THE NEXT EUROPEAN EMPIRE? Open Access

This paper proposes to use the concept of empire for the analysis the European integration process. From a methodological point of view, this concept is an invitation to reintroduce in EU studies the comparative approach as well as long-term history. This paper also leads to a detailed reflection on the concept of empire itself and proposes a modernized definition of it. The use of this concept enables to shed a new light on the EU and some of its policies. If the EU cannot be considered an empire, it can nonetheless be said that it is undergoing a process of imperialization. This raises the question of the transformation of the European project and to the consequences of this transformation.

The fifth enlargement has in many respects brought deep changes in the EU. This explains the rise of a new theoretical approach in order to conceive of the European polity and its integration process. Ulrich Beck and Edgar Grande probably pioneered the movement with their book Das kosmopolitische Europa (2004) and, two years later, Jan Zielonka's Europe as Empire (2006) developed in an unrelated research agenda the same idea: what if the EU were evolving towards an empire? That all three authors are ‘continental’ Europeans is certainly not irrelevant: the empires they have in mind are those empires, which appeared and disappeared on the European continent. These are very different from the British Empire, from empires in the Marxist theory, and from empires in a metaphoric sense. Beyond a common point of departure, both books focus on very different contexts and aspects of empires. Zielonka refers explicitly to ‘medieval’ empires – hence his concept of ‘neo-medieval empire’ – while Beck and Grande, less explicit, mostly refer to later European empires. While Beck and Grande, authors of the ‘second modernity’, focus on identity construction, post-nationalism, and cosmopolitanism, Zielonka is more interested in enlargement policies and in a view of the EU from the new member states. Together, both books shed complementary lights on imperialization trends of the EU and enrich the structural theory of empire (Motyl 2001) with which I will work in this paper.

The use of the concept of empire is not self-evident in EU studies, where scholars have often developed a specific terminology in order to explain the present European supra-national experience. Nonetheless, there are many reasons for using this concept. First and foremost, the use of the concept of empire will help show that, over the years, the initial European project has changed into something which, in certain respects, is opposed to what it was meant to be and might, if not endanger its capacity to survive, at least weaken its capacity to efficiently perform its tasks. Another reason for using the concept of empire is its capacity to enable comparison with related polities. Further, this concept enables to introduce long-term history in EU studies. This must be done cautiously: concepts that make sense today might be irrelevant for a remote era and vice versa. But it is possible to bridge the time-gap – sometimes, not always; in certain ways, not in all – between older and newer epochs in order to shed a new light on today's political life. Historical comparison is indeed important in societies which have very long histories and ground part of their collective memories and institutions in ancient times. Despite these theoretical and methodological advantages, the concept of empire is not an easy one to work with.

All too often used as a metaphor of the (political) ‘Bad’, it needs to be ‘hardened’ in order to become a useful and somewhat operational concept. By necessity, the aim of this paper will thus be two-fold. On the one side, I will show how this concept can be used in a useful manner in the context of the EU. For this purpose, it must be clearly stressed that my purpose is not to ‘force’ a classification of the EU in the category ‘empire’. Cropping the edges of reality in order to make it fit in a scientific category is always a deep mistake. So, indeed, I will show that the EU does not totally fit in the category ‘empire’. But neither does it in any other traditional category of polity – for instance ‘federation’ – and nonetheless, the concept of federation is still used and considered legitimate in EU studies. Pinder (1986) has been very convincing in proposing a shift in the way the concept of federation should be used in EU studies: instead of focusing on the result (i.e., federation), one should think in terms of process (i.e., federalization). Likewise, in this paper, I will look at the European integration process in terms of imperialization, i.e., transformation processes that introduce in the europolity features, which are considered characteristic of empires.

On the other side, this paper will aim at proposing elements of definition of the concept of empire. The great majority of authors writing on empires are either historians or specialists of international relations who usually try to convince their readers, now that the USSR has disappeared, that the USA is an empire. Among the historians writing on empires, only some of them deal with empires as polities. They roughly part in two groups, according to the type of empires they study. One group studies colonial empires (Cooper 2003), the other group studies ‘old type empires’ (Hintze 1962[1907]) such as the empires of the Habsburgs, of Napoleon, or of the Romanovs (Barkey and von Hagen 1997; Motyl 2001; Lieven 2002). On the whole, they study very different kinds of empires, which have collapsed at different periods and whose collapse produced different aftermaths. Given this scattered landscape of scholars of empires, one is not surprised to find out that no single definition emerges from the writings of these authors. Zielonka's attempt to develop a typology based on structural differences is very promising, but it differentiates different types of empires without having differentiated previously empires from other types of composite polities. In matters of EU, this point is important: why use the concept of empire when the concept of federation is also available? This paper will try to bring an answer to this question.

