International functionalism and democracy

Abstract

International functionalism as proposed by David Mitrany envisions non-territorial functional agencies to supplant the states system. Mitrany makes no provision for democracy in functional agencies. Instead, he assumes that the policies of international technocrats would be ‘technically self-determined’ and uncontested – a stance several critics deem antidemocratic and naive. A second, related criticism holds that even if functional agencies were formally democratic, democracy could not operate effectively since functional polities would be too ‘thin’ and fragmented to sustain democratic commitments among their members. The article qualifies the first charge and rejects the second. First, defined as an institutional decision-making principle, ‘technical self-determination’ is Mitrany’s add-on to the underlying functionalist logic, not an inseparable part of it. That logic instead holds that institutions work best if their form and scope of authority follow their function and that function-specific agencies therefore could meet some needs better than the state. Contra Mitrany, this does not privilege technocratic over democratic decision-making and it does not imply that functional agencies would be free from political conflict. Nor, second, would functional agencies necessarily be unsuitable for democracy in practice. Several strands of democratic theory suggest that even people who do not share a ‘thick’ communal identity can develop a commitment to meet shared needs democratically. If one accepts this, it loosens functionalism’s technocratic stigma, highlights its potential as a democratically viable alternative to both state-centric and supranational models of international order and broadens our conception of possible democratic futures.

Introduction

International functionalism remains a frequently cited chapter in the intellectual history of international integration but as a widely debated blueprint for the reorganization of the international system its popularity peaked long ago. This decline in interest was in part inspired by empirical developments such as the failure of the European integration project and United Nations specialized agencies to evolve into what international functionalists had advocated and to some extent predicted. Theoretically, meanwhile, the decline of functionalism as a dominant analytical paradigm in sociology and anthropology from the mid-1960s seemed to affect its international variant, too. Critics variously accused international functionalism of being economically reductionist, apolitical, teleological and technocratic (Long, 1993; for a classic overview see Pentland, 1973: Chapter 3). Some deemed functionalism’s project of transferring powers from national governments to international technocrats outright ‘impossible’ (McLaren, 1985: 140) as they thought that it underestimated the resistance this would spark among national governments and mass publics alike.

The past two decades have seen some resurgence of interest in international functionalism and a re-evaluation of the work of David Mitrany as its ‘founder’ and most influential exponent (e.g. Ashworth, 2017; Hammarlund, 2005; Holthaus, 2018; Steffek, 2015). Much of this work is inspired by the same underlying concerns that also animated Mitrany: a (qualified) scepticism of the state as a model of social and political organization and – by implication – a leeriness of centralizing and territorial models of international integration that seek (partially) to reproduce state-like structures at a supranational scale.

This article seeks to contribute to the current revival of interest in Mitrany’s work by examining its relationship with democracy. It focuses on two related criticisms that have often been levelled against international functionalism from this perspective. A first criticism highlights Mitrany’s failure to provide for democratic representation inside the functional agencies he envisions. This lack of democratic provision reflects Mitrany’s doctrine of ‘technical self-determination’, his claim that international technocrats could uncontroversially ascertain, prioritize and satisfy needs. Yet according to this line of criticism, Mitrany’s attempt to substitute democracy with ‘technical self-determination’ is illusionary since prioritizing and satisfying needs is inevitably contentious and thus political. Rather than being accepted as ‘technically self-determined’ and thus beyond rational contestation, policy outputs in non-democratic functional agencies would risk being seen as illegitimate impositions by democratically unaccountable international technocrats. The second criticism holds that even if functional agencies were furnished with formally democratic institutions these could not operate effectively in practice since functional polities would be too ‘thin’, fragmented and lacking in communal cohesion to sustain democratic commitments among their members. In such a conception, democracy could not be ‘unbundled’ and deterritorialized in parallel with an ‘unbundling’ and deterritorialization of political power as advocated by Mitrany.

The article qualifies the first criticism and rejects the second. It accepts the first charge to the extent that Mitrany is indeed naive in his assumption that prioritizing and satisfying needs could ever be ‘technically self-determined’, uncontested and in that sense apolitical. Yet defined as a principle of institutional decision-making, ‘technical self-determination’ is Mitrany’s add-on to the underlying functionalist logic, not an inseparable part of it. That logic suggests that institutions work best if their form and scope of authority follow their function and that function-specific agencies therefore could satisfy some needs better than the state. Contra Mitrany, this does not privilege technocratic over democratic decision-making within functional agencies and it does not imply that those agencies would be free from political conflict. More than that, assuming that non-democratic functional agencies would find it difficult to gain popular acceptance, democracy would likely constitute a need – or ‘functional requisite’ – in its own right which functional institutions would have to satisfy to be seen as legitimate by their members.

Nor, as the second part of the article argues, would their relative ‘thinness’ and lack of communal underpinnings necessarily render functional polities unable to engender shared democratic commitments among their members. Drawing on several strands of democratic theory, I argue that such an assumption (1) overestimates the strength of overarching communal attachments in many domestic democracies and (2) by extension exaggerates the extent to which such attachments are necessary for democracy. Democratic polities instead can be associative rather than communal, based on their members developing an awareness of shared needs and a commitment to address those needs democratically. Moreover, such commitments can form even in polities that are neither ‘multifunctional’ nor territorially based.

In its unwavering faith in technical progress and enlightened administrative rule Mitrany’s functionalism is a product of its time. But in its core assumptions it is prescient and profoundly worthy of intellectual resuscitation. Assuming that human needs are becoming ever more complex and differentiated, Mitrany proposes correspondingly differentiated institutions to meet those needs and a decoupling of political authority from territorial demarcations. International functionalism’s popularity has suffered among other things from a belief that democracy could not be diffused and deterritorialized alongside a diffusion of political power and that functional institutions could therefore not be democratic. By attributing such pessimism to an unduly narrow view of democratic prerequisites, the article seeks to free international functionalism from its technocratic stigma and to foreground its potential as a democratically viable alternative to both state-centric and supranational models of international order.

International functionalism

Mitrany uses the terms ‘functional theory’ and ‘functional approach’ interchangeably. His functionalism is nonetheless closer to an approach than to a (strictly non-normative) social science theory. Mitrany combines strands of utilitarianism, early interdependence liberalism and especially early British pluralism into an ambitious blueprint for the reorganization of the international system. Containing analytical as well as normative elements, this blueprint reflects Mitrany’s own lifelong role as both social scientist and political advocate.

