Abstract. Early studies on European socialisation hypothesised that social interactions crosscutting national borders and transcending the national level lead actors to shift their allegiances towards the European level. Although this is a key question for many studies on European socialisation, this paper argues that this hypothesis ignores that actors are embedded not in one, but in multiple social environments. This paper attempts to go beyond this so-called contact hypothesis; it tries to take into account the actors� multiple embeddedness and, based on quantitative interview data, it investigates under which conditions supranational role conceptions are adopted. In contrast with the idea that intensive Euro-level interactions cause actors to adopt supranational role conceptions or to supplement intergovernmental roles with supranational roles, it is shown that an extensive exposure to the European level does not lead to more supranational role conceptions. On the contrary, it seems that various domestic experiences seem to affect role conceptions more substantially than Euro-level experiences. �

Introduction

Socialisation is an important theme in many early and contemporary writings on European integration and European policy-making. It is central in the transactionalist approach � la Deutsch (�partial identification in terms of self-images and interest�), functionalism � la Mitrany (�functional loyalties from the experience of co-operation itself�) and neo-functionalism � la Haas (�a shift of loyalties, expectations, and political activities towards a new centre�). Recently, within the ranks of empirically oriented constructivists the socialisation theme has regained attention. Socialisation is the process by which views and practices are internalised (Checkel 1999; Risse and Wiener 1999) . It affects how actors see themselves and what they consider as their interests and duties. Key for a socialisation process to occur is the involvement of actors in existing, new, changing or transforming social environments. In the context of European integration, which can be considered as a process changing and transforming the European nation-states, it is then hypothesised that �Europe� causes a redefinition of political views and practices or �Europe� should become �internalised�. More in general, European socialisation refers to the adaptive learning processes of national organisational structures, and the individuals representing these, to a changing, or changed, and increasingly Europeanised environment.

 

Early neo-functionalist studies on European socialisation hypothesised a positive effect of �Europe�. Social interactions crosscutting national borders should lead to more sophisticated cognitive views, more favourable affective dispositions towards other nations and a more European orientation. Domestic political actors gradually socialise into the shared norms and practices of the EU, a process leading to a diffusion of those norms and practices within various national systems. Important in this approach is that views and practices are not fixed or exogenously given; views and practices are contingent, they change or transform as a result of changing contextual factors. Although highly divergent in many respects, some general observations regarding the older empirical literature on European/international socialisation are noticeable in this context. As some constructivists today, many of the writings in the seventies related their ideas on socialisation to social-psychological studies on attitude change and the first empirical studies were constrained by similar methodological problems as we face today. And it is worth noting that most early studies concluded that attitude change as a result of involvement in European/international settings, or contact, was very limited, non-existent or, due to methodological problems, difficult to grasp (Kerr 1973; Peck 1979; Scheinman and Feld 1972; Smith 1973; Wolf 1973) .

 

Also the recent literature remains highly contradictory regarding these matters. On the one hand, some argue that national officials involved in EU-level policy-settings, e.g. Council working parties (CWPs hereafter), are exposed to a spirit of co-operation and mutual understanding, an esprit de corps (Hayes-Renshaw and Wallace 1997; Wessels 1991; Westlake 1995) . In addition, it is claimed that, given the functional fragmented character of the Council, territorial bounded role conceptions should loose importance and that specialised expertise-based identities are becoming more preponderant. And in addition, the role of supranational institutions - the European Commission and the European Parliament - or the reliance on majority voting tends to diminish national coloured views and practices. On the other hand, others argue that such a loyalty transfer does not take place (Pollack, 2001), that it is substantially moderated by domestic institutional factors (Hooghe, 1999) and that �the identity evoked in EU level settings does not replace the identity evoked in national institution� (Egeberg 1999: 457) .

 

Most of these competing claims are still under-researched or as Egeberg puts it: �leading scholars in the field express quite opposite views on the issue without any empirical reference at all� (Egeberg 1999: 457) . What is more, many studies look for socialisation at the level of the state or state elites in aggregated ways and do not �go micro�, as Johnston writes (Johnston 2001: 505-506) . As a consequence the literature has remained somewhat disconnected and isolated from existing studies and the methodological rigour in the social psychological literature. Broadly speaking, this literature investigates to what extent and under which conditions relational density or frequent face-to-face contact leads to the emergence of more homogeneous and consensus-oriented attitudes and behaviour among social agents (Friedkin 1984; Linssen, Hagendoorn, and Matheuse 1996; Stephan 1987) .

 

This paper aims to add some empirical grounded and methodological sound knowledge to this debate. It also tries to control the potential effect of �Europe� for other, especially domestic-level, factors. More in concrete, we focus on the adoption of role perceptions among national officials involved in 89 Council working parties (CWPs hereafter). The end result of the European socialisation processes taking place in the CWPs is conceived as a change in beliefs regarding how one should act in the CWPs. Two role conceptions, a supranationalist one and an intergovernmentalist one, will be distinguished and investigated. It is not a-priori claimed that Europeanisation necessarily replaces or rejects national allegiances; whether supranational conceptions replace, supplement or rather contradict existing intergovernmental conceptions is partly a matter of empirical investigation. But it is not only an empirical issue. European socialisation not necessarily concerns the adoption of supranational role conceptions at the expense of intergovernmental conceptions or the mere shift of one category to another; it might also imply a re-definition or a re-conceptualisation of existing intergovernmental conceptions. Officials continue to define themselves as national representatives and they increasingly conceptualise their national interests as overlapping with a European common interest. Although theoretically such connection between role conceptions can be imagined, our research strategy does not tie both dimensions together. As a null-hypothesis, it is scientifically, we believe, more fruitful to consider both dimensions as separate empirical categories and then explore the correlational pattern.

 

In sum, this paper investigates the conditions that enhance or inhibit the emergence of specific role conceptions. Conditions referring to the CWP features (e.g. density of meetings, voting within the working groups�), current and previous professional experiences at both the EU- and the domestic level, and national contextual factors (e.g. national elite orientation�) are discussed as independent variables influencing the adoption of role conceptions. The next section presents a brief discussion on how these conditions fit in a more general theoretical argument regarding Europeanisation. It is argued that Council negotiators are embedded not in one, but at least in two social environments, a domestic environment and a European environment. The second section outlines the research-design and reviews the dataset. The paper�s empirical corpus is presented in three separate sections. First, by looking at recruitment processes we explore current involvement in CWPs, European/international professional experiences and domestic professional experiences of the CWP-officials. It is illustrated that the recruitment of those officials strongly involved in CWPs (so-called full-timers) is considerably different from the recruitment of those that are less intensively involved (so-called part-timers). Then, the measurement of role conceptions is discussed and it is explored how actors mix both role conceptions. The third empirical section delves deeper into the conditions leading actors to adopt supranational role conceptions. The evidence suggests that especially low levels, and not high levels, of involvement cause actors to supplement both role conceptions. It is also illustrated that how administrative actors think they should act at the European level cannot easily be disconnected from domestic factors. Instead of Euro-level factors, especially specific domestic experiences seem to play a role.

 

Theoretical issues and hypotheses to be tested

Neo-functional studies on European socialisation hypothesised, painting with a broad brush, that social interactions crosscutting national borders lead to a more European orientation. Involvement in European settings gradually socialises actors into European norms and practices, a process leading to the diffusion of those norms and practices within various national systems and a shift of loyalty to the European level. This traditional notion is the starting point for this paper. To this hypothesis an institutionalist perspective is added, which means that role conceptions are affected by the actors� multiple institutional embeddedness. Embeddedness implies a relational perspective on actors� views and practices (Granovetter 1985) ; what actors do and believe are not just single-actor-level choices, but may be affected by multiple relationships, exchanges and actor-constellations. For instance, in order to understand something about the European role conceptions of national officials we should not only pay attention to Euro-level factors, but we should control for their domestic political experiences as well. More in general, changing views and practices might be constrained or enabled by a myriad of institutional factors and, consequently, an involvement in one institutional setting does not necessarily result into more sophisticated cognitive or favourable affective dispositions. On the contrary, it might be that, because of previous socialisation experiences actors are constrained or not encouraged to adopt such new views. In many circumstances, changing roles is both difficult and unattractive - difficult because previous socialisation leads to habituation and inhibits actors to depart from initial practices; unattractive because change usually entails increased short-term costs and risks.

 

Here, it is argued that multiple embeddedness, and more precisely domestic experiences, might inhibit or enable the adoption of supranational role conceptions. Domestic socialisation experiences might be consistent or fit well with such supranational role conceptions so that actors shift their allegiances faster. In such a case the explanatory power of EU-level factors is not necessarily more important than domestic level factors. It is not only consistency between Euro-level experiences and domestic experiences that might ease the adoption of supranational role conceptions. Actors can also have bad experiences with the domestic realm (e.g. inter-organisational conflicts, mismanagement, weak co-ordination devices�), try to escape from this, �go native� and defend supranational solutions. Likewise, experiences with Europe can inhibit European socialisation. Increased familiarity can lead to more knowledge and inside information, cognitive sophistication, which does not necessarily lead to more positive affective dispositions (Egeberg 1999: 460; Pollack 1998) . Crucial is that actors bring with them attitudes and dispositions formed as a result of socialisation in multiple social environments. Socialisation processes encompassing learning that takes place prior to the focussed moment, here involvement in European affairs, should be taken explicitly into account.

 

Three institutional arguments lead us to expect no, or a rather weak, European socialisation (Harmsen 1999) . First, the �stickiness� of domestic institutions makes actors resistant for change and adaptation and this �stickiness� does not fade away because of Europe. Therefore, socialisation is moderated through institutional, political and cultural factors varying among the member states. Second, domestic institutions emanate existing power balances and modes of policy-making, which tend, in Bourdian terminology, to reproduce themselves in the face of new challenges. Or, appropriate ways of policy-making and existing patterns of inclusion/exclusion prevail instead of a fundamental transformation of politics. And thirdly, when actors become involved in new settings, they search for familiarity, i.e. they rely on already existing routines and practices. Thus, domestic actors �domesticate� the European integration process by employing a familiar repertoire of institutional practices.

 

These arguments have a �bottom up� character, that is, they focus on factors situated at the domestic level (B�rzel 1999) . Other studies adopt a �top down� approach. Institutional features situated at the European level constrain or enable European socialisation. Just as domestic institutions aim at reproducing particular views and practices, also European or international institutions may socialise actors into a specific set of roles. However, theories differ whether or not European experiences should lead to the adoption of supranational or intergovernmental role conceptions. Intergovernmental �top down� approaches are stressing the constraints, while neo-functional approaches are pointing to the enabling factors.

 

Intergovernmentalists argue that cross-bordering interactions sustain territorial allegiances and that several Euro-level institutional factors reinforce this (Hoffmann 1966; Hoffmann 1982) . The Council, which is the body par excellence via which national politics interacts with the EU-level, imposes a nationality label upon domestic officials participating in it. Here officials are delegated negotiators. They represent national views and their roles are predominantly intergovernmental roles; European policy co-ordination is primarily a bargain between member states representatives. Supranational institutional actors as the Commission or the Parliament facilitate these inter-state bargains; they have no autonomous impact. Although variance regarding role conceptions is expected to be low, intergovernmental role conceptions should dominate, existing variance should reflect that the main cleavage in the Council concerns a cleavage between opposing intergovernmental and supranational roles. Territorial, not organisational, functional or ideological, allegiances are supposed to predominate the CWPs.

 

While intergovernmental �top-down� theories expect �convergence� in the sense of strengthening the national executives, the preponderance of domestic socialisation sites and intergovernmental role conceptions, neo-functional approaches expect the opposite. Involvement in CWPs has a strong socialisation effect and induces a supranational role conception on officials. Or, at minimum, it leads to a substantial re-definition of existing intergovernmental role conceptions in the sense that these are supplemented by supranational roles. The spirit of cooperation and mutual understanding to which CWP-officials are exposed, it is argued, makes that they become more capable to go beyond the national interest and start to adopt a common European perspective. Also the functional fragmented character of the Council decreases the relevance of territorially bounded practices and views. Expertise-based and functional identities are becoming more important. And finally, as supranational institutional actors - the European Commission and the European Parliament - and majority voting are increasingly growing in importance, national coloured views and practices tend to loose importance.

 

If we want test some of these general claims at the level of the official involved in CWPs, we have to translate them into specific hypotheses and measurable variables. In the remainder of this section, some independent variables and their supposed effect on role conceptions are presented and briefly discussed.

 

Overview 1. Summary of independent variables and supposed effects

Independent variables	Effect on the adoption of supranational role conceptions or the supplementing of intergovernmental roles with supranational roles
Euro-level experiences
1. duration of involvement	+
2. intensity of involvement	+
3. density of involvement	+
Professional experiences
4. previous European/international professional experiences	+
5. previous domestic professional experiences	-
6. low organisational self-esteem	+
Features of the working groups
7. qualified majority voting in CWP	?
Features of the member-states
8. length of membership	+ or curvilinear
9. member-states with pro-European elites	+

 

A first set of hypothesis concerns experiences of officials at the European level. Early neo-functional writings on socialisation argued that �more contact among elites and citizens of the member states, and as more and more people come to participate directly in these decisions and to perceive them as mutually rewarding, these people will develop more favorable attitudes toward integration� (Kerr 1973: 47) . A key question concerns whether regular and systematic contact with European institutions stimulates more favourable pro-European role conceptions. Yet, �contact� is a difficult variable. Usually the duration of contact, length of attendance in CWPs, is assumed to be a proxy for involvement in processes that are presumed to socialise agents. But duration as a variable does not say very much about the quality of involvement or the social mechanism that leads to socialisation. It does not trace the interaction process between t₀ (entering the organisation) and t₀₊₁ (worked for a while or leaving). Nevertheless, contact with the European level is a necessary condition; without this little European socialisation will take place. However, duration of contact should not be considered as a sufficient condition. In addition to duration, it is expected that especially intense and dense interactions contribute to the adoption of supranational role conceptions. The more frequent, or dense, the interactions occur and the more time demanding, or intense, these are for officials, the more supranational role conceptions will emerge.

 

A second set of hypotheses concerns previous professional experiences that are likely to shape role conceptions. First, a strong European socialisation hypothesis implies that those actors having extensive previous professional experiences with other European or international policy-making settings have had more chance to socialise. They already know the codes of conduct in European/international policy-making settings and, as a result, they should adopt rapidly supranational role conceptions or they should supplement more easily their intergovernmental views with supranational conceptions. A second, competing, hypothesis concerns the impact of domestic socialisation experiences. Here, the idea is that CWPs basically aim at increasing or preserving national control over European policy-making processes. As a consequence, officials having extensive domestic policy networks because of various domestic professional experiences are parachuted in EU-level settings. Therefore, extensive national professional experiences should contribute to the adoption of intergovernmental roles and likewise they should constrain supranational role conceptions.

 

From an intergovernmetalist �top-down� perspective, the Council imposes or reinforces intergovernmental role conceptions upon state agents. Domestic officials are assumed to re-represent and defend national views and interests. However, the national position officials should represent is often not well-conceived and ambiguous because of weak and insufficient domestic co-ordination structures or unclear political guidelines. A �bottom-up� perspective implicates the domestic policy-making process affects how officials think they should act when they enter the realm of CWPs. Here, it is hypothesised that those who perceive deficiencies and shortcomings in their own political-administrative setting or those having a low organisational self-esteem, develop, as an escape route, more easily supranational role conceptions. The absence of strict and detailed mandates or a substantial political backing makes that these officials have more �freedom�. They experience less inconsistency between, on the one hand, their role as a national representative, which they find difficult to adopt because the �national interest� is ill defined, and, on the other hand, someone who actively promotes the European common good. In this case, it is not the European level that induces the adoption of supranational role conceptions. Rather, this possibility refers to a situation where officials are supposed to represent what they experience as a rather ambiguous interest and this gives them some room or discretion to develop supranational role conceptions (see on Commission officials also Egeberg 1996) . This outcome resembles Checkel�s second hypothesis, namely that the interaction processes or mechanisms through which socialisation takes places is more likely to be occur when �the persuadee has few prior, ingrained beliefs that are inconsistent with the persuader� message� (Checkel and Moravcsik 2001: 222) . A similar hypothesis is put forward by Marks and Hooghe when they write that �� those who already rely on effective national networks have an interest in intergovernmentalism. Officials from weak national networks have an interest in supporting supranationalism� (Hooghe & Marks, 2001: 158).

 

Thirdly we explore the possibility that some specific CWP-features affect role conceptions. More in concrete, it is investigated whether the voting procedure applied for issues dealt with in the CWP matters. Qualified majority voting can be considered as a factor diminishing intergovernmental role conceptions because we know that it stimulates national delegates to build more extensive communication networks (Beyers and Dierickx 1998: 308) . In this respect, it is a proxy for the above-mentioned density contact-variable, or the frequency of CWP-meetings. Yet, for various reasons it is somewhat too simple to suppose that qualified majority voting coincides with supranational role conceptions. Coalitions might be formed because of national interest calculations and they are not necessarily inspired by a pro-European role conception. In addition, a social environment wherein a unanimity rule exists does not necessarily induce intergovernmental role conceptions or realpolitik behaviour (Johnston 2001: 502-506) . In transparent multilateral settings, as the EU, concerns about images and status are important; systematic defection from co-operation might be a costly strategy. At least empirically, it is well known that within CWPs or COREPER, officials do not vote in a formal way and that even at the ministerial level formal voting is a rather unusual practice. Most observers report that in daily practice negotiators try to search consensus and avoid formal voting. For these reasons, although a positive effect of majority voting sounds plausible, it remains to be seen what the effect of voting on the emergence of supranational role conceptions should be.

 

Finally, we consider two conditions related to the member-state themselves. First, we expect that older member states� representatives adopt more supranational role conceptions than newer member states. This factor has a similar effect as duration, or length of attendance, in the sense that length of membership implies more experiences and, as a result, the adoption of supranational roles. However, length of membership might be collinear to some specific nation-state, or non-Euro-level, factors. The founding member-states - Benelux, France, Germany and Italy � formed the European institutions during an era characterised by a permissive consensus, their political elites laid the basis for the contemporary institutional infrastructure and they had more time to learn the peculiarities of European policy-making. If time or duration would play a role, then the countries that are part of the first enlargement, the first wave, should adopt more supranational role conceptions than the second wave countries. However, countries belonging tot the first wave - United Kingdom, Ireland and Denmark - were at the moment of their entrance reluctant about supranationalism and for this reason, although it cannot be verified in detail, their governments did probalby not appoint supranationalist officials in the CWPs. On the contrary, it is plausible that, given the general political climate of that moment, they charged the most intergovernmental officials with representation and negotiation tasks. Although these officials had more time to become socialised than those whom entering the EU at a later stage, they started off as more intergovernmental than the Southern states - Spain, Portugal and Greece - which became members in a second wave. If this would be correct, the evidence should reflect a curvilineair pattern; the founding members and the Southern member-states as most supranationalist and least intergovernmentalists, while the first wave should be intergovernmentalist and least supranationalists.

 

Second, most domestic officials negotiating in the CWPs received a substantial part of their training and political education within national political systems. It is there that they learned the values and orientations of the national political and administrative elite to which they belong. From this perspective, Euro-level factors only slighty affect the adoption of supranational role conceptions. More important is that the roles adopted by domestic officials are basically a reflection of values and ideas shared by the respective national elites to which they belong (Hooghe 1999) . This hypothesis views the national official as weakly socialised by Euro-level factors. On the contrary, the national official represents views and practices in accordance with national elite orientations and, consequently, the more a national political elite favours supranationalism, the more officials adopt supranational role conceptions.

 

Methodological issues and data-set

In this section, some methodological problems related to the operationalisation of the above mentioned hypotheses and our data-set are discussed. Socialisation refers to social mechanisms or process which leads to the adoption and internalisation of views or practices. Actors do not change their allegiances the very first moment they interact in a new environment; presumably some time is needed before new role conceptions are adopted. It is exactly this notion of �time� that raises methodological problems in terms of measurement and testing, problems that complicate empirical research. One way to research a socialisation hypothesis would be a panel design; interviewing officials before they enter a new organisation, followed by an interview when the actor is active in new organisation and, finally, if possible, an interview after the time the official left the organisation (e.g. returns to a domestic governmental agency). Unfortunately, such designs are rather difficult to implement in international organisations. Another valuable approach would be qualitative process tracing or carrying out some in-depth case studies. Although such designs are maybe somewhat easier to implement, few qualitative research projects capture the temporal process of interaction and, in general, most studies fail to control for domestic selection processes, selective recruitment or self-recruitment.�

 

Our project used evidence derived within the framework of a quantitative quasi-experimental research design. Its added value is that it can control scope conditions at the European level for domestic level conditions. Although this design makes a rigorous testing of some conventionally accepted assumptions regarding international socialisation possible, it is not yet the �ideal� data-set. The notion of social mechanisms or time is implicitly, not explicitly, incorporated in this design. It is assumed that the presence of specific scope conditions implies the likelihood that particular social mechanisms occur. For instance, a dense environment characterised by frequent interactions and a consensus-oriented atmosphere suggests that processes of argumentative deliberation instead of pure bargaining takes may place. When meeting with colleagues from other member-states on a daily basis, it is imaginable that social interactions tend to go beyond instrumental calculations by member-state representatives and that a European group identity or allegiance will emerge. More in-depth process tracing or panel-data would re-dress this deficiency of our data-set.

 

The observational units are individual officials who were, to a varying degree, involved in European policy-making, more in particular decision-making within the CWPs. It was decided that, given the large number of working groups (170), it would not be valid to rely on a small number of witnesses in order to establish the facts. However, at the time when the research project started (1993) detailed information regarding the officials and their involvement in the CWPs was not available in a systematic fashion. Therefore, we started, with a quasi-complete list of 170 working groups operating during the Belgian presidency of 1993, collecting detailed information regarding the Belgian officials who were active in these 170 groups. This list of CWPs was used during 28 qualitative in-depth interviews in order to screen all CWPs. On the basis of these interviews most CWPs were coded according to density of meetings (e.g. ranging from �never� to �several times a week�), policy area (based on the items discussed), type (e.g. cross-cutting policy fields, technical regulation in an existing policy field�), composition (travelling officials or Brussels based officials) and the applicable voting procedure for the issues on its agenda (qualified majority voting or unanimity).

 

For making distinctions according to the level of involvement in the CWPs, the sample of officials has been divided in three sub-groups, i.e. the non-Belgian multinational full-timers, the Belgian full-timers and the Belgian part-timers. Belgian part-timers are officials from sectoral ministries who participate in Council working-groups on a part-time basis. They take only occasionally part in CWPs and they combine this with other, mainly domestic, tasks. For these officials CWPs are secondary institutional affiliations. The Belgian and non-Belgian full-timers are involved in CWPs on a more regular basis and include diplomats at the Foreign Office or the Permanent Representation (PR hereafter) and officials of sectoral ministries attached to the PR. [1] These officials are supposed to be more exposed to the peculiarities of European policy-making than the part-timers. All these officials, part-timers and full-timers, were interviewed regarding their involvement into one particular CWP. The distinction between Belgian part-timers and Belgian full-timers was key in order to investigate whether different level of involvement lead to the adoption of supranational role conceptions. The two sub-samples of Belgian and non-Belgian full-timers enable to test the effect of similar variables in a large multi-national data-set. [2] �

 

Domestic and European embeddedness of full-timers and part-timers

In this section, we describe key independent variables, namely the European and domestic experiences of Council officials. Domestic and European experiences are strongly affected by recruitment patterns. The recruitment of professional diplomats, about 50% of the full-timers in the Belgian PR, is the formal responsibility of the Foreign Office (Kerremans and Beyers 2001) . In principle, Belgian diplomats work at the PR for a period of three, maximally four years. [3] However, the highest diplomatic positions in the PR deviate form this general pattern and these posts are unique in many respects. First, the traditional four-year rule does not apply to the Belgian ambassadors. They serve much longer periods at their post in the PR than one would expect from an average professional diplomat. Second, most officials at the highest position in the PR had previous, sometimes extensive, experiences with the European Union and its member-states. And thirdly, these �European� experiences correspond with an extensive experience and involvement in various domestic political networks.

 

Lets briefly illustrate these points by looking to the career of current Permanent Representative, Frans Van Daele. His internship (in 1971) brought him under the wings of Etienne Davignon who was at that moment political director at the Foreign Office (and who later became Commissioner). At that moment, Davignon played an important role in the emergence of the European Political Cooperation (EPC). After his training, Frans Van Daele started his career as a professional diplomat at the Belgian PR to the EU (1972-1977). In 1977 he moved to Athens where he was involved in the accession negotiations with Greece. In 1984 he returned to the Belgian PR to the EU. Between 1984 and 1986, he became spokesman for Leo Tindemans, the foreign minister at that moment. After diplomatic posts in Rome and New York, he returned (in 1993) to Belgium and became Deputy Political Director responsible for the implementation of the Maastricht Treaty and for ex-Yugoslavia. In the first semester of 1993, he was extensively involved in Belgium�s presidency of the European Union. In 1994 he became Political Director at the Foreign Office and member of the Belgian negotiating team in the 1996-97 Intergovernmental Conference. When Ambassador de Schoutheete retired in September 1997, Van Daele succeeded him.

 

Working at the PR as a diplomat is not just an ordinary thing; these functions are reserved for the best. Important also is that the resulting prestige of the job has political consequences. Domestic party political considerations are unambiguously and openly taken into account for the appointments (Dierickx and Beyers 1999: 200-203) . It is not that somebody of a particular political colour has more chances to become Permanent Representative or his deputy. But in the act of balancing different important and prestigious diplomatic functions among the different political colours and, importantly, between the two language groups, the PR is considered to be one of the �highfliers�. Consequently, the relation of these highfliers with domestic political parties is no secret at all.

 

The recruitment of sectoral attach�s at the PR and the part-timers happens in a somewhat different way and depends strongly on the ministry that sends its attach� to the PR. Also the appointment of part-time officials negotiating in the various CWPs takes place in a rather ad hoc manner. Neither the Permanent Representative himself, nor the Foreign Office is systematically involved in this process. It is something that happens inside the different ministries as a result of the interplay between the highest officials, directors-general and secretary-general. Our evidence suggests that the minister, his cabinet or interdepartmental co-ordination processes play a rather limited role in this recruitment process. In all cases however, full-timers (i.e. diplomats, attach�s at the PR and Foreign Office officials involved in European matters) are relatively high-ranking officials and merit, professional or personal interests seem to play a pivotal role here. Also relevant are previous experiences with the European Union in addition to negotiating experiences in various internationalised policy-settings.

 

Does this correspond with available quantitative evidence? We asked our respondents two open questions: �why did you choose a job in the public service?� and �how did you actually become involved in European policy-making?� and, broadly speaking, respondents mentioned two categories of factors. Some respondents said their entrance in the public service was a �pure coincidence of various elements and not necessarily a personal choice�, �that they had no alternative� or that they chose this job because it implicates �a fixed wage and income�. Regarding their involvement in European policy-making, respondents answered that �they were just available�, �that nobody else wanted to take up this job� or �that they became involved by coincidence�. These answers contrast with answers, both for CWPs and entrance in the public service, which refer to �professional experience�, �personal interest, �intellectual interest� or �political interest�. In general, it is noticeable that the first kind or answers, i.e. factors unrelated to personal interest or merit, are significantly more mentioned by Belgian part-timers (67% for the entrance of the public service and 72% for involvement in European policy-making) than by the Belgian full-timers (30 and 50%) or the full-timers of other member states (34 and 61%). On balance, merit, factors related to professional experience or general interest are more important for the recruitment of full-timers than the part-timers� recruitment.

 

What about the European embeddedness of officials, their current involvement in policy-making processes at the European level in general and their participation in CWP in particular? In general, the distinction between part-timers and full-timers is not significantly related to the duration - in years - of involvement in various CWPs, but it corresponds with the intensity of contact and the density of CWP gatherings in which officials were involved. Within the group of Belgian respondents the duration of involvement is slightly higher for the full-timers (12 years) than for the part-timers (11 years) and the non-Belgian respondents had, on average, spent nine years in CWPs. Yet, it concerns small and statistically irrelevant differences (F=1.69, df=2, p=.1872). It is more the intensity of involvement and the density of CWP meetings which differentiates part-timers from full-timers. Among the full-timers 80% of the Belgian officials and 89% of the non-Belgian officials devoted more than half of their time to meetings with CWPs. Only 42% of the Belgian part-timers spent more than half of their time to meetings with CWPs. [4] The density of CWP-meetings is substantially higher for the CWPs in which full-timers were involved than for the part-timer�s CWPs. Of the Belgian part-timers only 29% was involved in CWPs that met more than once per month. The proportion of Belgian full-timers and European full-timers involved in these densely meeting CWPs was respectively 70 and 87%. [5]

Table 1. Part-timers and full-timers compared regarding their European involvement

	Non-Belgian full-timers (n=106)	Belgian full-timers (n=30)	Belgian part-timers (n=65)
1. Duration of involvement �� (Average number of years involved in European policy-making)	9	12	11
2. Intensity of involvement �� (Percentage devoting 50% or more of their time to meetings with CWPs)	89%	80%	42%
3. Density of involvement �� (Percentage involved in CWPs which met more than once per month)	87%	70%	29%

 

Table 2 and 3 show some results regarding previous professional experiences of the officials interviewed. In general, these tables suggest that European/international professional and domestic professional experiences predict whether an official belongs to the group of part-timers or full-timers. As qualitative evidence already illustrated, full-timers follow a professional trajectory that has a quite European/international character; they have spent more of their professional career in other EU-member states, an indication of extensive �European� socialisation opportunities. Especially the Belgian full-timers seem to be characterised by both extensive European/international and domestic professional experiences. [6] And, when comparing the domestic and European/international social environments, it appears that for the Belgian full-timers domestic professional experiences seem to be extremely relevant. That Belgian full-timers have had more experiences with ministerial cabinets and have more affinities, as previous staff-member, with domestic political parties, confirms the importance of domestic political factors in the recruitment of Belgian full-timers (as compared to Belgian part-timers and non-Belgian full-timers). [7] Although this dataset is rather small in order to reach firm conclusions regarding previous experiences with interest groups, the quantitative evidence and the in-depth interviews suggest that most sectoral specialists at the Belgian PR have strong ties with the domestic interest groups of their sector. For instance, the two representatives from the agricultural ministry at the Belgian PR, have close links to the biggest Belgian agricultural interest association, the Boerenbond.

 

Table 2. Part-timers and full-timers compared regarding previous European/international professional experiences

	Non-Belgian full-timers (n=106)		Belgian full-timers (n=30)		Belgian part-timers (n=65)
	%	Years, average	%	Years, average	%	Years, average
1.     Having worked as a civil servant in an international organisation	7%	.67	3%	.08	3%	1.42
2.     Having been attached to an embassy in another EU member-state	10%	1.15	17%	3.63	0%	.00
3.     Having been attached to an embassy in a non-EU state	15%	2.21	23%	3.92	5%	4.67
4.     Having been attached to a diplomatic mission toward an international organisation	6%	.65	17%	1.31	0%	.00
5.     Having worked as a bureaucrat for the European Comission	3%	.10	3%	.23	3%	.29
Overall European/international professional experiences	36%	4.78	43%	9.17	9%	6.38

Index: The average number of years only concerns those officials that had the professional experience mentioned.

 

Table 3. Part-timers and full-timers compared regarding previous domestic professional experiences

	Non-Belgian full-timers (n=106)		Belgian full-timers (n=30)		Belgian part-timers (n=65)
	%	Years, average	%	Years, average	%	Years, average
1.     Having been attached to the personal staff or the cabinet of a ministers	33%	2.57	50%	2.70	18%	2.35
2.     Having worked for local authorities	3%	.66	3%	.05	6%	.27
3.     Having worked for regional authorities	5%	.35	3%	.35	8%	2.07
4.     Having worked for a domestic political party	2%	.12	13%	1.90	3%	.18
5.     Having worked for an interest groups; trade union	1%	.02	7%	.30	2%	.36
6.     Having worked for an interest groups; farmer�s union	2%	.24	7%	.13	2%	.23
7.     Having worked for an interest groups; employer�s union	2%	.12	3%	.25	2%	.05
8.     Having worked for an interest groups; SME�s	1%	.05	0%	.00	0%	.00
Overall domestic professional experiences	38%	4.14	67%	5.68	34%	5.50

Index: The average number of years only concerns those officials that had the professional experience mentioned.

 

Crucial here is whether European or domestic experiences shape the role conceptions. If European socialisation represents only secondary socialisation and if role conceptions are primarily shaped by various domestic experiences, it should be clarified in how far both domestic and European experiences are truly independent from each other. [8] Or, domestic professional experiences should not affect too much the extent to which actors have current or previous professional experiences with the European level. The aggregate data analysed above suggest that the extensive European involvement of, especially Belgian, full-timers corresponds to extensive domestic experiences. Does such an association also appear at the individual level? Or, if an official has much European experience, does he also have extensive domestic experiences? It would take us too far to present all correlations in detail. In general, it appears that, controlled for full-timers versus part-timers, the overall extent of current (density, intensity and duration) and previous European/international professional experiences (in average years, last row table 2) is not associated with prior domestic professional experiences (in average years, last row table 3). At the individual level, previous European experiences are, again controlled for a potential difference between part-timers and full-timers, not correlated to the current involvement into CWPs. [9] Only when considering Belgian officials more in detail some evidence suggests that particular domestic experiences affect specific (current or previous) European experiences for specific groups; it cannot be concluded that the general pattern of domestic experiences predicts the overall (current and previous) experience with Europe. [10]

 

Exploring role conceptions

After considering some key independent variables more in detail, this section explores and outlines the dependent variable, namely the supranational and intergovernmental role conception actors adopt. Supranational and intergovernmental role conceptions concern how actors believe they should act when they participate in European policy-making settings. Those with an intergovernmental role conception conceive European policy-making as an act of exchanging and balancing member state interests. A supranational role conception implies that Europe is an autonomous policy level primarily designed for finding policy solutions in the interest of a common, European, good.

 

During the interviews, the respondents were presented nine statements, each tapping into how they should act in the CWPs, and asked to specify their view on a scale ranging from to 1 (disagree) to 6 (agree). The higher the score on one of these items, the more the interviewee was convinced that he or she should act according to the role conception outlined in a particular statement. Table 4 shows all nine items, factor-analysing the answers. The first two columns present factor loadings for all respondents and, in order to check the similarity of the factor pattern, an additional analysis is carried out for the all the Belgian respondents and the multinational sample (including Belgian full-timers) separately. The first five items clearly correspond to a supranational role conception; officials should �draw up guidelines for the member states�, �develop a common policy�, �collaborate and look for common objectives� and �strengthen the role of European institutions�. Key words in the other four items are: �taking into account interests of other member states�, �defend national positions� and �exchange and balance national viewpoints�. These four items refer to a more intergovernmental role conception.

 

So far the analysis suggests the existence of two distinct role perceptions. What about the possibility that actors, to some extent, mix up role conceptions or supplement one role with another role? Or, are officials capable to separate one role from the other and thus keeping each role distinct? Do actors simultaneously combine the evoking of one role conception with another? Hypothetically three patterns in the data refer to different ways in which role conceptions might be combined. First, evoking supranational role conceptions might positively correlate with the adoption of intergovernmental role conceptions. In this situation actors, systematically mix up both role conceptions in the sense that evoking an intergovernmental role, or �staying native�, is supplemented with a supranational role conception, or �going native�. Second, a negative correlation means that both role conceptions are distinct, but in this case an outspoken supranational role conception implies that one tends to reject an intergovernmental role conception. Those who �stay native� will not �go native� (and vice-versa). Thus, a negative correlation suggests that the main cleavage in CWPs concerns a cleavage between those adopting supranational roles and those adopting intergovernmental roles. Then, both role conceptions are antagonised. Third, the absence of a substantial correlation should not be considered as the complete absence of supplementation or antagonising. It means that on average supplementation or antagonising is not taking place. If a substantial part of the dataset is extensively supplementing role conceptions and another part is antagonizing roles, the net correlation result is likely to be zero. Although supplementation or antagonizing is not the general pattern, it might take place in some specific cases. For instance, it might be that actors more extensively connected with the CWPs are more disposed to supplement both role conceptions, that is they see less a difference between defending national positions and working towards common European solutions.

 

In this section, we tentatively illustrate that actors can and do combine different role conceptions. In the next section, a more systematic analysis with independent variables is carried out. In factor-analysing the data, an oblique rotation, instead of imposing an orthogonal structure, was carried out. This enables to explore whether and in how far it is realistic to assume complete independence between the two role conceptions and the individual statements measuring these. This approach is useful since it allows observing the eventual supplementing or antagonising character of two, analytically distinct, role conceptions. If items cross-load on both factors or if a substantial inter-factor correlation is observed, then it is plausible to presume that both analytical distinct role conceptions are blurred.

 

A closer inspection of the factor-pattern in table 4 indeed demonstrates that both conceptions are quite, although not perfect, distinguishable. Although the first two columns present no substantial cross-loadings, it is not wise to look for differences within a heterogeneous dataset putting together Belgian part-timers, Belgian and non-Belgian full-timers. The other two columns suggest that within subsets of our data different patterns of supplementing exist. The positive inter-factor correlation for Belgian respondents (full-timers and part-timers) suggests that on average Belgian officials tend to supplement intergovernmental role with supranational roles (r=.28, p=.0059).� Also, the positive, and in three cases rather substantial, cross-loadings point into this directions. The outcome is less straightforward when we consider the full-timers. Although the inter-factor correlation is low and insignificant (r=.08, p=.3707), some cross-loadings suggest that antagonising rather than supplementing practices are occurring here. It is noticeable that the most �intergovernmental� statement referring to the defence �of our national position� points to an antagonizing conception. To conclude, there may be a pattern of antagonising, rather than supplementing, behaviour among full-timers, but the results remain weak in order to make firm conclusions.

 

Table 4. Supranational and intergovernmental role conceptions (factor loadings, procrustean oblique rotation)

	All respondents         N=203		All Belgian respondents, full-timers & part-timers     N=95		Multinational sample of full-timers (including Belgian full-timers in 13 CWPs) N=123
	supran	interg	supran	interg	supran	Interg
1.        In the working groups, the European Commission and the national representatives should take an active part in drawing up guidelines for member states' policies.	.70	.03	.74	.26	.64	-.07
2.        In the working groups the representatives should develop a strong common policy and lay down clear directions for the national govern�ments.	.72	-.14	.55	.01	.78	-.08
3.        The main task of the working groups is to look for common objectives and a common policy in collaboration with the different member states and with the European Commission.	.61	.06	.58	.27	.57	-.04
4.        In my opinion, in the working groups the European Commission has to arbitrate between the diverse national interests in order to make a common policy possible.	.44	.04	.55	.01	.43	-.01
5.        In my opinion, in the working groups we should work towards a strongly united policy which strengthens the executive role of the European Institutions.	.70	.03	.54	.13	.78	.11
6.        In my opinion the exchange of information on positions in working groups helps avoiding open differen�ces of opinion which might stand in the way of an agreement at the level of the COREPER, the SCA or the Council.	-.07	.38	.33	.33	.00	.41
7.        Information on national positions in working groups is useful in order to be able to take into account interests of the other member states when setting out our national policy.	.07	.40	.24	.33	-.01	.47
8.        In my opinion, in a working group we have to inform and defend our national position.	-.14	.40	.04	.59	-.34	.22
9.        According to me, we should inform the representatives of the other member states and of the European Commission about our national points of view helps facilitating higher‑level consultations.	-.02	.63	.15	.74	-.18	.48
Explained variance	33%		33%		33%
Inter-factor correlations			r=.28, p=.0059		r=.08, p=.3707

Index: Inter-factor correlation are pearson product moment correlations between additive indices (standardised). Similar measures generated by the PROC FACTOR statement in SAS yield comparableresults. Supran = factor referring to supranational role conceptions and Interg = factor referring to intergovernmental role conceptions. Factor-loading higher than .30 are put in bold.

 

Explaining role conceptions

One conclusion from the previous section is that on average role conceptions are not systematically and substantially supplemented or antagonised. On average, officials adopt a supranational or an intergovernmental role conception and they do not mix up both. However, it has to be recognised that supplementation or antagonising can take place within among specific categories of actors. For instance, especially extensively involved officials may be able to combine both roles. In general, if European socialisation takes place, then the supplementation of role conceptions, substantiated by significant inter-factor correlations, should occur among the extensively involved officials. Such contact variables should also affect the mere adoption of supranational role conceptions substantially.

Table 5. Supplementing of supranational and intergovernmental role conceptions (averages and inter-factor correlations)

	Average suprantionalism		Average inter-governmentalism		Inter-factor correlation
	BS-FPT	MNS-FT	BS-FPT	MNS-FT	BS-FPT	MNS-FT
Euro-level experiences
1. duration of involvement
��� - before 1986	.08	-.06	-.15	.22	.26	ns
��� - after 1986	-.04	.03	.26	-.14	Ns	ns
2. intensity of involvement
��� - less than 50% of time devoted in CWPs	-.07	.11	-.00	.11	.48	ns
��� - more than 50% of time devoted in CWPs	.06	-.02	.00	-.02	Ns	ns
3. density of involvement
��� - less dense CWP-meetings	-.05	.17	-.02	.03	.47	ns
��� - dense CWP-meetings	.07	-.02	.03	-.00	Ns	ns
Professional experiences
4. Previous European/international professional experiences
��� - no	.05	.07	.03	.19	.28	ns
��� - yes	-.22	-.12	-.11	-.30	Ns	ns
5. Previous domestic professional experiences
��� - no	.04	-.01	.10	.03	ns	ns
��� - yes	-.05	.01	-.12	-.04	ns	ns
6. Organisational self-esteem
��� - high	-.97	-.12	.25	-.12	Ns	ns
��� - moderate	-.27	-.02	-.11	-.00	.39	ns
��� - low	.29	.35	.08	.05	Ns	ns
Features of the working groups
7. Applicable voting power in CWP
��� - qualified majority voting (including mixed groups)	.10	-.02	-.07	-.09	Ns	Ns
��� - with unanimity	-.17	.03	.12	.13	.39	Ns
Features of the member-states
8. Length of membership
��� - original member-states (Benelux, France, Germany, Italy)	-	.17	-	-.17	-	.25
��� - first wave (UK, Denmark, Ireland)	-	-.46	-	.43	-	ns
��� - second wave (Greece, Spain, Portugal)	-	.13	-	-.10	-	ns
9. How pro-European are national elites1
��� - most pro-European elites (Belgium, Greece, Germany, Italy)	-	.26	-	-.18	-	ns
��� - moderate pro-European elites (Netherlands, Luxemburg, France, Spain)		-.03		-.12		ns
��� - least pro-European elites (Ireland, Denmark, Portugal, United Kingdom)	-	-.26	-	.30	-	ns
Full-timers versus part-timers (only Belgian sample)
�� - Belgian part-timers	.09	-	-.05	-	.36	-
�� - Belgian full-timers	-.20	-	.02	-	ns	-

Index: MNS-FT= Multinational sample of full-timers (N=123), BS-FPT= Belgian sample of full-timers and part-timers (N=95); Averages that differ significantly (according to an F-test) from each other are put in bold; ¹ = categorisation based on the Eurobarometer elite survey of 1996

 

Table 5 introduces some categorical distinctions related to the independent variables outlined in section one. Average supranational and intergovernmental role conceptions are presented across subgroups separately for multinational full-timers and Belgian full-timers and part-timers. The last two columns show inter-factor correlations within subcategories of the data. These results enable us to explore whether and to what extent role conceptions are supplemented within relevant subgroups. Table 6 correlates the two role conceptions with the independent variables, but instead of crude categorical distinctions as in table 5, the original not recoded variables are used.

Table 6. Supranational and intergovernmental role conceptions, bivariate correlations

	Average suprantionalism		Average inter-governmentalism
	BS-FPT	MNS-FT	BS-FPT	MNS-FT
1. duration of involvement (number of years)	ns	ns	ns	ns
2. intensity of involvement (5-point scale)	ns	ns	ns	ns
3. density of involvement (6-point scale)	ns	ns	ns	ns
4. previous domestic professional experiences (number of years)	ns	ns	ns	ns
5. previous European/international professional experiences (number of years)	ns	-.24	ns	ns
6. organisational self-esteem (additive index)	.32	.23	ns	ns
7. qualified majority voting in CWP (dichotomous)	ns	ns	ns	ns
8. length of membership (three waves)	-	-.24	-	.23
9. member-states with pro-European elites (Eurobarometer elite survey of 1996)	-	.31	-	-.19

Index: MNS-FT= Multinational sample of full-timers (N=123), BS-FPT= Belgian sample of full-timers and part-timers (N=95).

 

The discussion of the results starts with the inter-factor correlations in table 5. Given the absence of negative correlations, it can be concluded, that antagonising of role conceptions is not the usual pattern. In addition, the predominance of supplementing behaviour (instead of antagonizing) gains further confirmation by the observation that especially among the bigger member-states� officials (not in the table) and within unanimity CWPs both role conceptions are combined. When isolating the four member-states with the biggest voting power, a significant correlation is observed (r=.43, p=.0026). These findings are not irrelevant. They clearly refute the popular conception of European politics as an ongoing antagonistic struggle between the big member states. Rather the evidence suggests that especially the bigger member states, not necessarily the smaller member states, have adopted an attitude that aims to reconcile their national interest with a common European interest. It is also interesting to observe that unanimity does not go together with pure intergovernmental styles of European policy-making, but that especially under conditions of unanimity actors tend to supplement role conception. Or to put it differently, more majority voting does not necessarily lead to a shift of allegiances.

 

Furthermore, the evidence gives no strong support for the hypothesis that especially extensive experiences with Europe lead to the supplementing of role conceptions. On the contrary, in the Belgian sample we find considerable evidence for supplementation among part-timers (r=.36, p=.0033), among those without previous transnational experiences (r=.28, p=.0127), among those involved in less dense CWPs (r=.47, p=.0003) and among those who spent less than 50% of their time in CWPs (r=.48, p=.0010). In contrast, officials that are well involved in European matters are less likely to supplement these role conceptions. [11]

 

Does intensive, dense and prolonged participation in CWPs lead to more supranational and less intergovernmental role conceptions? In general, nor the bivariate correlations in table 6, nor the categorical distinctions presented in table 5, give strong support to the European socialisation hypothesis. What we observe are weak and unsystematic patterns of European socialisation. Some previous European experiences seem to moderate intergovernmental role conceptions. Among Belgian respondents, those officials that became involved before 1986 have a less intergovernmental role conception (F=4.00, df=1, p=.0484) and among the Euro-level full-timers the availability of previous European/international professional experiences decreases intergovernmentalism (F=7.25, df=1, p=.0081, see also table 6). But despite these effects, the evidence gives no strong and systematic support to the strong European socialisation hypothesis.

 

More systematic and significant results are found for hypotheses related to member-state characteristics and domestic socialisation. First, the results in table 6 demonstrate, as expected, that founding member states develop more supranational roles and less intergovernmental rules than the newer member states. However, the distinctions in table 5 nuance this general conclusion considerably. The curvilinear hypothesis, namely that founding member-states� representatives are more supranationalist (less intergovernmentalists), followed by the supranationalists (less intergovernmentalists) representatives of the second wave and that the first enlargement wave is most intergovernmentalists (less supranationalists), is confirmed. This outcome indicates that European socialisation is moderated by domestic factors and not by the duration of EU-membership. Second, we find substantial support for the hypothesis that adopted roles reflect to a significant extent the general elite attitude within a member-state. In general, if an official represents a member-state of which the societal elite is on average pro-European, the official tends to act according to a supranational role conception and reject intergovernemental roles. In contrast, if a CWP-official represents a member-state where the overall elite attitude is more Euro-sceptic, we observe the opposite. In this respect, domestic Council-officials are quite �representative�.� Finally, in both the Belgian and the European sample specific domestic experiences correlate with the adoption of supranational role conceptions. In order to measure the distance officials experience from their domestic politico-administrative system, seven Likert-scale items were combined in a scale labelled organisational self-esteem (see table 7). In general, a low organisational self-esteem is positively associated with the adoption of supranational role conceptions and especially in the Belgian case the perceived weaknesses of domestic networks contributes to supranationalism.

 

Table 7. Measurement of Organisational Self-Esteem (all 5-point scales, factor-loadings)

	Belgians (N=95)	Multinational sample (N=123)
1.        The internal co-ordination of the viewpoints of the different ministries in our country is chaotic.	0.68	0.76
2.        In the administrations of my country there is not sufficient training for officials who have to take part in negotiations at a European level.	0.59	0.48
3.        It often happens that I am not quite certain what point of view I should put forward in the working groups.	0.68	0.71
4.        For officials it is very important that the preparation of a policy is easily surveyable. The structure of the administration in our country does not always add to this effect.	0.69	0.60
5.        Most member states prepare themselves more thoroughly for the negotiations in Brussels than we do.	0.52	0.56
6.        I only learnt how I had to deal with complicated European dossiers as I went along. �	0.49	0.56
7.        I always get very clear instructions from my ministry or my department as to what position I should take up.	0.59	0.61
Explained variance	37%	38%

 

Table 8 Supranational role conceptions, organisational self-esteem and involvement in CWPs (Belgian sample, average supranational role conception)

	Organisational self-esteem
	High (N=30)	Middle (N=37)	Low (N=28)
Full-time involvement (N=30)	-.35	-.39	.19
Part-time involvement (N=65)	-.25	-.13	.74

 

Lets illustrate this last finding in a more succinct manner. We already know that the distinction between part-timers and full-timers corresponds with embeddedness in CWPs. An alternative strategy for investigating the effect of involvement, therefore, uses the distinction between part-timers and full-timers as a proxy for involvement. In general, the average organisational self-esteem is somewhat higher (-.19) among Belgian full-timers than among Belgian part-timers (.09), but this difference is statistically insignificant (F=1.61, df=1, p=.2076). A covariance analysis with one interaction term (involvement*organisational self-esteem) revealed that a low domestic organisational self-esteem stimulates supranationalism (F= 5.29, df=2, p= .0067). Table 8 displays the average supranational role conception according to self-esteem and involvement. It appears that not only involvement makes a substantial difference. Rather the evidence suggests that the lesser one has been involved in Europe (i.e. part-timers), the more a low organisational self-esteem (i.e. a negative view about domestic co-ordination and policy-preparation) leads to the adoption of supranational role conceptions. Domestic factors, and not the involvement in CWPs, stimulate supranational role perceptions. At least in the Belgian case, some specific domestic political factors, and more precisely weak domestic co-ordination networks are a factor leading to the adoption of supranational allegiances.

 

Discussion and conclusion

It has to be recognised that the evidence reported in this paper does not lead to general conclusions regarding other international organisations or, with regard to part-timers, to other member states. The exploration of European socialisation in CWPs does not reject or confirm the possibility that other EU-level settings � e.g. committees under the realm of the Commission - may function as sites of European socialisation. Nevertheless, our general observation that domestic level factors matter considerably more than European experiences is confirmed by a large array of empirical studies and makes us feel confident regarding this research outcome (for a review of some literature see Goetz 2000) . Also the limitation of the data to only Belgian part-timers has to be considered. By focusing on the Belgian case we may have underestimated the fact the Belgian administrative elite�s socialisation has been completed a long time ago and that it takes Europe �for granted� (Green Cowles and Risse 2001: 236) . �Europe� is not that new and so Belgian actors are maybe not pressed to adopt new roles. In this sense, our results should not be interpreted as a rejection of Trondal�s findings (Trondal 2002) . Trondal�s sample included officials belonging to nation-states for which the EU is, compared to Belgium, relatively new. In addition, regarding Belgium one might also object that the variance with respect to supranationalism is, given the overall domestic political consensus regarding supranationalism, rather low and so that not very much has to be explained.

 

However, it remains to be explained why within the Belgian sample about 20 to 30% adopted an intergovernmental role conception. Also it is, despite the limitation of our data, still worth noting that especially among the weakly involved officials the supplementing of role conceptions takes place and that most conditions related to European socialisation experiences did not have the hypothesised effect. In contrast, factors referring to domestic socialisation experiences had much more impact. More in particular, in the Belgian case it appears that weak domestic policy co-ordination foster supranational role conceptions. Although additional comparative research is needed in order to confirm the robustness of these findings, it is noticeable that as Trondal and Egeberg observed, also weak, not necessarily strong, European ties can lead to Europeanised role conceptions.

 

The research outcome on organisational self-esteem is not the ultimate confirmation of Checkel�s second hypothesis; if there are few prior ingrained positions, then socialisation takes more easily place (Checkel and Moravcsik 2001: 222) . However, it is clear that the results with respect to organisational self-esteem, interpreted as the absence of prior ingrained positions, is substantial enough for warranting further empirical and theoretical consideration. Also much of our qualitative data, that could not be reported here, points in this direction. From the qualitative interviews we learnt that among Belgian officials complaints about the increasing complexity of the domestic political system prevailed. The interplay between different agencies was considered to be inefficient or absent, the involvement of a large number of ministries �increases the probability of bad instructions�. One respondent said that Belgian negotiators most of the time work without detailed instructions, but that they have adopted one big instruction: �do what you think that is appropriate, but try to improve European decision-making, do not make enemies and ensure that you have a positive working relation with everyone�. Almost all stressed that the appropriate way to deal with Europe is �to comply with the views of the Commission� or �to follow the Commission�s ideas, because these usually fit what is good for Belgium�. In general, domestic political control was considered to be weak or even completely absent.

 

To conclude, bold claims concerning the �Europeanisation� of the nation-state - in terms of growing convergence or homogenisation � are misleading and often empirically incorrect. Even for actors who are regularly involved in European policy-making, European experiences basically function as as secondary socialisation processes not as primary socialisation process. Noticeable is that the recruitment of full-timers is driven by both previous professional experiences with the European Union and its member states and domestic professional experiences. However, the strength of current and previous European involvement is not the most important factor for the emergence of supranational role conceptions. Less, instead of strong, involved officials supplement role conceptions and especially domestic experiences cause the adoption of supranational role.

.
Literature

Beyers, Jan, and Guido Dierickx. 1998. The Working Groups of the Council of Ministers : Supranational or Intergovernmental Negotiations? Journal of Common Market Studies 36 (3):289-317.

B�rzel, Tanja A. 1999. Towards Convergence in Europe? Institutional Adaptation to Europeanisation in Germany and Spain. Journal of Common Market Studies 37 (4):573-596.

Checkel, Jeffrey. 1999. Social Construction and Integration. Journal of European Public Policy 6 (4):545-560.

Checkel, Jeffrey, and Andrew Moravcsik. 2001. Forum Section: A Constructivist Research Program in EU Studies. European Union Politics 2 (2):219-249.

Dierickx, Guido, and Jan Beyers. 1999. Belgian Civil Servants in the European Union. A Tale of Two Cultures? West European Politics 22 (3):198-222.

Egeberg, Morten. 1996. Organization and Nationality in the European Commission Services. Public Administration 74 (Winter):721-735.

Egeberg, Morten. 1999. Transcending intergovernmentalism? Identity and role perceptions of national officials in EU decision-making. Journal of European Public Policy 6 (3):456-474.

Friedkin, N.E. 1984. Structural cohesion and equivalence explanations of social homogeneity. Sociological Methods and Research 12:235-261.

Goetz, Klaus. 2000. European Integration and National Executives: A Cause in Search of an Effect? West European Politics 23 (4):210-231.

Granovetter, Mark. 1985. Economic Action and Social Structures: The Problem of Embeddedness. American Journal of Sociology 91 (3):481-510.

Green Cowles, Maria, and Thomas Risse. 2001. Transforming Europe: Conclusions. In Transforming Europe. Europeanization and Domestic Change, edited by M. Green Cowles, J. Caporaso and T. Risse. Ithaca: Cornell University Press.

Harmsen, Robert. 1999. The Europeanization of National Administrations: A Comparative Study of France and the Netherlands. Governance: An International Journal of Policy and Administration 12 (1):81-113.

Hayes-Renshaw, Fiona, and Helen Wallace. 1997. The Council of Ministers. Edited by N. Nugent, W. E. Paterson and V. Wright, The European Union Series. London: MacMillan.

Hoffmann, S. 1966. Ostinate or Obsolete? The Fate of the Nation State and the Case of Western Europe. Daedalus 95:826-915.

Hoffmann, S. 1982. Reflections on the Nation-State in Western Europe Today. Journal of Common Market Studies 21 (1-2):21-38.

Hooghe, Liesbet. 1999. Supranational Activists or Intergovernemental Agents. Explaining the orientations of senior Commission officials towards European integration. Comparative Political Studies 32 (4):435-463.

Johnston, Alastair Iain. 2001. Treating International Institutions as Social Environments. International Studies Quarterly 45 (December):487-515.

Kerr, Henry H. 1973. Changing Attitudes Through International Participation: European Parliamentarians and Integration. International Organization 27:45-83.

Kerremans, Bart, and Jan Beyers. 2001. The Belgian Permanent Representation to the European Union: Mail Box, Messenger, or Representative? In The National Co-ordination of EU Policy, edited by H. Kassim, A. Menond, B. G. Peter and V. Wright. Oxford: Oxford University Press.

Linssen, Hub, Louk Hagendoorn, and Ludo Matheuse. 1996. Changing nationality stereotypes through contact : an experimental test of the contact hypothesis among European youngsters. In Democratic socialization and conflicting loyalities in East and West: Cross-National and Comparative Perspectives, edited by R. Farnen, H. Dekker, R. Meyenberg and D. B. German. London: MacMillan.

Peck, Richard. 1979. Socialization of permanent representatives in the United Nations: some evidence. International Organization 33 (3):365-390.

Pollack, Mark. 1998. Constructivism, Social Psychology, and Elite Attitude Change: Lessons from an Exhausted Research Program. Paper read at Conference of Europeanists, at Baltimore, MD.

Risse, Thomas, and Antje Wiener. 1999. 'Something rotten' and the social construction of social constructivism: a comment on comments. Journal of European Public Policy 6 (5):775-782.

Scheinman, Lawrence, and Werner Feld. 1972. The European Economic Community and National Civil Servants of the Member States. International Organization 26 (1):121-135.

Smith, Keith A. 1973. The European Economic Community and National Civil Servants of the Member-States - A Commet. International Organization 27 (4):563-577.

Stephan, W.G. 1987. The contact hypothesis in intergroup relations. In Group processes and intergroup relations, edited by C. Hendrick. Beverly Hills: Sage.

Trondal, Jarle. 2002. Beyond the EU membership-non-membership dichotomy? Supranational identities among national EU decision-makers. Journal of European Public Policy 9 (3):468-487.

Wessels, W. 1991. The EC Council : the Community's decisionmaking center. In The new European Community. Decisionmaking and institutional change, edited by S. Hoffmann and R. O. Keohane. Boulder Co: Westview Press.

Westlake, Martin. 1995. The Council of the European Union. London: Catermill.

Wolf, Peter. 1973. International Organization and Attitude Change: A Re-examination of the Functionalist Approach. International Organization 27 (3):347-371.