Seven characteristics appear in most definitions of empires: empires are polities, they are composite, they consist of a dominating center and of dominated peripheries, they are temporally unstable, they follow a rationale of territorial extension, are authoritarian, and last but not least, they diffuse an imperial identity and culture – some speak of the mission of the empire. These will be dealt with in order so as to offer a progression from the description of the structure and components of empires to policy-oriented aspects. Testing each of them on the EU will be the occasion to reflect on them and either confirm or deny their relevance in a ‘modern’ definition of empire.¹

The fact that empires are polities is probably the most important component of the concept of empire because it is what differentiates them from supranational economic networks and from hegemonic relations of power. Authors writing on the subject often use terms like ‘state’ (Hintze 1962[1907]), or ‘political system’ (Eisenstadt 1961), or ‘polity’ (Ferguson and Mansbach 1996; Tilly 1997) to describe imperial modes of institutionalization of the political power.

For Motyl,

Hardt and Negri (2000) belong to a very different tradition of thought. Unlike historians, they claim not to study past empires, but an emerging one:This definition is tempting because it contains some of the characteristics of empires. However, closer attention to the internal rationale of their would-be empire shows that they describe above all an economic phenomenon. According to their description, what is emerging is a worldwide economic network, which has no center, because ‘imperialism is over’ (2000: xiv). But suppressing the idea of a center suppresses eo ipso one of the essential characteristics of an empire. In fact, what Negri and Hardt analyze already has a name, which was coined in 1974 by Wallerstein: a ‘world-economy’. Wallerstein defined this as a zone of economic interrelations, precisely one without a political center because as soon as a political center emerges, the ‘world-economy’ becomes an empire. The existence or absence of a political center indeed makes a very big difference, which can be illustrated by the fact that there is no such thing as a citizen of the WTO, whereas there are EU citizens. Of course, this argument is a bit tricky but it gives an insight into the major difference between a supranational economic network and an empire. It also suggests why Negri's and Hardt's object cannot be considered an empire.

Of course, in its first years, the European polity used to be referred to as an ‘international organization’. But recent papers now describe the EU as a ‘polity’ (Offe and Preuss 2006). The evolution of the vocabulary of EU scholars reflects the evolution of the EU itself. Indeed, the progressive transformation towards more political integration, and to a certain extent towards more statehood, has been a constant goal of some European leaders (and a constant repellant for others). All in all, throughout the years, we have experienced the development of a common legal system, the abolition of internal borders, the creation of a citizenship of the Union (albeit incomplete), the existence of a common currency (in 16 of the member states), and the election of a common parliament. Of course, the EU still lacks some major elements of statehood, such as a legal personality, the recognition of its statehood by other states, a common police, a real common foreign and defense policy. But it is clearly no longer an international organization. If one agrees on defining empires as polities, and since the EU is now considered a polity (in many respects even a ‘state-related’ type of polity), the EU clearly satisfies the first characteristic of empires.

If we go back to Wallerstein, we read that the emergence of supranational orders in the European zone covers a great deal of the European history – going back to the Middle Ages, or even before, if, like Hintze, we start with the Roman Empire. Wallerstein points out that for the last few centuries, the European continent has known many empires that never really lasted very long, unlike China, which succeeded in creating a long lasting empire. His thesis is that while China has successfully created a supranational political order, what is typical of the European supranational culture is on the one hand the establishment of a supranational economic order – his famous ‘world-economy’ – and on the other hand, the failure of all attempts to create a long lasting supranational political order.

Let us take a closer look at the idea of the failure of a supranational political order and see how it can help us shed a different light on the European integration process. The starting point is the acknowledgment that European empires have risen and fallen. But there has also been a tendency to repeat these experiences over the centuries. Each major state has had a turn in trying to create and control a supranational political order, until – according to Hintze's thesis (1962[1907]) – the lust for control was exported outside of the European territory, giving birth to what he called ‘new imperialism’ – i.e., colonial empires – and divided the planet into different zones of control, each being claimed by one major European state. If we consider these imperial histories from the view of the European continent and not as unrelated national histories, they form a European trajectory of repeated supranational experiences in which the EU can appear as the most recent attempt. However, this time, an unprecedented recipe changed the constitution of the core. Until now, the core consisted of one state that extended its zone of political control, to either neighboring or overseas territories/states. The EU has based itself on a totally different conception of the core: it is a core consisting of all that exerts its control on all. In other words, all member states are part of both the core and of the periphery of the EU. Thus, there cannot be a geographical localization, since the core cannot be reduced to one state. Of course the EU has one major ‘capital’ – Brussels – often used by the man on the street to designate the EU. But nobody ever thought that Belgium was the core of the European Communities with all the rest reduced to peripheral status. This is so because the EU's core is not a place but a stratum. It is an institutional and decisional setting to which various sectored pieces of sovereignties have been (and still are being) transferred, and which has the power to impose decisions on all member states. Where the sovereignty has been transferred, the ultimate decisional power lays in the hands of the central organs – the Council, the Commission, the Parliament, and the Court of justice. Elsewhere, the member states remain sovereign.

At first sight, my conception may come close to that of Zielonka's neo-medieval empire (2006), since it stresses the multiplicity of the core of the EU's system of governance. But Zielonka's neo-medieval empire is characterized by competing and overlapping sources of power as well as jurisdiction, by the fragmentation and privatization of authority, the absence of sovereignty, and multiple and asymmetrical allegiances (Spruyt 1994: 36–9; Zielonka 2006: 10–11). In this sense, Zielonka's neo-medieval empire is hierarchically ‘flat’. But the EU is neither ‘flat’ nor – and here I agree with Zielonka – a Westphalian hierarchical ‘pyramid’. Tightly elaborated around the concept of sovereignty or, better, ‘late-sovereignty’ (Walker 2006: 19), the EU is rather a complex ‘heterarchical’ (Walker 2006: 4) pyramid with several summits. Indeed, the system of governance of the EU organizes a functional share of the ultimate power between possible actors and can sanction reluctant ‘players’ thanks to its central Court of Justice. This reminds of the schema core/periphery, even if the core presents an atypical composition because of the functional rather than the usual structural, absolute, and monist mode of distribution of power.

The distinction between core and periphery is not just a question of borders delineating political territories and entities. Imperial cores accumulate instruments of domination and protection because empires are ‘responsible for administering and defending a huge land and population mass’ (Wallerstein 1974: 61). The political and administrative system of the cores are more efficient, they have a stronger army technologically, strategically and demographically – which means a lager population, more advanced indestry, and more powerful economy.

In many respects, the first member states of the EU built a rather homogenous zone. One may contend that until the enlargements to the South, there was no periphery. Quite contrary to imperial patterns, the European project integrated rich states and sometimes states that were richer than the founding members. It was a ‘club’ of rich states – a ‘club’ because it was made of more or less equals. With, first, the integration of Spain, Portugal, and Greece; then, the political choice to enlarge towards the East; and further the disputed decision to start accession discussions with Turkey, and in the future with the Balkans, the European Union started to accept a strong discrepancy between what now appears to be a core and a periphery – one needing large scale financial transfers in order to catch up with the economic standards of the core² and political monitoring until accession.

The 2004 and 2007 enlargements, the next one(s), as well as the spirit of the ENP, reveal the strategic use of promises of integration and of economic boosts in order to encourage, canalize, and to a certain extent force political transformations, thus linking economics and politics in ten of the 12 states that acceded in 2004 and 2007 and in potential future member states, where democracy and state institutions are fragile.

The European project turns out to be, at least at its margins, a political tool meant to diffuse a political culture. One might speak of a ‘zone of political influence’, but this also reminds us of the ‘civilizing mission’ claimed by earlier empires. This choice, which costs the core the price of integrating states with lower economic and industrial capacities, has produced and will continue to produce a differentiation between two zones, which progressively appears to be similar to the traditional division between core and periphery. Several new features contribute to and reveal the slow but deep dissociation between core states and peripheral states. One immediately thinks of the Treaties of Nice and Lisbon, and of the Constitutional Treaty, which have given an explicit institutional framework to this differentiation: ‘enhanced cooperation’ enables states wishing to go further in the integration process to do so, leaving the others outside, and thus increasing even more the difference between core and periphery, while at the same time homogenizing the composite core even more. Restrictions on the freedom of movement of workers from the new member states constitute a second and equally significant signal. The EU has decided to reintroduce against the will of the new member states the principal of frontiers – albeit on a temporary basis – in order to protect its economic core.³

At this stage of the analysis, the EU now appears to have a core that is distinguished in two dimensions, both of which correspond to traditional characteristics of classic imperial cores. The first one is the institutional and political setting – ‘Brussels’ as opposed to the member states, if one will. The second one, which normally coincides with the first one, but does not in the complicated case of the EU, is what makes up the rest of the core: the economic, industrial, military, and demographic center of power. This second dimension is not made up of a single supranational entity, but of a group of stronger and/or bigger states, the list of which may vary according to the ‘power’ considered. The consequence of this is that it creates eo ipso a second type of periphery. From the standpoint of institutions (the ‘first dimension’ of the core), all member states are to some extent peripheral zones (EU regulations are mandatory to them), but if one considers the other dimension some of these states are peripheral while others are not.

The existence of a core and a periphery leads to the question of the relationships between them. Motyl stressed the importance of this point and even made it a major characteristic of empires. For him, vertical relations characterize empires: peripheral zones do not have relations between themselves. They only have relations with the core, which appears like the hub of a rimless wheel; the peripheries stand at the end of each spike and are not linked by a rim. These relations may be either centripetal (when the dominated periphery pays its tribute to the core) or centrifugal (when the core ‘protects’ its periphery). But on the whole, the core must be in a position where it can enforce decisions on the peripheries and peripheries must ‘serve’ the core and have limited relations between themselves. In Motyl's theory, the ideal ‘result’ – the empire – is a rimless wheel (no relations between the peripheries). In addition, two ‘processes’ are likely to occur: the appearance of a rim (i.e., the development of relations between the peripheries) or the disappearance of a rim (i.e., the atrophy of these relations). The first process corresponds to a phase of imperialization or emergence of an empire, while the second one to one of desimperialization or collapse of an empire. When it comes to desimperialization, Motyl is essentially interested in the disintegration of empire: the components become autonomous states again. But the history of some federations – e.g., the German one – shows that desimperialization can occur through the transformation of the relations both between the peripheries and between the peripheries and the core, whereby the power is no longer centralized by the core at the expenses of the periphery, but shared between the different levels of government, which continue to form together a single polity. Elazar (1987) uses the metaphor of a ‘matrix’ to describe this situation where different levels of government are in charge of different sectors of governance, and considers it as typical for ‘federal arrangements’.

One can certainly contend that the core has gained much more power than it initially had. The increase of sectors where qualified majority voting has become the rule and where co-decision is required show that asymmetry has increased in favor of the center. But some scholars of federalism now contend that asymmetry, which was originally considered an explanatory factor to the failure of federations, is actually a tool for federal stability (Burgess 2006: 209–25). It is thus considered a normal feature of federations. So, on the basis of Burgess, I will contend that asymmetry characterizes both federations and empires. The difference between both types of polities is however not a difference in nature – asymmetry or not – but one in degree – empires experience a much greater degree of asymmetry than federations. At first glance, the unusual level of asymmetry that characterizes the EU brings it clearly on the side of empires: ‘there is no historical precedent for the creation of a multinational, multicultural and multilingual federation composed of 15 to 20 established national states’ (Burgess 2000: 39).

However, the mode and instruments of governance still rely to a significant point on the member states. A core and periphery might be emerging but the core is still weak, and what's more, weaker that the member states in crucial policy areas like budget, foreign policy, justice, police, where the imperial pattern does not apply. According to Motyl's model of empire, the importance of periphery–periphery relations does not allow to speak of a strait imperial pattern for the EU. The system of governance at play in the EU still corresponds too much to Elazar's matrix, which is characteristic of federal arrangements and not enough to Motyl's rimless wheel.

The disappearance of empires probably explains the great number of studies analyzing the fall of empires. Since they appear to fall very often, one is not surprised to read that empires are temporally instable (Taagepera 1997). Motyl (1992) even contended that empires contain in their very nature the causes of their own fall. Tilly (1997), on the contrary, contends that empires are the oldest, the most stable and the most frequent form of political organization in the past ten thousand years. All in all, however, the arguments elaborated by scholars of empire around these temporal aspects do not make it clear why temporality should be a defining characteristic of empires. First, empires last more or less and seem to be qualified empires regardless of their duration: Alexander's empire lasted 11 years, Rome several centuries. Second, to rise, undergo transformations, and fall, is not specific to empires.

If one really wants to tackle the question of the duration, one can do it from a more stimulating angle. When dealing with the question of the core, I had already outlined the idea that the present form of the European core can be interpreted as the product of the previous imperial experiences. Of course, the individual empires have risen and fallen, but the idea of supranationality is very old on the European continent: empires have been a feature characterizing the European continent during a period stretching from the Roman Empire to the Soviet Empire – more than two thousand years. Let us also recall that European empires have often lasted several centuries – a longevity that some European states still haven't matched. But above all, the shared experience of the Roman Empire has had deep consequences both on the characteristics of the single European states, and on the European continent, since after the fall of this empire, what was left was a coterie of states that both knew each other and had known the experience of belonging together to the same composite polity. And indeed, since then, the European zone has known several attempts to recreate empires often claiming to be the heir to the original Roman Empire – the famous translatio imperii. The link to the ‘Holy Roman Empire’ was used for purposes of political symbolic and legitimacy: the empires of the Caesars called their emperors Kaisers and Czars, both derived from the name ‘Caesar’; the symbol of the Roman imperial eagle traveled over time and across borders, from one empire to another (and still symbolizes today's Germany and Austria, albeit with two heads instead of one in the Austrian case). Interestingly, each imperial attempt was initiated by states that were powerful in their time. In other words, the core of European supranationality migrated within the continent and was experienced by most of today's big and middle-sized EU member states. Last but not least, the fact that these states were still the cores of empires when the EC/EU was founded has weighed on it since its very beginning (Muller 2001).

The mobility of empires and their rise and fall are most certainly linked with the logic of territorial expansion, which characterizes empires. For various reasons ranging from economic necessities to strategies of defense, empires tend to expand ever more and sometimes end up competing with the next neighboring empire over territories situated between them.⁴ This characteristic is all the more important, as it marks a major difference between empires and federations: empires expand, federations do not.⁵ Empires have moving borders; federations are territorial polities and thus have fixed borders. A consequence of this might be the existence of territorial zones, which are not part of the empires but have intense (above all economic) relationship with them: the ‘marches’.

The territorial criterion raises the weighty question of the geographical borders of the EC/EU. Since 1951, it has expanded – or ‘enlarged’ in ‘euro-speak’ – several times and still has no officially definitive borders although it seems to be an accepted idea that enlargement will stop at some point. The ‘no’ of the French and Dutch electors to the Constitutional Treaty expressed in part the desire to stop integrating states (or populations) that EU citizens do not identify with for some reason. One thinks immediately of the Turkish case, but in fact the enlargement to the East was already enough to trigger fears (for instance the recurrent argument of the ‘Polish plumber’ in the French campaign against the Constitutional Treaty). Hence, the (temporary) restriction of the freedom of circulation of workers from the new central and eastern European member states.

But the question remains: where will Europe stop? Charles de Gaulle once said that Europe should stretch from the Atlantic to the Urals. Jorge Semprún and Philippe de Villepin (2005) contend that Europe is an idea. But there are no self-evident answers. De Gaulle's conception was obviously inspired from the doctrine of ‘natural borders’, whereas Semprún's and Villepin's is driven by an abstract understanding of political communities. Today's EU has taken a mixed path (cf. Art. 49 and 6(1) of the TEU). Though there is still no official doctrine of borders, the European Neighbourhood Policy (ENP) has marked an important step in this direction. This policy aims to create a buffer zone between the EU and the ‘outer world’ by developing a network of privileged relationship with countries surrounding the EU (Aliboni 2005; Del Sarto and Schumacher 2005). The zone created by the ENP is intended to ensure a curious zone not just of political and economical stability but also of political and economic compatibility with the EU, since countries aimed at by this policy are offered economic aid provided they adopt the economic and political principles promoted by the EU; roughly put, market economy and democracy. The belt created by the ENP will be a soft border of countries that come close to the EU but will not be granted membership.

All in all, the territorial criterion applies very well to the EU, which has known five enlargements in half a century and will most probably experience others, even if the fifth enlargement has made it clear that the EU will have to be more restrictive in the future. The ENP, which can be considered a consequence of this enlargement and an attempt to set the final borders of the EU, is another element bringing the EU close to the imperial pattern. Because it creates a territorial zone between the EU and the ‘outer-world’, the ENP delineates a belt of countries, which are like the ‘marches’ or the ‘semi-periphery’ of empires. Let us recall here that the ENP was originally called ‘Wider Europe’, an expression which was at first awkwardly translated in French by the imperial sounding ‘La grandeEurope’ and then by a more neutral ‘L'Europe élargie’ …

The concept of empire enriched by that of ‘semi-periphery’ enables also to read differently the tormented paths followed by the states situated between the EU-15 and today's Russia: since the implosion of the USSR, we have witnessed a shift of a greater part of the periphery of the Soviet empire to the EU's zone of influence. Motyl's theory (2001) on the collapse and revival of empires applies here quite well. Former Soviet peripheries which were strong enough (i.e., CEE countries) could detach more effectively from the core that had collapsed (Russia), but not being strong enough to continue on their own, they have gone to the next strong core: the EU. The peripheries, which were not strong enough, have remained within the zone of influence of the weakened Russian core, while some uncertain cases like Ukraine have been the theater of a clash of imperial influence – a clash which has increased with the revival of the Russian core-state. Within Motyl's theoretical framework, the Orange Revolution of December 2004 can be analyzed as a consequence of the delineation of the new border between the occidental and the oriental European ‘empires’ – a border, which now passes through the Ukrainian territory, hence the crisis between the camps of the pro-European Yushchenko and the pro-Russian Yanukovych.

Most scholars agree on a perception of empires according to which they are authoritarian and on the fact that empires do not claim this aspect because they legitimize themselves as the only peaceful solution for a better world. The fact that empires are authoritarian states would rather be the consequence of imperial Realpolitik, so to speak. Motyl is one of the rare scholars to transform this acknowledgement into an element of definition: ‘More than a simple dictatorial multinational state, an empire is a highly centralized, territorially segmented, and culturally differentiated state within which centralization, segmentation, and differentiation overlap’(Motyl 1997: 21). I will take a strong position and contend that defining empires as ‘dictatorial states’ is wrong. Motyl waters down his own argument himself when he reveals what he means by ‘dictatorial’: his ‘understanding of empire, in its exclusive emphasis on the core-periphery relationship, has nothing specific to say about the regime either of the core polity or of the peripheral polity’ (Motyl 1997: 21). So the components of the empire need not be dictatorial and the only space left for it is the transmission belt, i.e., the organization of the relations between the core and its peripheries (Doyle 1986). Indeed, for him, the ultimate three defining characteristics of empires are: ‘1. a distinct core elite and a distinct peripheral elite; 2. a distinct core population and a distinct peripheral population; and 3. a dictatorial relationship between the core elite and the peripheral elite’ (Motyl 1997: 22). But this is not sufficient to justify speaking of a dictatorial regime. Dictatorial or ‘authoritarian’ regimes are said to be so because of the kind of relation of power exerted on the population – not just the regional elite. Second, he relativizes the authoritarian parameter when mentioning the formation of the Austrian Empire. Indeed if some empires were formed through a violent process (war or invasion), others have also emerged using voluntary agreements or, like the Habsburg Empire, a highly strategic marriage policy.

Obviously not at ease with this parameter, Motyl has chosen a much more prudent solution in his later work. In Imperial Ends (2001), keeping to his structural approach, he contends that the USSR was the ‘world's only totalitarian empire […] as pure a structural example of both empire and totalitarian as one can imagine’ (2001: 49). This theoretical position is ‘safer’, but it is blind to the question of ideology and civil rights (the liberal dimension of democracy), a criterion regularly used to qualify regimes, ‘democratic’ or ‘totalitarian’. It also tends to dissociate ‘empire’ and ‘totalitarianism’, since the USSR, in which both coincide, is presented as an exception. In the end, the difficulty to operationalize this parameter into a solid and convincing defining characteristic leads me to reject it. In fact, this whole discussion is fallacious.

Lieven (2002) and Beck and Grande (2004) are those who help us pose the question in different terms. Lieven does explain, ‘empire is by definition not a democracy’ (2004: xi). But he adds that ‘since very few societies indeed before the 19^th century were democratic in this sense, this distinction only acquires salience in modern times’ (2004: xi–xii). Lieven thus reminds us that empires are pre-democratic polities. This is important: the last European empires that were erected and called themselves thus were created before the first democratic wave and disappeared in the aftermaths of the democratic wave that occurred after WWII. Unable to adapt, these empires did not survive the diffusion and consolidation of liberal democracy. Afterwards, the only two polities which were called empire were the USSR and, later, the USA. Specialists of the Soviet state admit that it was an empire (Lieven 2002; Motyl 2004), but the case of the USA is much more debated. Motyl's theory, which I use in this paper, does not allow considering the USA an empire. So we are left with a political form, which seems to be extinct and for which the question of democracy/authoritarianism is anachronistic.

Beck and Grande, who have chosen to use the concept of empire in the case of the EU were aware of the difficulty to do so, because the concept of empire brings along with itself negative connotations. They therefore proposed to ‘modernize’ it and suggested the use of the concept of ‘post-hegemonial empire’ (2004: 85).⁶ In doing so, they brought us on the right theoretical path, but I think that they did not go far enough. Offe and Preuss (2006) have proposed the concept of ‘republican empire’ but consider it an oximoron. Zielonka missed the discussion when he argued the possibility of using the adjective ‘democratic’ next to the adjective ‘neo-medieval’ while avoiding the bothering topic ‘democracy/empire’ (2006: 16, 119). Motyl's structural theory of empire – because it is structural – has little to say about the type of regime used within this structure. Hence my proposition: like the concepts of state and monarchy have been ‘modernized’ in order to follow the progressive democratization of societies, it should be possible – at least at the level of theory – to accept the same possibility for the concept of empire. Empires have already been deeply affected by the diffusion of the sovereign modern nation-state. We should also accept the possibility that they be equally affected by the diffusion of democracy. As states can be pre-democratic, democratic, or authoritarian, and monarchies absolute or parliamentarian, empires can be pre-democratic, authoritarian, or, why not, republican, or democratic. Whether ‘post-hegemonial’, ‘republican’, or ‘democratic’ we would be left with the task of finding an example of this contemporary type of empire and this would point out at the EU.

Refuting the necessity to consider authoritarianism a characteristic of empire is compatible with the structural theory I am using in this text. It also opens up the door for the use of the concept of empire in democratic contexts. This new option provides us with a new concept, in the category of composite polities, which enables to grasp notions to which the concept of federation is blind – such as territorial expansion or territorial marches.

Identity problems are often mentioned among the reasons leading to or linked with the fall of empires, so identity should be an important variable for the study of empires. Yet it is seldom transformed into an operational variable. The difficulty in doing so could explain why this is so, but my guess is rather that scholars are heirs to the definition of the state provided by Max Weber: an organization that monopolizes the use of legitimate physical coercion. This isn't wrong, but in the era of the Rechtsstaat and of liberal democracy this monopoly actually works best when it remains a latent threat – to be carried out only if need be. A state cannot function solely with the use of coercion. It needs some support that can be obtained without asking for it. This is where another kind of monopoly comes into play: the monopoly of definition of a legitimate collective identity. The identification of citizens with their state means that they accept it and leads them to act for it. Collective identity is therefore a very strategic tool. If efficiently wielded, states can more easily avoid exerting physical violence, be respected by their citizens, and obtain more from their citizens than under the use of constraint.

Seen this way, one understands a problem faced by all empires. The fact that they are a type of polity in which one cultural group (the ‘core’) exerts political domination over others (the ‘periphery’) necessarily raises the question of the legitimacy of the empire and of the dominant society. At a certain point, most empires have been weakened by the fact that their population didn't identify sufficiently with them – in other words by a lack of legitimacy. Let's be clear: there is probably no direct causality between claims of alternative identity and the fall of empires. Other factors must come in play to explain their collapse, e.g., external pressures or the costs of empire. However in recent centuries the rhetoric of nationalism was always present, at least to catalyze mobilization against empires. Deàk (1992: 225) has tried to show how the Habsburg Empire was able to prolong its existence by giving cultural legitimacy to peripheral groups, and how, on the contrary centralization and ‘great Russian/Slavic nationalism may well have contributed to the […] crisis of the Soviet army and hence also to that of the Soviet empire’. His argumentation thus functions in both directions: collective identity wisely handled can prolong empires, whereas disregarding it may well contribute to their breakdown.

Collective identity in empires intertwines two levels of identity. The ‘level one’ is ‘supranational’ and takes the form of a civilizing mission justifying the creation and extension of the empire over the dominated territories. In order to surpass the particularisms of the imperial components, it must take the form of a universalistic ideology, usually claiming to promote peace and/or a better world. The ‘level two’ is that of national/regional identities. Here, the empire does not intervene as a producer of identity content, but of solutions that respect regional identities well enough so as to be accepted by the peripheral subjects or citizens. These two levels can be combined in many ways, usually implying respecting local languages, religious or philosophical beliefs, and other cultural practices and institutions.

Like the very existence of European empires, ‘the idea of Europe as a community with common values and institutions is an old one’ which ‘in the Middle Age […] was identified with Christendom’ and whose ‘component parts [shared] a common legacy in the Roman Empire’ (Keating 1999: 427). The place of Christendom and more generally religion in the European identity has become problematical nowadays. The secularization of some European states (such as France) precludes a strong commitment to religion, let alone any particular one. In fact, the EC/EU has always used another kind of cement for the ‘level one’ of identity: the (imperial sounding) promise of peace and wealth on the tormented European continent. The booklet Europe in 12 lessons, written in 2003 by Pascal Fontaine,⁷ a former assistant to Jean Monnet, explains this to European citizens:

Besides a set of values, the EU's ‘level one’ of identity also proposes symbols such as a flag and an anthem. But the EU also makes a lot of efforts to protect the ‘level two’ of identity. In this respect, the linguistic policy is most interesting. Always worrying that the EU could be interpreted as an attempt by stronger states to swallow up smaller ones, EU officials and leaders have always made major efforts to exhibit the cultural plurality of the EU among others through the maintenance and protection of languages. That was the object of the first regulation of the European Council dated October 6, 1958. It established the principle of one official language per member state. The EU has always maintained this principle although it was often challenged. Thus, when Neil Kinnock presented the strategy for the Commission's interpretation service in preparation for the 2004 enlargement, he introduced his memorandum with the following statement: ‘The forthcoming enlargement of the European union presents a major challenge […]. Multilingualism will remain a founding principle of the European project’.⁸ Within the institutional setting of the EU, the multiplication of languages has firstly lead to the creation of two separate directorates dedicated, respectively, to translation (DGT) and interpretation (SCIC). Second, the European civil service is multilingual with individual bilingualism as the minimal linguistic requirement for recruitment. The reason why these requirements and services were set up can be found on the EU's website:These two quotations reflect the conception that languages are not just means of communication, but also political tools. Neil Kinnock makes it very clear: ‘It is widely agreed at institutional level nowadays that the use of the national languages is essential for the legitimacy of the Union’.¹¹ For the DGT, multilingualism enables democracy in general and participatory democracy in particular (‘the public […] must be able to play an active role’). The SCIC, on the other side, provides its services essentially to states and parliamentarians. Hence the stress on representative democracy: EU citizens can send ‘their best experts’ and not ‘their best linguists’ – a powerful formulation. Making sure that all communications with the civil society are systematically proposed in all languages enables everybody to see that indeed the EU does not endanger their cultural identity.

In this paper, I have argued that the concept of empire can be used to analyze the EU (see Table 1). Following Motyl's structural theory, I have defined the concept of empire as a composite polity made of a core and peripheries, which are bound by vertical and asymmetrical relations. I have added two policies, which directly produce or are the direct product of the structural elements of the definition of empires: the territorial policy of expansion and the identity policy. Lastly, I have argued that authoritarianism as well as temporal instability should not be considered defining characteristics of empires.

Given these elements of definition, if one tries to analyze in terms of result, the EU cannot be qualified an empire because one major characteristic does not apply properly. However, if one analyzes in terms of process I contend that the recent evolution of the European polity reveals clear aspects of imperialization. The reasons for this rest in the successive enlargements and in particular the fifth one.

The intention of this paper is not normative: whether the imperialization is good or not is not to be decided here. But it is possible to confront the implications of imperialization to the self-definition of the EU. The way the EU has been dealing with multiculturalism and national identities over the decades has remained true to its original intention. Here, imperialization appears to be a mere prolongation of the initial European project and puts the EU on the side of composite polities having understood that multiculturalism in a composite polity is a tool of legitimacy and thus of longevity. Things are different when it comes to asymmetry. One must raise here the question of whether the growing asymmetry in the EU is compatible with the rough equality, which underlies the relations between the first member states, all the more as the next enlargements will increase the present degree of asymmetry. In time, such an asymmetry could be a source of serious contestation within the EU. Here imperialization does not appear as a prolongation but as a source of transformation of the European project. Increasing asymmetry being the result of enlargements, European decision makers should rethink the policy of enlargement. Indeed, ‘polities that expand faster also tend to contract sooner’ (Taagepera 1997: 480). This does not mean that enlargements should be stopped, but their frequency should at least be reduced. Firstly, this would grant the new member states time to ‘blend’ into the EU and thus to reduce the overall degree of asymmetry. Second, citizens would have more time to identify with the new EU. If the process of imperialization seems to bring about transformations in the European project on which EU decision-makers should seriously reflect, it is by no means irreversible. History has shown in the cases of the USA and of Germany that empire can also be a stage before the emergence of another type of polity, which was in these two cases a federation.

MagaliGravier is Associate Professor at the International Center for Business and Politics at the Copenhagen Business School. Her present research areas are the EU's staff policies, the European integration process, and the study of loyalty. She recently published ‘The 2004 Enlargement Staff Policy: the case for representative bureaucracy’, JCMS, 2008, 46(5): 1025–47; and Good Bye Honecker! Identité et Loyauté dans les administrations est-allemandes (1990–1999), Paris: Presses de Science Po, 2008.