Despite their disagreements, participants in this debate remain committed to the view that arguments about what liberalism is, are arguments about the content of liberal nonnative theory. But it is not clear that we should feel bound by this account of liberalism. Indeed, one could argue that this view of liberalism is anachronistic in the sense that thinkers in the classical cannon of liberal thought were not simply, or even largely, engaged in doing normative political theory.¹ Given the dominance of a particular kind of analytical normative theory within the contemporary academic study of political theory, it is easy to reinterpret the arguments of these classical thinkers as normative theory. But the historical point of liberalism has been precisely its transformative or even revolutionary character, and many liberal thinkers were engaged substantially in intensely political arguments about social transformation. This is not to say that normative political theory is unimportant to liberalism; rather it is to say that we should recognise how much of this theorizing, at least historically, has been animated by broader political objectives.

One attempt to get beyond the idea of liberalism as normative political theory has been put forward by Barry Hindess (Hindess 2002a, 2002b, 2004). He argues that liberalism can be understood as a ‘project of government’. By this he means that liberal thought has always been concerned with the problem of governing. Hindess uses the terms ‘government’ and ‘govern’ here in ways derived from the work of Michel Foucault (Foucault 2001x, 2001b). Foucault drew attention to the emergence of a new concern with governing populations that emerged in eighteenth-century social thought. This concern arose, so Foucault suggested, as a result of the discovery of the ‘social’ and the ‘market’ as autonomous realms with their own internal dynamics. The problem of government was then the problem of how to govern (regulate, order and control) without disturbing the dynamics ofthese realms – particularly that ofthe market. What distinguished liberalism was its commitment to governing as far as possible through the promotion of certain kinds of free activity and the cultivation among the governed of suitable habits of self-regulation (Hindess 2004). Hindess expands and develops Foucault’s account of liberalism as a ‘rationality of government’ in two important ways. First, Hindess is concerned to balance the books as it were, by drawing attention to the myriad ways in which liberal thinkers have advocated the use of coercion (rather than just regulated freedom) to govern. Second, and very importantly, Hinders locates the liberal desire to govern within the distinctions liberal thinkers have always made between the ‘civilized’ and ‘uncivilized’ or ‘developed’ and ‘less developed’ (Hindess 2002b). Thus for Hindess, a central part of liberalism has been its international or cosmopolitan vision, but this vision has been guided by a set of concepts that locate people and peoples on a spectrum – from uncivilized to civilized, less-developed to developed – that justifies the use of all kinds of coercive governmental techniques on those who cannot govern themselves.

Hindess puts forward two justifications for developing this account of liberalism (Hindess 2004). First, he argues that seeing liberalism as a project of government illuminates the actual practices of governing evident within liberal states (and within the attempts of these states to govern other places) (see also A. Barry et al. 1996). Second, he argues that seeing liberalism in this way provides a fuller and more powerful account of the work of central figures in the liberal tradition (Hindess 2004). There is a great deal to be gleaned from Hindess’s account of liberalism. As this book proceeds certain similarities with his account of liberalism will become clear, especially the concern with the techniques and strategies of governing within liberal thought, and the stress on the cosmopolitan nature of liberal thought and practice. This book also follows Hindess’s justifications for conceiving of liberalism as more than simply a body of normative political theory. It argues that the focus on liberalism as a project of social transformation helps illuminate the actual practices of the World Bank, and it argues that this conceptualization of liberalism makes better sense of some of the key thinkers within the liberal tradition, as well as helping to illuminate some of the key tensions and ambiguities within liberal thought. In other words, thinking of liberalism as a project of social transformation helps connect political thought with political practice in illuminating ways, and helps make better sense of liberalism than the view that liberalism is a body of normative theory. Where we part company from Hindess is in the stress placed on the idea of social transformation rather than government. To be sure these are not contradictory views; the exercise of government is an essential part of the project of social transformation. Yet we place more emphasis on the ‘kinetic’ elements of liberalism; its restless and relentless desire to remake the world in its own image. It is this that ultimately underpins the liberal project. To pursue this argument we remain for the time being at the level of theory. But we treat liberal theory not simply as normative theory, but as an extended reflection on the desirability and possibility of social transformation.

The investigation proceeds in three stages. It is not too controversial to say that liberal theorizing has been overwhelmingly concerned with four spheres – the state, the economy, ‘civil society’ and the individual – and with the appropriate relationship between these spheres. Initially, we look at liberal theorizing about the state, the economy, and civil society. Obviously there is no space to undertake a detailed exposition of liberal thought in all these spheres; instead we identify some of the central tensions and ambiguities that characterize liberal thought about these spheres. We then move to the heart of liberal theorizing – its stress on the individual as the source of justification for liberal ends and as the central component of liberal theoretical strategies. But again here the concern is to show the typical ambivalence and ambiguity that characterizes liberal thought about persons. On the one hand, and in their more philosophical mode, liberal thinkers have tended to want to ground liberal thought in traits that are deemed ‘natural’ in one way or another: autonomy, the ability to reason, or ‘the desire to truck, barter and exchange one thing for another’ as Adam Smith put it. On the other hand, and in their more sociological mode, many of the same thinkers were profoundly impressed by how malleable people were – by how influenced they actually were by religion, custom and tradition; so much so that many of these supposedly ‘natural’ traits were absent from social life. Finally, we focus on those often neglected aspects of liberal thought that more obviously identify it as a political project. In particular, liberal thinkers advocated a series of techniques designed to change the way people thought and acted. These techniques rely on the fact that people are influenced by things like religion, custom and tradition. Thus, the promise of liberalism as liberation from various ‘oppressive’ institutions and practices is revealed to be insertion into new forms of discipline.

There is an extensive debate over exactly what the basis of the argument for a neutral state is, and liberals are characteristically ambivalent on this point (Galston 1982; Kymlicka 1989). On one hand liberals have appealed simply to the ‘facts’ of pluralism to demonstrate the desirability of a neutral state. As Locke says, ‘men in this world prefer different things and pursue happiness by contrary courses’ (Locke 1976: 123). In a related fashion, Rawls has suggested that neutrality is to be understood as apolitical not a metaphysical principle; the liberal state simply responds to the already existing pluralism in modern society (Rawls 1985, 1993a). On the other hand, liberals have very often relied on philosophical arguments, rather than political ones, to justify the neutral state, and in so doing have advanced conceptions of the good (Galston 1982; Sandel 1982). Hobbes and Locke both endorsed what would come to be called the ‘bourgeois’ goods of peace, comfort and security (Salkever 1990: 170). In doing so they endorsed these values as particularly appropriate for individuals. There is a good case for arguing that Rawls also has a particular conception of the good for persons at the heart of his theory (Sandel 1982). As Brian Barry has argued, ‘there is no way in which nonliberalscan be sold on the principle of neutrality without first injecting a large dose of liberalism into their outlook’ (B. Barry 1990: 54). That is, arguments for the neutrality of the liberal state actually rest on substantive conceptions of the desirable way of life; or to put it another way, the neutrality principle applies between ways of life that are the expression of the choices of individuals already committed to the goods of liberalism (Galston 1982: 625).

This ambiguity about the grounding and extent of the neutrality principle points in the direction of the other vision of the state in liberal thought – that of the strong state. For while it might be right that all liberals think there should be domains of social life over which the state has no legitimate authority, they also want to see these domains created and sustained, and the only agency capable of doing that is the state. The strong state must to a certain extent be disengaged from social interests and certainly not overwhelmed by them; it must in other words, be ‘autonomous’ from society, at least up to a point, because it is only by being autonomous that the state is not prone to capture by social groups who would threaten the achievement of liberal ends. And this is where we should locate the anxiety that many liberal thinkers have had about the form and extent of democratic arrangements. This is especially clear, for example, in the work of John Stuart Mill (Mill 1972). This suggests that there is within liberalism a vision of the state as capable of imposing and maintaining a certain kind of social order – a liberal order. On this view it is quite impossible for the liberal state to be neutral and indifferent to values; rather, it must actively interfere in what people believe and how they live, even to the extent of inculcating certain kinds of values and dispositions. As Rawls put it, ‘in agreeing to principles of right the parties in the original position consent to the arrangements necessary to make these principles affective in their conduct’ (Rawls 1971: 515). Such elaborate processes of transformation of both institutions and practices require not a weak state, but rather a state constituted in the form of a bureaucratic apparatus with all the capacity for social surveillance and social control which that makes possible.

The tension between the idea of a weak state and the idea of a strong state is explicable in terms of the liberalism as a project of social transformation. To put it bluntly a strong state is necessary to construct and defend liberal institutions and practices, but once these are established and secure the state can afford to be ‘neutral’. All of this might seem some way from the World Bank, but as we shall see, it is not. For these ambiguities about the state – weak but strong, neutral but partial, accountable but not captured – are exactly replicated in the way the World Bank thinks about the role of the state in development. And the central role that the state is understood to have in the liberal project is again exactly replicated in the World Bank’s account of good governance.

In terms of the pursuit of private economic interests, the economic realm is conceived as a realm of individual freedom. It provides a space for individuals to pursue and satisfy their private interests. This connects with liberal thought both in terms if its individualism and its commitment to forms of ‘neutrality’: individuals are the best judge of their own social welfare, and the market provides an institutional expression of the diversity of individual choices (Peacock and Rowley 1972). And it is because people are the best judges of their choices that the state should not intervene in the market to constrain those choices unnecessarily. The great historical ‘trick’ of liberal economic thought has been to combine this account of the economic realm with an account that stressed how the pursuit of private interests is congruent with, and can advance, the public good (Gauthier 1986). The first and most obvious way in which this is thought to happen is through the operation of the market mechanism. ‘The superiority of market society over its predecessors and rivals is manifest in i...