In sketching out international functionalism, its limited relationship with structural functionalism as a broader analytical paradigm in the social sciences is a useful starting point. Mitrany loosely echoes structural functionalists such as Durkheim and Parsons in arguing that social evolution and technical innovation engender a continuous differentiation of needs, roles and functions and thereby make individuals more interdependent. For both Mitrany and many structural functionalists, differentiation and interdependence define the logic to which social institutions must adapt. At that point, however, Mitrany parts company with structural functionalism. For structural functionalists, functional adaptation typically serves to meet presumed needs such as ‘cohesion’, ‘integration’ or ‘reproduction’ which they attribute to correspondingly abstracted needs-holders such as ‘social systems’, ‘social structures’ or ‘societies’. ¹ This focus on presumed systemic needs and needs-holders helped spawn well-known charges that structural functionalism reifies social structure and ‘dehumanizes’ needs in addition to being tautological, charges that helped propel structural functionalism’s rapid decline in popularity from the mid-1960s. ² Mitrany’s international functionalism, by contrast, ‘is built from the opposite direction, focusing on the needs of individuals, their relations to one another and also to the [. . .] arrangements that help serve those needs’ (Steele, 2011: 19–20). Needs for Mitrany are meaningless unless they are attributed to individual human beings while institutions serve no justifiable purpose other than meeting those needs. In that sense, Mitrany’s functionalism is individual-centric both normatively and analytically.

A second defining assumption in Mitrany’s functionalism is that people are best able to cooperate in areas which directly affect them and in which they share expertise and that such cooperation can evolve organically and without central coordination. His belief in people’s innate cooperative inclinations places Mitrany in a tradition of early 20th-century British pluralists including guild socialists such as G. D. H. Cole and liberals such as Leonard Hobhouse who was his academic mentor. Though they differed in the extent of their radicalism and transformative ambition (Holthaus, 2018), pluralists concurred that inside existing (liberal) states patterns of voluntary association and cooperation were already evident and ripe for expansion. These states were constituted not by singular ‘peoples’ or ‘societies’ but by a ‘complex of associations held together by the wills of their members’ (quoted in Long, 1993: 360; also see Wright, 1978. For pluralism-inspired models of associative democracy, see Hirst, 1994). Like Mitrany after them, many pluralists sought to deepen such associations and extend them across state boundaries, gradually submerging states in transnational webs of cooperation and interdependence.

Mitrany’s main problem with the state-centric status quo is what he sees as a mismatch between evolving human needs and the state’s institutional and territorial rigidity. If needs are becoming ever more differentiated and interdependent yet states remain bounded and static institutional ‘bundles’, states’ ability to meet needs must steadily be diminishing. In its broad contours such a critique is not of course unique to Mitrany. It resonates in various proposals to bypass the state’s limitations through international integration, defined as a process through which existing states (partially) subsume themselves into larger overarching units. Mitrany, by contrast, rejects such merger solutions. He fears that attempts to build supranational unions would spark resistance from sovereignty-conscious governments and risk becoming side-tracked by issues other than human needs such as sovereignty and ‘member state rights’. Attempts to settle such issues through (quasi)-constitutional divisions of power in turn would reproduce the state’s structural rigidity at a supranational level. Besides, supranational integration projects would reproduce rigid territorial divisions. New political-territorial lines of inclusion and exclusion ‘would bind together some interests which are not of common concern to the group [of participating societies], while they would inevitably cut asunder some interests of common concern to the group and those outside it’ (Mitrany, 1966 [1943] 69). To overcome the rigidity and territorial divisions associated with existing states, Mitrany advocates new institutional arrangements that would grow organically to reflect evolving needs and interdependent relationships while remaining flexible and ‘allergic to set formulae and forms’ (Mitrany, 1975: 255):

If one were to visualize a map of the world showing economic and social activities, it would appear as an intricate web of interests and relations crossing and recrossing political divisions – [. . .] a map pulsating with the realities of everyday life. They are the natural basis for international organizations: and the task is to bring that map, which is a functioning reality, under joint international government [. . .]. The political lines will then in time be overlaid and blurred by this web of joint relations and administrations. (Mitrany, 1948: 358–359)

By ‘international government’, Mitrany does not mean a government in the form of a unified set of institutions governing over a bounded territorial space made up of several (formerly) sovereign states. As discussed, he rejects the divisions and rigidities this would (re)create. As an alternative, Mitrany (1948) instead advocates the gradual construction of a web of non-territorial, ‘full-fledged but specific and separate functional agencies’ (p. 360). The form and remit of these agencies as well as their policies would follow the principle of ‘technical self-determination’, which Mitrany hails as functionalism’s ‘cardinal virtue’ (Mitrany, 1966 [1943] 72):

The functional dimensions [. . .] determine themselves. In a like manner, the function determines its appropriate organs. It also reveals through practice the nature of the action required under given conditions, and in that way the powers needed by the respective [functional] authority. The function, one might say, determines the executive instrument suitable for its proper activity, and by the same process provides a need for the reform of the instrument at every stage. (Mitrany, 1966 [1943]: 72–73, emphases in the original)

Mitrany suggests that functional agencies should start out in strongly technical and ‘low politics’ areas of administration as this would be relatively uncontentious. Even so, his aim is not to separate the technical from the political so much as to ‘fuse’ the two. Even issues that are traditionally highly politicized would ultimately become subject to widespread expectations of functional needs-satisfaction. At that stage they would stop being contentious and, in that sense, cease being political (Ashworth and Long, 1999: 8). The assumption is that for as long as states retain a monopoly on tackling societal problems, their citizens tend to view these problems through distortive ‘state lenses’, tainted by concerns for national sovereignty, power and prestige. But once needs had been disaggregated and responsibility for their satisfaction diffused to bespoke functional agencies, the appropriate technical solutions would become readily obvious to all. In that way, Mitrany believed that politics could be separated into layers of relatively uncontentious functional service provision. As functional institutions would satisfy needs more effectively than the state, their legitimacy would grow and they would benefit from a gradual shift of popular loyalties (Taylor, 1968: 398).

Mitrany shares a strong interest in centralized planning widespread among his contemporaries. Accordingly, he focuses more on the upward delegation of powers to functional agencies ‘above’ the state than on their devolution. Yet such a centralizing slant is not intrinsic to the functionalist blueprint. Some needs are such that the corresponding functional agencies would be global and universal, counting all people on earth as their constituents. Other agencies would be local. Still others would be global or continental but not universal and their membership would be geographically dispersed (e.g. encompassing particular industries or professions) (Ashworth, 2017: 7). To describe international functionalism as a blueprint for regional integration therefore is doubly problematic. Neither would it be confined to any particular region (though it might commence at a regional or sub-regional scale), nor would it integrate smaller units into a larger supranational entity. What Mitrany advocates instead is a fully-fledged qualitative reconfiguration of political space. The states system would become overlaid with a network of functional agencies as states’ erstwhile functions would become disaggregated and diffused upwards, downwards and sideways.

The fate of the state in this potential functional transformation process remains ambiguous. When wearing his hat as a pragmatic political advocate, Mitrany reassures states that functional agencies would coexist alongside them and that national governments would retain control over which powers would be transferred. This emphasis on state control established a conservative, ‘minimalist’ interpretation of international functionalism as essentially a framework for inter-state cooperation (Imber, 1984). Elsewhere, Mitrany advances a more transformative and utopian reading of international functionalism. That reading emphasizes international functionalism’s expansive logic and its concern for the needs of individuals as opposed to social constructs such as states – a normative orientation that squares with Mitrany’s individualism and leads one commentator to categorize him as a ‘cosmopolitan functionalist’ (Steffek, 2015). In this interpretation, individuals – not states – would be the constituents of functional agencies whose legitimacy would grow in proportion to which they met individual human needs (Ashworth, 2017: 7).

International functionalism’s plausibility depends in large part on whether one accepts its underlying claims (1) that human needs are becoming increasingly differentiated and (2) that functionally disaggregated institutions could best meet those needs. If one accepts both assumptions, Mitrany’s approach drives a potentially powerful critique not just of the territorially bounded and institutionally ‘fused’ state but also of integration projects which partially seek to reproduce state-like structures in Europe and elsewhere. Accordingly, for international functionalists the European Union’s (EU) enduring preoccupation with how power should be divided between the national and supranational levels only highlights its supposedly anachronistic character. It flies in the face of what Mitrany sees as continuously evolving and differentiating human needs which he believes necessitate a disaggregation of institutions and a transcendence of territorially defined political units and ‘levels’, be they national or supranational. Put differently, Mitrany’s functionalism resembles many alternative blueprints for international (re)organization in that it seeks to impose a form of order on an anarchical international system. However, it envisions that order without new spatial demarcations and without an overarching centralized hierarchy. ‘A world based on functional organisations remains an order without an orderer’ (Ashworth, 2017: 12).

International functionalism has often been mischaracterized as a close intellectual ancestor of neofunctionalism in the study of International Relations. In reality, the link between them is fairly tenuous. The two overlap to the extent that neofunctionalists broadly accept Mitrany’s notion of gradualism which they sought to theorize much more systematically than Mitrany did through the concept of spillover. In broad, spillover would occur as initial integrative moves would produce functional pressures for further integration. In parallel with this, initial shifts of power to supranational institutions coupled with political engineering by supranational officials would lead domestic interest groups (and ultimately also mass publics) to transfer their expectations and loyalties to an emerging supranational centre, which in turn would facilitate further transfers of authority (Pentland, 1973: Chapter 4; Tranholm-Mikkelsen, 1991). Also, like Mitrany (see below), neofunctionalists were largely unconcerned with democracy in international institutions (while assuming that interest group–based spillover could probably only happen where integration involved pluralist liberal democracies). However, a crucial difference between Mitranian functionalism and neofunctionalism is that in their quest for a non-normative social-scientific framework to capture integration in (mostly) Western Europe, neofunctionalists abandoned Mitrany’s structurally transformative ambitions. They envisioned integration to produce a merged territorial entity, possibly some kind of supranational state. This is the diametrical opposite of the deterritorialized and ‘unbundled’ functional agencies Mitrany had in mind. Its centralizing slant accordingly led Mitrany to dismiss neofunctionalism as a version of ‘semi-functionalism [. . .] stuck firmly in the old sovereign-territorial concept of political organisation’ (Mitrany, 1971: 535, emphasis omitted; also see Navari, 1995). Mitrany’s transformative ambitions were thus lost even on many of his successors who claimed for themselves the functionalist mantle.

Functionalism and democracy

A well-rehearsed line of criticism accuses Mitrany’s functionalism of being technocratic and based on an overly consensual conception of politics (Long, 1993; McLaren, 1985; Pentland, 1973: Chapter 3). Few commentators discuss Mitrany specifically in relation to democracy, however. Those who do tend to give short shrift to Mitrany’s democratic credentials, highlighting the far-reaching absence of democracy from his blueprint further discussed below (Long, 1993). Even so, Mitrany’s stance on democracy seems ambivalent. On the one hand, he outlines a sophisticated and prescient critique of what he sees as the deficiencies of democracy inside industrialized states and proposes ‘functional representation’ as a partial antidote. On the other hand, he remains vague on what he means by ‘functional representation’ while his blueprint for functional agencies contains no tangible democratic elements.

Mitrany’s critique of domestic democracy again broadly echoes various strands of early British pluralism. Pluralists argued that a person’s political interest is defined by their needs. Effective representation must thus be in relation to those needs and the institutions tasked with satisfying them. As Cole puts it, ‘[r]epresentation must be “specific and functional,” not “general and inclusive”’ (quoted in Wright, 1978: 233). Democratic states fall short on both counts. Because domestic democratic representation is both ‘general’ and ‘inclusive’, citizens help decide many issues in which they have no stake and little knowledge just as their ability to shape policy outcomes in which they do have a stake is diluted by the equal participation of those who do not – a mismatch which pluralists believed compromised the quality and effectiveness of democratic representation (Holthaus, 2014: 725). This was compounded by the ‘multifunctional’ and territorially circumscribed nature of domestic governments, which pluralists believed made them too blunt an instrument to meet many needs fully. In short, in this interpretation domestic democratic systems prevent the effective articulation of needs ‘upwards’ from needs-holders to governments just as governments are structurally inhibited from fully responding to those needs.

Mitrany echoes such an interpretation. Moreover, he argues that the expansion of the welfare state in industrialized liberal democracies after the Second World War further reduced the quality of democratic representation in those democracies. Reflecting the increasingly technical nature of welfare governance, power was continuously moving upwards from local to central government as well as sideways from elected representatives overburdened with the complexities of modern government to unelected technocrats populating a myriad of expert bodies. Both shifts, according to Mitrany, diluted citizens’ democratic control still further and threatened to degrade domestic democracy into a substantively empty series of electoral rituals with actual power moving to actors beyond democratic reach (Mitrany, 1955, 1971). The paradox in this, Mitrany observed, is that the expansion of the welfare state empowers citizens economically but disempowers them politically. It means ‘the more government for the people, the less government by the people’ (Mitrany, 1971: 540, emphases added) as citizens would find it difficult to control increasingly remote and technocratic modes of administration.

For Mitrany, ‘functional representation’ in functional agencies would offer an effective antidote to this crisis of representation inside domestic democratic systems. Bypassing the state, functional agencies would connect needs-holders directly with policy-makers in the relevant functional areas, ensuring a more effective articulation of needs and greater institutional responsiveness to those needs. Functionalism, he contends, ‘offers a valid remedy for the growing power and insulation of the administrator. It brings together those who know with the things they know, and in which they can both initiate policy and judge its performance’ (Mitrany, 1971: 540).

Mitrany’s account of ‘functional representation’ raises important questions. Even if one accepts his claim that technocratic governance undermines democratic representation within states but not within functional agencies, it is far from clear that the representation he has in mind would actually involve all individual members of a given functional agency. In several passages, Mitrany seems to want to limit representation to vaguely defined stakeholders or – as in the quote above – to equally ill-defined cohorts of ‘those who know’ while leaving it unclear whether holding a given need would furnish someone with sufficient stake or knowledge to participate. A notion of democracy as something less than universal representation reveals itself most clearly when Mitrany discusses how functional agencies would work in practice. He makes no provision for broad-based popular participation in policy formation or in the selection of officials. Instead, Mitrany insists that functional administration would be of a ‘purely technical nature’ (Mitrany, 1966 [1943]: 80). Functional agencies would be staffed by technical experts recruited based on their expertise by selection panels themselves composed of unelected experts. The same technocratic makeup would define several administrative layers that would monitor and to some extent coordinate the work of individual functional agencies (Mitrany, 1966 [1943]: 73–76). To the extent that there would be any broader-based representation at all, it would be in the form of policy consultations with stakeholders from the relevant functional sectors. These would represent fields such as industry and science and would advise managers in functional agencies. As a model, Mitrany (1966 [1965] repeatedly praises the International Labour Organization’s system of formalized employer and trade union involvement, crediting it with providing an ‘effective process of democratic representation’ (p. 206). In reality, such selective and highly indirect representation is a far cry from universal democratic representation based on one-person-one-vote. Mitrany also fails to show (or even consider) how it would avoid further privileging the very vested interests whose influence a system of democratic accountability presumably is meant to constrain (Long, 1993: 375). In short, in Mitrany’s functional agencies, determining, prioritizing and satisfying needs would be the preserve of technocrats with at best limited representation from the functional sectors concerned. There is no provision for elections, referendums or other ways of involving mass publics directly in the decision-making process.

Seen in this light, Mitrany’s stance on democracy and ‘functional representation’ reveals the influence of early British pluralists on his thinking but also the limits of that influence. As suggested, both Mitrany and many pluralists doubted the state’s ability to fulfil needs optimally and to provide for effective democratic representation. Both wanted to organize society into functionally delineated associative bodies cooperating across and blurring national divisions. Yet contrary to Mitrany, most pluralists were impeccably wedded to democracy. What animated them was not just (and not even primarily) a quest for more efficient needs-satisfaction à la Mitrany. More importantly, they believed that functionally delineated cooperative associations would enhance human freedom and allow for grassroots democracy beyond what they thought possible within both the modern state and the hierarchical structures of corporate capitalism. Mitrany shares this belief in the inadequacies of domestic democracy but his functional blueprint offers no democratic alternative. As Mitrany ‘sanitized the functional approach [by his intellectual predecessors] in his construction of an academically viable thesis [. . .] he de-radicalized it’ (Long, 1993: 376), leaving democracy as an important casualty.

Mitrany’s omission of democracy from his functional blueprint thus contrasts with his concern for democracy inside states. At the same time, it aligns with other elements in his thinking, above all with a particular understanding of ‘technical self-determination’. As implicit in the earlier discussion, Mitrany gives two distinct meanings to ‘technical self-determination’ even though he does not clearly separate between them. In a first reading, ‘technical self-determination’ represents a structural organizing principle. It demands that the form, size and membership of functional agencies should reflect the technical dimensions of their functional remit – in other words, that form should follow function. This raises many practical questions but the underlying logic is coherent in principle: tools often work best if they are designed for the specific purposes for which they are being used. According to this rationale, if needs are becoming ever more differentiated, so too should the institutions used to satisfy those needs. In a second reading which Mitrany adopts parallel to the first, by contrast, ‘technical self-determination’ denotes a kind of operating doctrine for decision-makers in functional agencies. After ascertaining needs as ‘a matter of factual audit’ (Mitrany, 1948: 358), functional technocrats would decide how to satisfy them based on their technical dimensions alone. The implication is that this would allow functional agencies to dispense with political conflict and with democracy as a means of resolving it. If (1) people have needs which objective and incorruptible technocrats can uncontroversially ‘audit’, (2) these needs, once sufficiently disaggregated, prescribe functions best able to satisfy them and (3) these functions prescribe policies which are technically rational and politically non-contentious, there would be no room, need or demand for democratic deliberation and decision-making. There would be nothing left to contest, debate and decide. Without political conflict, democracy serves no purpose.

Yet the claim that functional governance could be ‘technically self-determined’ and free from conflict is problematic and, not surprisingly, is a staple target for many of Mitrany’s critics. Even if technocratic ‘needs auditors’ could uncontroversially separate ‘real’ needs from mere interests, preferences, urges and desires (many of which would undoubtedly be couched by their proponents in the language of need), a world in which needs invariably outstrip resources available to satisfy them fully would require choices between needs of prima facie equal validity. Because they would advantage some and disadvantage others these choices would be contentious. And even if some needs could be prioritized uncontroversially, different means are often equally functional in satisfying a given need without there being ‘one best way’ (McLaren, 1985: 143). Cultural norms, ideology and material self-interest are among the factors that may help shape how different actors prefer to address a given need. Finally, even if functional technocrats could use relatively uncontroversial technical criteria when prioritizing and satisfying a need, this would not necessarily extricate them from political controversy. Rather, their choices would remain subject to our tendency to place ‘small’ functional issues with their correspondingly ‘small’ technical solutions back into a wider political or ideological context. That context is often defined by competing notions of justice, fairness and the common good that are neither ‘technical’ nor ‘functional’ because they do not exist as technical answers to functional problems in the first place. The creation of an integrated water utility in the Republic of Ireland a few years ago and some of the largest street protests in Irish history which it sparked is a fascinating example of how attempts to depoliticize a project by framing it as a purely technical response to a functional challenge can go awry. ³

In short, disaggregating political problems into their functional component parts does not necessarily neutralize their contentious nature. Nor is there much to suggest that politics is morphing into a form of uncontentious public service provision as Mitrany implies or that moving it from the state to functional agencies could ever make it so. If one accepts this, it raises three interrelated questions. First, is a belief in the inevitable endurance of political conflict compatible with the functionalist logic or does it negate that logic – in more practical terms, could functional agencies operate even in conditions marked by continuous political conflict? If so, second, is democracy the standard by which people in functional agencies would expect to address such conflict? If so, third, could functional agencies meet those expectations by operating democratically? The following sections take up these questions in turn.

Political conflict and democratic expectations

To tackle the first question, it is useful to recall the earlier distinction between two broad claims Mitrany is making (but not clearly separating). The first claim encapsulates the core principle of functional organization by stating that non-territorial function-specific agencies could meet some needs better than the state. The second claim is that decision-making within those agencies would be ‘technically self-determined’ and uncontested. As argued, that second claim is deeply problematic but it is not a logical corollary of the first. To contend that functional agencies could govern better than the state in some areas does not specify how they would govern. It does not imply that policy-making in functional agencies would be or would have to be ‘technically self-determined’ and free from political conflict. It is logically consistent to accept Mitrany’s assertion that functional agencies could meet some needs better than the state while rejecting his claim that their policies would be uncontested.

Nonetheless, even if one accepts this it does not show that functional agencies would necessarily have to be democratic. Could they not instead be governed by, for example, autonomous expert committees or enlightened functional autocrats? On this question it is again useful to recall the central functionalist tenet that institutions should be shaped by the needs they seek to satisfy, that is, that their form should follow their function. Those needs would include primary human needs such as health and education. But they would also include needs related to the operation of the relevant needs-satisfying institutions themselves, that is, needs linked to the selection, prioritization and implementation of needs, functions and policies. Those operational needs would likely include democracy, at least to the extent that rank-and-file members of functional agencies hailed from domestic democracies and had internalized democracy in its two basic dimensions: as a procedural norm determining the ‘input-oriented legitimacy’ (see Scharpf, 1999, pp. 7–21) of political decisions and as a mechanism to hold political institutions to account. Surveys consistently show that in many democracies around the world citizens’ support for democracy as a governing principle is strong and ubiquitous, despite frequent misgivings about its practical operation (Norris, 2011: Chapter 5). Where citizens embrace democracy as a core political value, they would likely perceive a transfer of powers from (democratic) domestic governments to (non-democratic) international functional agencies as an act of democratic disempowerment. They would likely perceive it in this way even if functional agencies featured (partially) effective non-democratic accountability mechanisms (e.g. some kind of Hong Kong-style system of administrative meritocracy). Such non-democratic functional agencies assuring mass publics that democracy was unnecessary since its technocrats would flawlessly ‘audit’ needs and heed ‘functional requisites’ would likely fall on profoundly deaf public ears. The erosion of public support for the EU in several member states partially due to its perceived ‘democratic deficit’ supports such an assumption. It suggests that national publics may retain democratic expectations even with respect to institutions that operate in an international (albeit in this case supranational rather than functional) context especially if these institutions directly and visibly affect aspects of their lives. Citizens do not necessarily become indifferent to democracy simply because the institutions concerned operate outside the state. ⁴

To summarize, Mitrany’s claim that functional agencies could meet some human needs better than the state neither assumes nor demands the disappearance of competing values, interests and preferences. It does not imply that the policies of functional agencies would be ‘technically self-determined’ and it does not call for technocratic rule. On the contrary, democracy would likely be the standard by which rank-and-file members of functional agencies expected to resolve political conflict, at least if they hailed from democratic states.

If one accepts this, it leads to the third question posed above, the question of whether functional agencies could operate democratically in practice. That question has two related parts. First, could individual functional agencies sustain democracy? If so, second, could people democratically participate in many different functional polities simultaneously?

Could functional agencies be democratic?

Modern democracy has evolved largely within the borders of the territorially bounded, ‘multifunctional’ state. That makes it difficult to contemplate how democracy might take shape in function-specific, non-territorial agencies. A useful starting point is to ask what kind of dispositions members of a polity must share to make democracy possible. In broad and abstract terms, the answer is relatively uncontroversial. Democracy presupposes at minimum that most members of the polity believe (1) that they confront issues that affect them collectively, (2) that they are united by an overarching common good that transcends their immediate individual self-interest (even if they might disagree on what that common good entails) and (3) that democracy should be adopted as the procedural norm when responding to shared challenges and seeking to further the common good. All three dispositions must be held individually as well as intersubjectively: members of the democratic polity must know that other members share their commitment to democracy and to an overarching common good. Echoing such an interpretation, some theorists classify democratic dispositions as a form of ‘social commitments’ defined as trusting, empathetic, ‘non-instrumental and infused with emotion or affect’ (Lawler et al., 2009: 3). Democratic deliberation and decision-making therefore have more exacting prerequisites than bargaining and negotiation among strategically interacting and self-interest-maximizing individuals (Axtmann, 2006; Honohan, 2001).

By many accounts, the democracy-enabling dispositions just described can form a self-reinforcing relationship with the practice of democracy. Partaking in democratic deliberation, witnessing others abide by democratic decisions and observing democratic accountability mechanisms at work strengthens citizens’ commitment to democracy while reinforcing their trust in the democratic dispositions of others. Many theorists believe that such a mutually reinforcing relationship can operate in many different settings. Jürgen Habermas (1991), for example, famously argues that it could take hold in a supranational EU polity as in his view it requires neither shared national affiliations nor cultural homogeneity. At the same time, according to Habermas and many other theorists, the commitments that democracy requires as outlined above can only emerge if two main conditions are met. First, members of the polity must share an overarching communal identity. Second, their democratic community must be particularistic, that is, socially bounded and demarcated from others. ⁵

The first assumption builds on an influential conception according to which people acquire communal identities by internalizing the relevant social categories (Irish, Canadian, etc.) and as a result come to feel and act differently towards fellow ingroup members compared with how they feel and act towards outgroups. Their feelings towards the ingroup become infused with solidarity and trust and they evaluate collective choices partially with reference to a perceived common good rather than just their individual welfare. Both dispositions benefit democracy in the ways discussed. For this reason, communal identifications and democratic commitments are often held to be inexorably linked while democracy without communal underpinnings is deemed inherently fragile (Axtmann, 2006; Etzioni, 2007; Hogg and Abrams, 1988: Chapter 2; for a critique see Theiler, 2012).

The second claim (that a democracy-sustaining community needs to be socially demarcated) has two broad rationales. They correspond to the Janus-faced role of social boundaries as enabling both inclusion and exclusion. On one hand, boundaries serve to define the demos, that is, those who can participate in and are bound by democratic decisions. On the other hand, boundaries signify demarcation from outsiders, which some theorists deem a social-psychological prerequisite for democracy. In linking collective self-differentiation to democracy they draw on various theoretical traditions that emphasize the role of outgroups in the constitution of social identities (Barth, 1969; Hogg and Abrams, 1988). Briefly summarized, their chain of reasoning broadly goes as follows:

• A. Democracy requires social commitments and trust and thus a communal identity to engender these (as per rationale outlined above).

• B. Like individual identities, communal identities are relational in that they are partially defined by their difference from others – every ‘we’ needs a ‘them’. Without a shared awareness of difference from outgroups, a shared ingroup identity cannot exist.

• C. Ingroup identities thus are sustained in part through collective ‘boundary work’, that is, through the affirmation of (actual or imagined) differences from others who are not part of the democratic ingroup.

• D. Consequently, a democratic community’s members must collectively self-demarcate from outsiders, thereby engendering the shared social commitments that make democracy possible.

Different theorists develop this argument in different ways, sometimes relying on additional claims such as the alleged need for salient ‘others’ as a source of recognition (for a critical overview see Abizadeh, 2005). Moreover, they disagree on whether communal self-differentiation necessarily has to be antagonistic (Brewer, 2001). Those differences not withstanding, a conception of democracy as by necessity particularistic and other-dependent is widespread and influential. Some historical sociologists draw on it when linking modern nationalism as an internal identity-building and external boundary-affirming phenomenon to the expansion of democratic practices, such as a widening of the democratic franchise (see Cederman, 2001). Other theorists use the same logic to rule out cosmopolitan democracy as a global polity could not demarcate itself from any other human community not already subsumed by it (Habermas, 2001: 107; but see Abizadeh, 2005 and, more broadly, Koenig-Archibugi, 2011 for critiques).

In sum, many theorists believe that democracy is inexorably communal and other-dependent. Therefore, they deem the socially demarcated identity community the ‘logical space for democratically united citizens’ (Habermas, 2001: 107). If true, such a conception would seem to rule out democracy in Mitrany’s functional agencies. First, those agencies would unite socially, culturally and geographically disparate individuals around nothing more than particular shared needs that might be quite peripheral to their everyday lives. Their substantive ‘thinness’ would make it hard for functional polities to engender ‘thick’, multidimensional and absorbing communal identifications comparable to those often associated with modern nationalism and other collective identity projects such as ethnicity and institutionalized religion. Second, functional polities would find it difficult to solidify their internal democratic cohesion through ongoing external self-demarcation from – invariably – other functional polities. Such ‘boundary work’ would be inhibited by the socially and culturally diverse, strongly overlapping and fluid membership of functional polities. Individuals would belong to many different functional agencies and move in and out of them in line with their evolving needs. Moreover, since the overall functional system would depend on cooperation among the various functional agencies, their scope for mutual self-differentiation would be curtailed still further. All these factors would distinguish a functional from the existing states-based international order. A state-centric international system in which power is ‘bundled’ and legitimized within largely autonomous sovereign units with little overlap and fluctuation in membership can tolerate or might even depend on these units solidifying themselves in contrast (if not in outright opposition) to one another (Mercer, 1995). By contrast, a functional international system which disperses power among functional agencies whose shape and membership are continually shifting and that are interdependent, overlapping, interlocking and cooperating would likely experience such oppositional identity-building as antithetical and, if attempted, might easily be rendered dysfunctional by it.

For those reasons, a communal identity-dependent and other-contingent conception of democracy rules out democracy in functional agencies. The question is thus whether such a conception exhausts the range of viable democratic possibilities. In what follows, I argue that it is overly narrow and restrictive, starting with the assumption that democracy requires a communal identity.

Democratic communities and democratic associations

Though not arguing specifically with reference to international functionalism, several scholars in recent years have contemplated how democracy might (and to some extent already does) flourish in conditions less demanding than those just outlined. Their arguments diverge, but in broad terms they suggest that democracy and other trust-dependent social relationships reside on a spectrum. One end of that spectrum is marked by a widely internalized and socially demarcated ‘communal we’ which can cultivate democratic commitments in the ways just discussed. The other end of the spectrum features a form of ‘associative we’. It is defined by individuals believing not that they ‘belong together’ so much as that they ‘belong to’ a project or endeavour to which they are jointly committed (Mason, 2000: 127 ff., emphasis added). What sustains such ‘associative wes’ is a belief among people that they share particular interests, needs or challenges and that they should address those cooperatively. As Iseult Honohan (2001) suggests, even inside states such metaphors of association – or, in her terminology, of citizens as ‘colleagues’ – tend to capture relations between members of the polity better than more ambitious analogies (e.g. those that depict national polities as imaginary ‘families’ writ large). Nonetheless, even such relatively limited and unidimensional associations can sustain trust and commitment-dependent practices, including democracy. Like colleagues, members of a democratic polity must be committed to their common democratic ‘project’. Their relations must go beyond purely instrumental interaction among mutually indifferent strangers. But unlike members of more closely-knit communities, the limited scope of the ‘project’ on which they cooperate limits the necessary depth of their collective ‘we’ and of their commitments towards one another. Such citizens form, if you will, a limited ‘we’ that straddles the conceptual boundary between Gemeinschaft and Gesellschaft.

Conceptualizing citizens in democracies as metaphorical ‘associates’ or ‘colleagues’, then, implies a degree of mutual trust, solidarity and shared commitments. But it also implies that such sentiments may be quite limited and conditional. Empirical evidence supports such an interpretation. Intriguingly, for example, the share of citizens willing to ‘fight for their country’ tends to be smaller in democracies than in other political systems (e.g. 83% in the Middle East and North Africa (MENA) region and 71% in China but 44% in the United States, 25% in Western Europe and 11% in Japan (WIN/Gallup International, 2015)). These figures are open to interpretation. Nonetheless, they broadly suggest that even (and, it would seem, especially) in consolidated democracies most citizens do not hold beliefs such as that they are part of a ‘national family’ deserving of unconditional commitment and self-sacrifice. To stay with the analogy, fighting and sacrificing oneself for one’s family or close friends is one thing but doing the same for one’s ‘colleagues’ is quite another. These findings also imply, that the limited nature of people’s commitments to their national polities does not impede the democratic functioning of those polities. They help dispel overly demanding and idealized notions of democratic underpinnings which in reality even many consolidated democracies do not live up to.

Focussing beyond the state, ‘downgrading’ the minimum requirements for democracy in this way helps explain why many existing associations unite individuals around shared needs or interests and are democratically governed without representing ‘thick’ communities in the sense discussed earlier. Examples include many professional, academic and scientific associations, cooperatives, nongovernmental organizations and other types of voluntary associations. Even as of today, many of these associations are transnational and they are not territorially demarcated. They unite people who may otherwise hold little in common around a particular cause, interest, need or shared predicament. Members of the International Association of Mathematical Physics, for example, are scattered all over the world. They share an interest in the study of mathematical physics and often probably little else; their ties are relatively limited and unidimensional. Nonetheless, these ties are strong enough to sustain various forms of collegial interaction and cooperation among members of the learned society. These include presenting papers and acting as discussants at conferences, disseminating information about one another’s activities and paying membership dues. These ties and commitments also suffice to sustain the association’s democratic governance. Outvoted members and defeated candidates for executive positions generally accept democratic outcomes as legitimate. Where such associations remain limited in scope and purpose, their democratic functioning is not impeded by the limited, ‘collegial’ relationships among their members.

But do not even such limited associative or ‘collegial’ formations need outgroups against which they can demarcate, define and constitute themselves just as many national polities have constituted themselves in opposition to one another? Here, it is useful to return briefly to the earlier distinction between (1) boundaries as demarcating an internal realm of collective (democratic) agency and (2) boundaries as enablers of collective self-differentiation from external others. Boundaries in the first sense exist in every associative context, defined by the association’s activities, functions and purpose. For example, the International Association of Mathematical Physics concerns itself with mathematical physics while political science lies outside its boundary. Yet this does not mean that such associations inevitably maintain their identities through self-differentiation across those boundaries. As Arash Abizadeh (2005) argues in an important essay, conceiving of political identities and other forms of social solidarity as inexorably other-dependent as many theorists do risks conflating definitional and causal arguments. In a purely definitional sense, any collectivity that does not encompass all of humanity faces people who do not belong to it and with whom it shares boundaries in the first sense discussed above. For example, since not all humans on earth are political scientists, the term ‘we political scientists’ excludes some people (those who are not political scientists). Yet this does not necessarily mean that these outsiders are causally significant in bringing the collective category ‘political scientist’ into existence and for its members to identify with it. While every collectivity that does not comprise all of humanity faces ‘others’, these may not be salient ‘others’ against which the collectivity constitutes itself and strengthens its ingroup identity. What constitutes, say, political scientists, industrial workers, lone parents, public transport users, farmers, nurses and mathematical physicists as potentially meaningful collective categories able to sustain various forms of collective action and organization are in the first instance their members’ shared needs, interests, preferences and predicaments, not the existence of outsiders from whom they could collectively self-differentiate. As argued, in some cases collective self-differentiation does indeed occur and where it occurs it may strengthen ingroup identities (e.g. in cases of oppositional state and nation-building). However, this still does not show that differentiation from outgroups is the only way a polity can sustain democracy-enabling commitments, contrary to what several strands of democratic political theory suggest. As in the examples just discussed, a democratic ‘we’ can instead be sustained as it were from within, by people sharing particular needs and interests and a commitment to address those democratically. Some collective identities (such as certain types of nationalism) may depend on mutual ‘othering’ and self-differentiation, but democracy as a form of political deliberation and decision-making does not.

Challenges to democracy in functional agencies

The argument so far has been that there are no a priori obstacles to democracy in associations that lack ‘thick’ communal identities and ingroup-constituting outgroups. In principle, this also applies to Mitrany’s functional agencies. Even so, functional agencies would face challenges that many other types of democratic associations do not encounter to the same degree. Many of those challenges would emanate from their transnational character coupled with their non-voluntary membership and lack of easy exit options. For example, a functional agency might impose obligations on its members which some of them perceived to conflict with their respective national loyalties and obligations (or vice versa) and this could weaken their commitments to the functional agency in question. ⁶ Though not arguing with respect to democracy, Mitrany implies that functional agencies would tend to prevail over national polities when loyalties came into conflict. As argued, he sees people essentially as rational welfare maximizers who would support functional agencies based on their ability to satisfy needs more effectively than the state. Besides, he assumes that the functionally restricted scope of functional agencies would limit such conflicts in the first place. Because these agencies would not be of the same type as states (i.e. they would not be political-territorial units in their own right) the two would be more complementary to each other and generate fewer tensions compared to supranational integration projects that juxtapose different political-territorial units. ⁷

Regardless of whether one accepts Mitrany’s optimism in this respect, one factor to bear in mind is that even where national and functional loyalties did come into conflict most people would not find such tensions unusual. Rather, loyalties can conflict in any setting in which people participate in different institutions or associations, be they transnational, supranational or purely domestic. For example, some citizens in secular states may experience tensions between their obligations as citizens and their religious beliefs. Even though such conflicts may touch on deeply held moral convictions they typically do not make people renounce their allegiance to their religious communities or national polities (or both). Especially in culturally divided domestic systems, institutions often help prevent torn loyalties from becoming politically divisive by employing a measure of restraint, ‘constructive ambiguity’ and selective opt-out opportunities regarding policies that conflict with some members’ commitments to other institutions or with their core beliefs (Theiler, 1999). Such practices (which may or may not be formalized) enable individuals to deviate from some of the institution’s prescribed norms and obligations without renouncing their loyalty to the institution per se. By maintaining a degree of pragmatism and flexibility similar to what is being practised in many culturally divided domestic settings, functional agencies could contain potential tensions between functional and national attachments as well as between obligations arising from people’s membership in different functional agencies.

A further challenge to democracy in functional agencies would be their tendency to turn parts of their electorates into permanent minorities. This is best illustrated through a comparison with democratic domestic systems. In domestic democracies, voters decide through elections or referendums on issues across the entire range of policy areas – health, education, transport and so on. Consequently, for most voters, being a regular democratic ‘loser’ on some issues is counterbalanced by ‘winning’ on other issues; almost every voter is part of a winning majority some of the time. By contrast, in functional polities all voting would be on issues in the same relatively narrow (functionally defined) area – only transport, only health, and so on. Assuming that individuals tend to vote in similar ways on similar issues over time, voters in functional polities might find themselves regularly outvoted without being able to counterbalance these ‘losses’ with ‘wins’ on other issues within the same polity. Continuously ‘losing’ in this way might alienate them from the functional democratic process. If voice is ineffective and exit is blocked, loyalties risk weakening. Nonetheless, even people who were continuously outvoted in a given functional agency would very likely not be consistent ‘losers’ system-wide as those ‘losses’ would be counterbalanced by ‘wins’ in other agencies to which they simultaneously belonged. As a functional system developed and people joined a growing number of functional agencies their ‘wins’ and ‘losses’ would become more likely to average out over time and they would find it easier to develop a cumulative sense of their political efficacy.

Apart from these more general questions, a given functional agency’s ability to sustain democracy would also in part depend on how strongly its members experienced shared needs and expected them to be addressed democratically. As noted, democratic expectations would likely be strongest among those individuals who hailed from domestic democracies and had internalized democracy as a standard of political legitimacy. A further factor would be how well an agency’s democratic institutions were tailored to the specific functional context concerned. In many functional agencies, democracy might work best by combining traditional representative elements in the form of elected assemblies holding executives to account (and perhaps popular referendums) with more deliberative components. These could reflect the functional issue area at stake and the agency’s demographic and geographic characteristics. Technologies enabling virtual deliberation (e.g. in ‘electronic town halls’) could play an important role, especially in those functional polities whose members were dispersed over wide geographical areas.

So far, the question of what democracy in functional agencies could look like in practice has received almost no attention. This reflects widespread doubts that international functionalism is even compatible with any kind of democracy in the first place. My aim in this and the previous section has been to address that second, more fundamental issue. It is possible to conceptualize democracy in a functional setting where the collective democratic ‘we’ is constituted by an awareness of shared needs and a commitment to meet those needs democratically. Although Mitrany envisioned functional agencies to operate without democracy, nothing in their nature would make them inherently unsuitable for it.

Could people democratically participate in multiple functional polities?

The previous section has argued that functional agencies could sustain democracy in principle. It did not examine whether people could democratically participate in several functional agencies simultaneously. This is an important question, since in a functional international system every individual would belong to many different functional agencies. To define more clearly what this would involve, a comparison between a potential functional international system and existing multilevel domestic democracies is useful. In multilevel domestic systems, citizens belong to an overarching (e.g. federal) democratic polity and to various subordinate (e.g. regional and local) polities. These polities tend to possess relatively clearly delineated competences. Moreover, the different polities and their associated institutions are nested and citizens’ political identifications are – typically – concentric. Both conditions mitigate potential tensions and loyalty clashes. For example, Switzerland’s federal parliament may pass a law which a voter in the Canton of Zurich opposes as she perceives it detrimental to Zurich’s interests. Despite this, she likely considers the law legitimate for two main reasons. First, she respects the constitutional division of powers in Switzerland. Second, she identifies as a member not only of the Zurich polity but also of the Swiss polity to which she applies the standard that parliamentary majority decisions are democratically legitimate. Where polities, institutions and identities are nested in this way, divided loyalties are in a sense built into the system and their potentially divisive consequences are absorbed by it. In a functional international system, too, every individual would belong to many polities of various scopes, sizes and compositions. However, these would be far more numerous than the two or three polities inhabited by each voter in typical multilevel domestic democracies and they would not be hierarchically nested. Nor, by definition, could political identifications be concentric since a functional international system would lack a centre. The question therefore is how the fragmented, non-nested and decentred nature of a functional international system would affect people’s democratic participation.

This question can be subdivided into two component parts that pertain to the cognitive and affective dimensions of political participation respectively. First, democracy is cognitively demanding for voters. It requires that they keep abreast of issues and arguments, be able to attribute positions to parties and candidates, broadly grasp what various institutional actors do and so on. The greater the number of democratic polities in which an individual participates, the more complex and thus cognitively challenging participation tends to become and the greater the risk of ‘polity overload’.

Whether voters in a functional international system could manage its complexity would partially depend on how many functional polities they would join. Ten? One hundred? One thousand? Mitrany offers few hints, consistent with his belief that functional agencies would evolve in line with constantly changing needs. Even so, the earlier argument is again relevant here. A functional international system would have to satisfy not just immediate human needs (e.g. health, nutrition and environmental sustainability) but also needs linked to the operation of the relevant needs-satisfying institutions themselves. As was argued, this would likely include the need for these institutions to be democratic to enjoy popular legitimacy. Likewise, the overall system would need to remain sufficiently simple and transparent to enable (and motivate) voters to participate. One way to help achieve this would be by limiting the proliferation of functional polities. Though not arguing from the perspective of democracy, Mitrany (1966 [1943]himself suggests a measure of horizontal integration and consolidation with agencies in related policy areas being subsumed under a joint coordinating body (pp. 73–75). If those coordinating bodies rather than each functional agency separately became subject to democratic control (or at least if each functional agency received a relatively broad policy remit), the proliferation of functional polities could be curtailed and their members would be less likely to suffer cognitive ‘polity overload’. In that way, a functional international system could reign in institutional complexity without abandoning basic principles of functional organization.

In addition to its cognitive requirements, democracy also demands that voters form affective commitments and trust towards one another and their polity at large. Previous sections argued that people could develop such allegiances with respect to individual functional agencies. They also suggested that functional agencies could reduce the risk of those allegiances coming into politically divisive conflicts with other loyalties their members might hold. What has not yet been examined is whether political loyalties could become sufficiently ‘fractionated’ (Pentland, 1973: 85) for people to remain politically committed to many different functional agencies simultaneously. From a social-psychological perspective, no fundamental obstacle would seem to stand in the way of this. Already as is, people typically belong to various social collectivities to which they are affectively committed to various degrees. These range from deeply internalized and multidimensional communities (such as close-knit families and circles of friends) to much looser and unidimensional associations discussed earlier (e.g. professional bodies) with many intermediate forms. Informed in part by the highly influential social identity theory and its various offshoots (Hogg and Abrams, 1988), many social psychologists posit that group loyalties grow out of learning and socialization processes that are continuous and expansive. Joining additional groups does not necessarily weaken attachments to those groups to which one already belongs. Several researchers even suggest a positive relationship between various social attachments with existing group loyalties serving as ‘springboards’ for people to develop commitments to additional groups. In the EU, for example, citizens who identify strongly with their nations on average also feel more attached to the EU (Duchesne and Frognier, 1995). If people’s social attachments are neither zero-sum nor ‘rationed’, their membership in various functional polities might simply expand their repertoire of social affiliations without overburdening their ability to commit.

But what if their membership in multiple functional polities became bound up with conflicting preferences in voters? As a member of an environmental protection functional agency, for example, someone’s sense of the common good might make her champion cuts in carbon emissions whereas as a member of a tourism functional agency she might oppose restrictions on air travel. Such tensions would likely be inevitable but they, too, would not be unusual. In any political system, people typically hold some mutually conflicting beliefs and preferences partially conditioned by their various social roles and affiliations. Even so, how people would express these preferences and their political consequences would be different in a fragmented functional compared to a unified domestic system. Inside domestic democracies, citizens tend to vote for those political candidates and parties they believe to align most closely with their own preferences on a combination of issues. This forces voters to compromise and may strain their value systems but it is inevitable since no party or candidate will fully match a voter’s preferences on every single issue (Krosnick, 2002). By contrast, in a functional international system voters could maximize their preferences in every functional polity of which they were a member. As in the example above, a voter might want to fight climate change and promote tourism and thus vote to restrict air travel in the environmental protection functional agency while simultaneously voting for candidates who promise to protect air travel in the tourism functional agency. Voters maximizing their mutually conflicting preferences in this way could lead to different functional agencies adopting mutually incompatible policies without there being a clear mechanism of reconciling them. It is not obvious who would ‘meta-aggregate’ conflicting preferences across functional sectors and agencies, just as it remains unclear who would allocate resources between (as opposed to within) different functional agencies. International functionalists have done little to contemplate this problem while sceptics might posit that the alternative is between two unpalatable outcomes. A first option would be purely intergovernmental methods of system-coordination with national governments jointly harmonizing the work of functional agencies. Such an arrangement would veer towards minimalist conceptions of international functionalism as little more than intergovernmental cooperation discussed above and it would carry dubious democratic legitimacy (Hammarlund, 2005: 128). Alternatively, reconciling competing preferences across functional polities could involve some type of overarching meta-coordinating body. Yet whether democratic or not, such an overarching institution might end up looking much like a supranational state – or at least an ‘orderer’ of some kind – and thus undermine central functionalist tenets from the other direction.

Transcending this ‘intergovernmentalism versus quasi-state’ dilemma would likely involve horizontal cooperation between functional agencies, perhaps accompanied by some measure of overarching institutionalization and democratic oversight. In places, Mitrany (1966 [1943]: 73–76) touches on such ideas for system-coordination (though again minus a concern for democracy) but to pursue this further would go beyond the scope of this article. What this section has tried to argue is that a functional international system would present individuals with various challenges due to its fragmentation and lack of political centre. However, nothing in principle would prevent people from dividing their democratic commitments among the various functional polities to which they would simultaneously belong.

Conclusion

Mitrany’s international functionalism is compatible with democracy both in principle and in practice. Its core proposition is that function-specific agencies could meet some needs better than the state in a context of continuously differentiating human needs and growing social interdependence. That claim builds on the notion that institutions work best if their form and scope of authority follow their function. Contrary to what Mitrany suggests, it does not inherently privilege technocratic over democratic decision-making within functional agencies and it does not imply that those agencies would be free from political conflict. In other words, Mitrany’s failure to provide for democracy in functional agencies is his add-on to the underlying functionalist logic, not an intrinsic part of that logic. Regarding the practical viability of democracy inside functional agencies, it is probably true that their members would not develop ‘thick’ communal identifications. Yet several strands of democratic theory supported by empirical precedent suggest that even people who do not share an overarching communal identity can develop an awareness of shared needs and a commitment to meet those needs democratically.

The article was not a general manifesto for international functionalism. It did not show that Mitrany’s approach is empirically possible or desirable in other respects and it leaves many questions unanswered. For example, while it argued that functional polities could be democratic, it is not clear how democratic outcomes could be reconciled between different functional agencies and sectors and how resources could be allocated system-wide. Those are among the questions that the (still relatively few) scholars who are presently rediscovering the functionalist tradition in International Relations will need to address.

Nonetheless, if my argument is valid it removes an important obstacle to taking Mitrany seriously. His international functionalism looks beyond the state as the central building block of international order and beyond the supranational centralization of power to its diffusion. International functionalism’s appeal has suffered in part from a belief that it is inherently technocratic and incompatible with democracy both conceptually and in practice. By arguing against such a conception, the article has sought to free international functionalism from its technocratic stigma and highlight its potential as a democratically viable alternative to both state-centric and supranational models of international order.