who directed both meetings), and all participants in the two meetings for constructive suggestions. We are also grateful to Mark Gray, Amy Verdun, Jeppe Tranholm-Mikkelsen, Lars Bo Kaspersen, for detailed comments on an earlier draft. We would also like to thank the two anonymous referees of the EIoP for their helpful comments.

(1) For a more balanced approach, see Edwards and Pijpers (eds.) 1997.

(2) For a similar view, see Westlake 1998.

(3) For critiques of intergovernmentalism in European integration theory, see for example Dehousse and Majone (1994), Forster (1998), Mattli and Burley (1995), Matlary (1993), Wincott (1995), Marks et al. (1997) and Wind (1997).

(4) Structuration theory has been introduced to the field of International Relations by Wendt (1987) and Carlsnaes (1992). For a very useful exchange of views on the value of structuration theory, see Hollis and Smith (1991) and Wendt (1991). On the use of structuration theory in the context of European integration, see Snyder (1990) and Christiansen (1997). On the relationship between scientific realism and structuration theory, see Wendt and Shapiro (1997).

(5) http://europa.eu.int/en/agenda/igc-home/.

(6) For a discussion of the use of external pressures on the integration process as a determining influence see Zimmerling (1990).

(7) Interviews with officials from Member State administrations, Council Secretariat and Commission who participated in the negotiations leading to the Amsterdam Treaty, Brussels, April 1998.

(8) In what became an infamous draft in preparation for the Maastricht summit, the then Dutch Deputy Foreign Minister Piet Dankert proposed a Treaty which was widely perceived as too 'federal' and which was dismissed by the IGC in favour of the previous Luxembourg Draft – a step backwards both in terms of the progress made during the conference and in terms of Dankert's own federalist ambitions.

(9) We do not want to create artificially watertight dividing-lines between 'legal' and 'political' actors. Clearly the ECJ is to some degree also a 'political power' (Alter, 1995), while Member State governments sometimes find it hard to refrain from playing the part of judge, as the example of the Barber Protocol demonstrates (Curtin, 1993; Gialdino, 1995). What we can say is that political actors like the governments and the European Parliament have not recognized – at least not explicitly – that the treaties constitute a European constitution. But there certainly is a tacit understanding that the Union's legal order has been constitutionalized by the ECJ and by the interpretative legal community.

(10) Formally, ten IGC's have taken place between 1961 and 1997: Extending the area of territorial implementation of the association system of the countries and territories to the Dutch West Indies, 13 November 1962 (OJ L 150, 01.10.64); The Treaty of Brussels fusing the executives of the Commission and Council, 8 April 1965 (OJ 152, 13.07.67); Treaty of Luxembourg, 22 April 1970 (OJ L 2, 02.01.71) – Budgetary measures; Treaty of Brussels, 10 July 1975 (OJ L 91, 06.04.78) – Budgetary measures; Amendment of the status of the European Investment Bank, 22 July 1975 (OJ L 359, 31.12.77); Exclusion of Greenland from the territory of the Community, 13 March 1984 (OJ L 29, 01.02.85); Single European Act (OJ L 169, 29.06.87); Treaty on European Union, 7 February 1992 (OJ L 224, 31.08.92); Treaty of Amsterdam, 2 October 1997 (OJ C 340, 10.11.97). The Treaty of Paris (1951), Treaties of Rome (1957), Treaty of the Saar (1958) and the Convention on certain Institutions common to the European Communities (1958) were modified without recourse to the procedure laid down in the treaties and were thus not formal IGC's. (Article 96 ECSC, Article 236 EEC). We are very grateful to Mark Gray for providing this information.

(11) See Weiler (1993) for an argument about the permeability of the pillar structure of the Maastricht Treaty.

Contents: