Social complexity and the emergent state

Abstract

Many political philosophers assume the state can coherently reform a society's legal system to realize just, society-wide distributive outcomes. Gerald Gaus invoked social complexity to highlight the limitations of this ambition. Complexity theory holds that interdependent social interaction in large-scale societies leads to unpredictable outcomes. For Gaus, complexity constrains what the state can accomplish. The state does not know how to reform the legal system to achieve ambitious distributive goals. However, Gaus did not model the state itself as a complex system. This is the contribution of this paper. I argue that the state is not a unitary agent that comprehensively oversees the legal system. Rather, the state is a network of interaction between myriad political agents and the legal system is an emergent outcome of this interaction—the result of human action but not an overall design.

Keywords

complexity the state Gerald Gaus justice legal centralism

“We need to cut off the King's head: in political theory that has still to be done.” Michel Foucault

Introduction

In one familiar view, the task of political philosophy is to identify ambitious principles of social justice to be realized through the reform of society's major institutions. This view assigns the task of reform to the state. As David Miller (2003: 6) puts it, when considering theories of social justice, political philosophers assume that:

… there is some agency capable of changing the institutional structure more or less the way our favored theory demands. It is no use setting out principles for reforming the basic structure if in fact we have no means to implement these reforms. The main agency here is obviously the state: theories of social justice propose legislative and policy changes that a well-intentioned state is supposed to introduce.

I will call this the standard view of political philosophy's task because it is held widely (e.g. Dworkin, 2002; James, 2005; Meckled-Garcia, 2008; Nagel, 2005; Rawls, 1996: chap. 7; Valentini, 2017). The standard view makes two critical feasibility assumptions. First, the state is a unitary authority that can coherently design and oversee a society's legal system. Second, the state knows how to design the legal system to realize demanding principles of distributive justice.

In various works, Gerald Gaus (2021, 2016, 1998) has invoked social complexity theory to criticize the standard view. A social system is complex if the autonomous behavior of its constituent actors is interdependent. This interdependent behavior gives rise to emergent outcomes that are very difficult to predict and irreducible to the constituent actors’ intentions. To paraphrase Adam Ferguson (1995): emergent outcomes result from human action but not design.

In Gaus’ telling, complexity arises principally in the economy and civil society. An economy is a complex system because the interdependent decision-making of myriad economic agents gives rise to emergent and unpredictable distributive outcomes. Gaus grants that the state has the unitary authority to design the legal system. However, on his score, the interdependency of economic behavior means that the unitary state does not know which design(s) of the legal system will realize demanding principles of distributive justice, such as Rawls’ justice as fairness or luck egalitarianism.

This paper advances one step further than Gaus by invoking complexity to question whether the state is a unitary authority. Complexity does not merely reveal constraints on what the state can accomplish but leads us to reconceive it as an entity. The state is a complex network of interaction between many political agents, including the various branches of national governments, many kinds of subnational governments (city mayors, provincial authorities, etc), unelected administrative agencies and politically engaged civil society associations. The legal system emerges through the activity of these myriad agents and its overall shape cannot be reduced to their individual intentions or a group-level corporate intention. In other words, the legal system itself results from human action but not the intentional design of a unitary authority. I call this the emergent conception of the state, in contrast to the unitary view prevalent in political philosophy (e.g. Lawford-Smith, 2019; Stilz, 2011), and make the case for why we should favor the emergent view.

I also outline the implications of the emergent view for the pursuit of social reform. While it acknowledges that particular state actors hold the authority to amend parts of the legal system, the emergent view rejects the notion of a singular, cohesive entity capable of systematically designing and managing the entire legal system. Still, I will argue that political complexity does not imply fatalism concerning the prospect of improving the legal system holistically.

It is well known that complex social systems can self-organize in the absence of a central governor if the constituent actors subscribe to the right sort of rules when making otherwise independent decisions. For example, rules of property and exchange allow the market economy to self-organize in the absence of an economic central planner. I will argue that constitutional rules perform an equivalent function in the emergent state. This implies that we can improve the legal system (as a whole) indirectly by reforming constitutional rules. However, the goal of such constitutional reform would be relatively modest. Our ambition would be to improve the self-organization of the political process rather than prescribe the precise nature of the legal system that emerges from an intricate and contingent web of political interactions.¹ Finally, I will conclude the paper by highlighting the implications of the emergent view for some wider debates within political philosophy.

Two feasibility assumptions of the standard view

I begin by explaining why the standard view is committed to the two feasibility assumptions I outlined just above. The standard view calls for outcome-oriented legal reform, where a reformer seeks to realize a specific social outcome via legal reform. Let us define legal rules as “[t]he patterns of interaction that govern and constrain the relationships of individuals” (North et al. 2009: 15) and posit that interaction under legal rules gives rise to social outcomes. To take a very simple example, a rule banning littering would hopefully lead to the social outcome of less littering. So, assuming ought-implies-can, a normative theory calling for a social outcome must meet two feasibility conditions:

There must be a duty bearing agent with (i) the authority to initiate or reform rules relevant to attaining the social outcome and (ii) who knows how to set those rules effectively to realise the desired social outcome. Call these the authority and realisation conditions respectively.

These two feasibility conditions are distinct. Suppose our goal is to reduce littering in a small town. Our duty-bearing agent would need the authority to initiate rules concerning littering and know how to design such rules effectively to reduce littering. Someone who knew how to frame an effective anti-littering rule but was not the relevant municipal official could not fulfil our goal. Vice-versa, neither could the relevant municipal official if they did not know how to formulate an effective anti-littering rule.

The standard view calls for outcome-oriented reform because it calls for legal reform to realize just distributive outcomes. However, it will prove helpful to my argument to stress the kind of social outcome the standard view seeks to realize. We can usefully distinguish between two kinds of such outcomes. A systemic outcome refers to the overall functioning of the social order. It is the product of how many components of the social order interact with one another. A discrete outcome refers to the functioning of a relatively isolatable part of the social order. Of course, this distinction refers to a continuum rather than a strict dichotomy. The littering rate in a small town is an example of a (very) discrete outcome. Examples of systemic outcomes include the society-wide distribution of income and wealth, anthropogenic climate change or an economy's overall industrial and sectoral composition.

According to our two-step feasibility model, the realization of a discrete outcome requires a duty-bearing agent with the authority to reform a comparatively limited array of institutional rules. In contrast, the realization of a systemic outcome requires a duty-bearing agent with the authority to reform a comparatively extensive array of rules precisely because the agent seeks to change an outcome resulting from the interaction of many components of the social order.

The standard view seeks a systemic outcome when it calls for a society-wide distribution of income, wealth, opportunities and so on. Take John Rawls’ theory of justice as a paradigm case. Rawlsian justice does not evaluate the discrete outcomes of particular institutions considered in isolation from one another, be these the labor market, the pension system or the central bank's monetary policies. Rather, Rawlsian justice evaluates how these and other major rules form an interconnected legal system—the “basic structure” in Rawls’ famous terminology—whose combined operation profoundly and differentially affects individual life chances (Rawls, 1996: 258, 1999: 6–7). The isolated operation of specific rules does not determine each person's expected, lifetime income and wealth. Instead, it is the combined effect of myriad institutional rules, including education provision, tax law and market regulation, among many others, that does so. Hence, Rawls’ difference principle aims to determine the society-wide distribution of income and wealth via coordinated reform of the overall legal system.²

Accordingly, defenders of the Rawlsian approach emphasize that social justice evaluates the legal system holistically. For example, Thomas Nagel (2005: 130) writes that the “legal framework” is evaluated by principles of social justice, “not [legal] act by act, but … the system as a whole.” Michael Blake (2001: 279) too says the task of political philosophy is to “justify the commands of [the] legal system as a whole.” Nor is the Rawlsian approach unique in this regard. Other principles of distributive justice (such as luck egalitarianism or the capabilities approach) also evaluate, and require the reform of, the legal system as a whole.

Let us now apply our feasibility model directly to the standard view:

There must be a duty-bearing agent with (i) the authority to design the legal system and (ii) who knows how to design the legal system to realise the systemic outcome of society-wide distributive justice.

It is worth reflecting on the broad scale of oversight that the authority condition requires in this context. I do not think this is sufficiently appreciated in the literature. The pursuit of a society-wide distribution of income and wealth (setting aside other goods) could require the coordinated oversight of rules as wide-ranging as national tax rates, local tax rates, education policy, regional housing policy, national infrastructure policy, immigration policy, competition policy, mental health provision, early year intervention policy, discrimination law, the regulation of product and capital markets and carbon policy. All these policies could interact in complex ways to affect the distribution of income and wealth and so all would need to fall under the authority of the duty-bearing agent.³

I will hereafter speak of the unitary authority condition to describe the overarching institutional authority that the duty-bearing agent would need to have on the standard view. As the David Miller quotation suggested, the state is seen as the obvious candidate for the duty-bearing agent. This would imply that the state is a cohesive agent with the authority to form and execute coordinated plans to design the overall legal system. In other words, the unitary authority condition would hold if the state met a key condition for the unified collective agency, namely, that it could form and act upon group-level intentions.⁴ I am here observing that the standard view is at least implicitly committed to modelling the state in this way. But several political philosophers do so explicitly (e.g. Lawford-Smith, 2019: passim; Pettit 2014: 98; Stilz, 2011: 195).

The realization condition is the second, additional requirement. The state not only needs to be an agent with unitary authority over the legal system. It needs to know how to configure the legal system so that interaction within the market and civil society leads to the desired, society-wide distributive outcome.⁵

Let me stress again the distinctiveness of these two feasibility conditions. One could grant the unitary authority condition but question whether the unitary state knows how to use its comprehensive authority to realize and maintain a society-wide distributive outcome. This, as we shall see, is Gerald Gaus’ position. Alternatively, one might grant that a unitary authority could attain a systemic distributive outcome if it did exist but deny that it does. In the global justice debate between statists and cosmopolitans, we can interpret statists as taking such a position concerning demanding principles of global distributive justice. Statists argue there is no agency with the authority to design an integrated, global legal system as would be necessary to attain distributive justice at that scale (Blake, 2001; Freeman, 2007: 444–45; Meckled-Garcia, 2008; Nagel, 2005). To my knowledge, statists do not argue that a global unitary authority would be unable to realize global distributive justice if it did exist.⁶ Nor, of course, do they question the ability of individual states to use their unitary domestic authority to realize distributive justice within their own societies.

The complexity account of this essay calls into question both feasibility assumptions of the standard view. I accept Gaus’ claim that an agent with the unitary authority to design the legal system would not know how to realize society-wide distributive justice. But I additionally argue that there is no agent with such authority. On the other hand, I concur with statists that no agency has the authority to design a unified global legal system. But I go one step further and claim no such authority exists in the domestic setting either.

Gaus’ complexity-based challenge to the realization condition

In this section, I outline Gaus’ complexity-based challenge to the standard view. As I mentioned, Gaus’ critique only challenges the realization condition because he models complexity as arising principally in the market and civil society. He does not consider whether complexity arises within the state and the political process. Before discussing Gaus’ critique explicitly, it will prove helpful to define a complex social system and explain how the market economy meets that definition. Like Gaus, my account of the complex market will draw significant inspiration from Friedrich Hayek. Later, I will argue that the state is also a complex system and consequently that the legal system is an emergent phenomenon rather than the design of a unitary authority.

Social complexity in the market

Scholars have used complexity to model a wide range of natural phenomena, from the global climate to the workings of the human brain. We are concerned with complex social systems. Such a system has five key properties (Gaus, 2021: secs. 17–19; Hayek, 1967a, 1967b; Lewis, 2012, 2016)

It is composed of agents whose decision-making is independent yet interdependent. Each agent decides for themselves how to behave, but their decisions affect the behavior of other agents.

This interdependent decision-making gives rise to emergent systemic outcomes that are not intended by the agents taken in isolation, the agents understood as collective actors, nor any central coordinator of the system. They result from human action but not an overall design.

These emergent outcomes are difficult to predict in any fine-grained way.

Complex systems are self-organizing rather than chaotic. Their emergent systemic outcomes are orderly and meet functional desiderata despite being unplanned in their totality because the relevant agents subscribe to procedural rules. We can understand procedural rules in contradistinction to outcome-oriented rules because their function is not to realize a particular social outcome. Their task, rather, is to promote the self-governance of a complex system.

We may be able to indirectly improve the systemic outcomes of a complex system by reforming the procedural rules. However, such reform would only make the emergence of systemic outcomes that meet functional desiderata more likely. It would not be outcome-oriented in the sense of intending to realize a fine-grained or specific social outcome.

Now, let us see how the market meets these five conditions.

The decision-making of market actors is inter- and independent. Each individual makes her own economic decisions while reacting to the decisions of others. Your employment, consumer and investment decisions influence my economic decisions which, in turn, influence the economic decisions of other actors, and so on. Many such iterative economic interactions give rise to emergent outcomes such as the society-wide distribution of income and wealth.⁷ We cannot trace such outcomes to the intentions of the many agents who are nonetheless causally responsible for them.

The interconnected behavior of market actors, where each decision is influenced by and, in turn, influences others, means that aggregate market outcomes are very difficult to predict in a fine-grained way. Even small changes in the behavior of some actors reverberate throughout the rest of the system. Consequently, our efforts to predict society-wide distributive outcomes will be subject to error inflation (Gaus, 2016: 80). Just a minor error in predicting the behavior of some actors will create yet larger errors in our predictions of the behavior of other actors further down the society-wide causal chain. This tendency for our errors to compound sternly constrains our predictive powers over complex systems.⁸

This is not to suggest we have no predictive leverage on complex social systems. Rather, our predictions are subject to very wide margins of error. We can reasonably forecast the range of systemic market outcomes, but more fine-grained predictions within that range will be beyond us (Hayek, 1967a: 10–14). An analogy with a natural complex system, the weather, can helpfully illustrate this point. It is very difficult to make accurate, fine-grained predictions of the weather just a few weeks into the future. Nonetheless, we know there is a limit on the range of possible weather outcomes. We can say with a very high degree of confidence that we will not see a hailstorm in Cairo or a tornado in London in the foreseeable future.

Society-wide distributive outcomes in the market can be functional in certain ways despite being emergent and unintended. Here I refer to the market's tendency to increase prosperity, discover and diffuse new technologies via creative destruction, incentivize individuals to change occupations given changing circumstances (albeit often painfully) and so on. This raises a social scientific puzzle, namely how decentralized market activity can lead to functional emergent outcomes, as opposed to anarchy of production, in the absence of deliberate central planning.⁹

In Friedrich Hayek's well-known account, the market system can self-organize so long as market actors adhere to common rules while making independent decisions in specific situations.¹⁰ The most important of these rules specify rights of property and exchange. I call these procedural rules to contrast them to outcome-oriented rules. Procedural rules do not direct market agents to promote a social goal but rather condition or constrain their otherwise independent decision-making. For instance, property rights assign jurisdictional rights over resources and rules of exchange specify how agents may modify their jurisdictional rights. These rules do not specify what kind of transactions agents should undertake nor how they should use the resources under their control. Such substantive decisions are left to the discretion of individuals and firms.

Procedural rules promote the self-organization of the market system in various ways. Property rights moderate the interdependencies of the economy when they do a reasonably good job of internalizing externalities. Consequently, many actions carried out within the confines of one's property have little to no impact on others and vice versa. This reduces (though by no means eliminates) the variance of macro-level economic outcomes, thereby promoting a more stable framework for interaction (Gaus, 2016: 198–201). Additionally, Adam Smith (2009: chaps. 1–3) famously observed that secure property rights and rules of exchange incentivize individuals to find specialist occupations in a functional division of labor without central direction.

Hayek identified two further ways in which rules of exchange and property further self-organization on the market. First, secure property rights enable the market to be a “discovery procedure” (Hayek, 2002). Entrepreneurs can experiment with putting new goods and services on the market or finding cheaper ways to offer existing goods and services. The successful experiments spread throughout the system via imitation without a central planner. Consider, for instance, how Apple's innovation of the smartphone changed the mobile phone industry. Second, rules of property and exchange generate market prices that aggregate and transmit otherwise dispersed information about consumer demand for, and scarcity of, various goods. These prices tacitly coordinate the behavior of myriad market actors by, inter alia, signaling to economic producers how they can best meet consumer demand and to consumers about the need to economize on certain goods (Hayek, 1945).

Finally, for this section, Hayek did not rule out a role for public policy in improving the systemic outcomes of the market.¹¹ Public policy can refine the procedural rules that underpin market self-organization by, for example, amending property laws to internalize externalities more effectively. However, consistent with our limited predictive powers over complex systems, such policies aim only to adjust the range of potential emergent outcomes by making more likely outcomes favoring prosperity or innovation. Unlike outcome-oriented reforms, they do not aim for any specific distributive outcome. That would depend on how the contingent process of interdependent market interaction unfolds.

Gaus’ focus on the limits of the unitary governor's knowledge

Gerald Gaus’ complexity-based critique of the standard view takes the unitary authority condition for granted. In his final work, The Open Society and its Complexities, he enquires into what kind of self-governance is possible in a complex system. Gaus (2021: 124) borrows Jennan Ismael's (2016: 19) definition of self-governing systems. These:

… are systems in which at least some organised activity is the result of a centralised process that involves the sharing of information and the formation of an overall plan and deliberate coordination of joint activity. In a self-governing system . . . at least some of the information distributed throughout the system is collected, synthesised, and used to fuel a decision procedure that plays a role in guiding the system's behavior. (Emphasis added)

Gaus’ framing of his inquiry accepts that a complex society has a unitary “governor”—Gaus nearly always uses the singular noun—who tries to guide the system's behavior via “the formation of an overall plan and deliberate coordination of joint activity.” Gaus appears, then, to grant that all formal legislation stems from a unitary governor. On his account, complexity and emergence arise in the manner citizens—qua participants in the market and civil society—respond to the governor's legislative designs. Hence, his critique of the standard view only questions whether the unitary state meets the realization feasibility condition.¹²

As we saw earlier, the realization condition requires the governor to create a legal system of outcome-oriented rules to attain a predetermined society-wide distribution of income, wealth, opportunities, and so on. Gaus argues that the governor does not know how to do this (2021: sec. 19.2 & 23). The essential problem is that the social justice literature calls for “relatively specific” (Gaus, 2021: 142) distributive results that maximize the benefit to the least advantaged, ensure that reward tracks merit, or compensate for luck and so on. And, given the third key property of a complex system, the governor cannot reliably predict which design(s) of the legal system will realize such fine-grained, society-wide distributive outcomes.

Insofar as the governor is concerned with influencing the systemic outcomes of the market order, she is better placed to design and enforce procedural rules of the game, such as property rights, to facilitate self-organization (2021: sec. 29.1). But, stresses Gaus (2021: 238), echoing Hayek, setting the rules of the game is not to pursue particular distributive outcomes: “Given excellent rules of the game, it can be played in diverse and surprising ways.” Such an institutional framework “allows an amazing array of social states, many of which we cannot yet imagine.”

This said, Gaus does think outcome-oriented legal reform is much more feasible if the desired outcome is closer to the discrete end of the spectrum, whereby the governor attempts to modify the outcomes of a specific “sector” of the social system, such as health or education (2021: 207). While daunting epistemic obstacles remain, these efforts may sometimes be successful.¹³ Such discrete goals include the supply of specific public goods (2021: sec. 22.4) and the provision of a social minimum (2021: 144). He is most optimistic about the feasibility of “micro-level governance” (2021: sec. 28.2). This involves conducting randomized control trials to identify ways to improve concrete institutions on the ground—schools, hospitals, neighborhoods, sanitation networks, etc.

Gaus thinks the governor can pursue discrete outcomes more feasibly than systemic ones because error inflation attenuates when we modify a relatively isolatable part of a complex system. The smaller the sub-system we intervene in, the fewer the interdependencies between actors, reducing the potential for predictive errors to compound and escalate (2001: 231).

In sum, Gaus grants, at least for the sake of argument, that formal legislation stems from a unitary governor but stresses how complexity constrains what they can achieve. The governor should set procedural rules of the game insofar as they aim to influence the market system as a whole and restrict outcome-oriented legal reforms to the pursuit of relatively discrete goals. They should not try to realize fine-grained, society-wide distributive outcomes. However, Gaus does not explicitly consider whether complexity arises within the state and the political process, such that the legal system is itself the emergent product of social interaction rather than the design of a unitary authority.¹⁴ This is the account I will advance in the next section. On my account, both the legal system and the systemic distributive outcomes to which it gives rise are emergent phenomena. The legal system does not stem from a unitary governor. It is the unintended result of the activity of plural “governors” and politically engaged citizens.

A complexity-based challenge to the unitary authority condition

I will now show how the five key properties of a complex system I outlined earlier apply to the state and the political process more generally. I first need to clarify my terminology. By the term “legal system,” I refer to first-order rules that directly govern the conduct of citizens rather than second-order “constitutional” rules, which govern how political agents reform the first-order rules. I will argue that while we cannot directly engineer the legal system as a whole, we can influence it indirectly by reforming constitutional rules.

Independent yet interdependent decision-making

My emergent account of the state takes inspiration from the multi-level governance literature within political science. That literature has documented a dispersal of rule-making authority from national governments to other decision-making centers within the state and non-state actors (Bache and Flinders, 2004; Hooghe and Marks, 2003). It seems to be a secular trend that decision-making authority within democracies becomes increasingly disaggregated over time.¹⁵

It is worth outlining a typology of the political agents who can exercise, to varying degrees, independent authority over the legal system. I distinguish between three classes of political agents. The first class consists of formal decision-making offices within the state. Here is a further subdivision of such formal offices:

Traditional branches of the national government. This category consists, of course, of the legislative, executive, and judicial branches.

Subnational governments. This category consists of regional governments, state governments (in a federal system), city governments, municipal governments, self-governing cultural and ethnic minorities, and so on.¹⁶

Administrative agencies. This category refers to unelected, specialist state bodies. They play a pervasive role in the governance of consolidated liberal democracies. They are responsible for a diverse array of functions, from providing services such as central banking to regulating various activities to providing legal rulings (Vibert, 2007: 20).¹⁷ Summarizing their influence, Vibert (2007: 7) writes:

The air we breathe, the water we drink, the food we eat, the electricity we use, the phone calls we make, the value of the coins and banknotes in our pockets, our access to media, the disputes we get involved in, are all influenced in basic ways by their activities.

International government organizations. The European Union, the International Monetary Fund, and the World Trade Organization are among the more well-known entities with substantial decision-making power independent of national governments.

The second class of political agents is the electorate. Citizens, in their capacity as voters, affect law-making by selecting some—though certainly not all—important decision-makers within the state. I would stress that voting in national elections is just one avenue through which citizens affect the political process. They can also vote at the subnational level and, in the European Union at least, at the international level. In other contexts, citizens vote directly through referendums.

The third class of political agents consists of politically engaged civil society associations. Some such associations have the explicit purpose of influencing the political process, such as political parties and campaign groups. Many other civil society associations, such as commercial firms or churches, do not have a political purpose primarily but nonetheless exert political influence.

It is not difficult to see how decision-making by these political agents will be interdependent and marked by feedback effects. Here is one stylized example. Suppose an independent central bank increases interest rates, raising the cost of borrowing for the central government, which responds by cutting back national spending. The central government's decision has, inter alia, consequences for local government actors. Perhaps they receive smaller transfers from the national government or otherwise must fill the gap left open by reduced social provision by the national government. They thus increase local taxes. The combination of higher local taxation and national austerity affects citizens’ political behavior in terms of how they vote and/or engage with the political system via civil society associations. The modified behavior of citizens will, in turn, lead to changes in central and local government. Perhaps they will be voted out of office or be forced to adjust their political behavior to avoid being voted out. Other political agents will, in turn, respond to these events. The critical point is that myriad political agents are making independent yet interdependent decisions in response to one another.

Emergent legal system

The first-order legal system is the emergent product of these independent yet interdependent political decisions. We cannot trace the overall result to the intentions of the political agents individually, the political agents understood as a unified collective agent, or a supreme authority within the state (I will say more on the latter two points anon.) I stress that the overall shape of the legal system is the emergent phenomenon. To be sure, particular actors have the authority to reform particular parts of the legal system. In the United States for instance, the Federal Reserve can modify interest rates, the President might change foreign policy, the Federal Drug Administration can determine which drugs can be legally sold, the National Rifle Association can influence gun legislation, and so on. However, the way such disaggregated yet interdependent rule-making initiatives combine to form an overall legal system is emergent (in addition to the society-wide distributive outcomes produced by the legal system). In other words, there is no agency with the authority to oversee the legal system as a whole. By analogy, market actors have the authority to make particular economic decisions. However, none have the authority to determine the overall distribution of income and wealth because that is the emergent product of many economic decisions.

Range predictions of the systemic legal outcomes of the political process

On the emergent state view, the precise legal system that any society converges upon is the contingent result of how political interaction plays out over time. Nonetheless, we can make range predictions of the legal systems that emerge. For example, all contemporary liberal democracies are variants of welfare-state capitalism, combining protection for procedural market rules with more outcome-oriented rules designed to offer social protection and regulate specific activities. However, a wide range of legal systems are compatible with this general characterization. Consider how the legal systems of Japan, France, Denmark and the United States differ substantially in specific details even though all accurately fit the general descriptor of “welfare-state capitalist societies.” These legal differences reflect how the contingent course of political interaction has played out differently in these countries. Error inflation constrains our ability to predict the specific legal systems that emerge from the political process just as it constrains our ability to predict the society-wide distributive outcomes produced by those systems.

Procedural rules and self-organization

Recall the puzzle of emergence on the market. We need to explain why such systemic market outcomes can be functional, at least in some respects, absent an economic central planner. We saw that procedural rules of property and exchange facilitate decentralized self-organization on the market in various ways. An emergent account of the state faces an analogous puzzle. We know that particular political agents have authority to reform or initiate particular laws, be these procedural rules of property and exchange or discrete outcome-oriented policies. But an emergent account of the state needs to show how the legal system that emerges from these many decisions can be self-organizing rather than chaotic in the absence of an overarching authority.

We can conceive of the complex political process as governed by a distinctive set of procedural rules. I call these constitutional rules to distinguish them from first-order legal rules that directly govern citizens. Constitutional rules perform a function for the political process like that performed by rules of property and exchange for the market process. Rather than dictating specific reforms, these rules condition how political agents make independent decisions. We can make a stylized distinction between two kinds of constitutional rules.

First, jurisdictional constitutional rules divide decision-making authority within the state, mirroring how property rights divide control over economic resources. Such constitutional rules may divide authority by function, distinguishing between the legislative, judicial, executive and administrative branches. Alternatively, they may divide authority geographically, differentiating between international, national and subnational jurisdictions. Second, contestatory constitutional rules govern how citizens interact with formal state offices. These include voting rules governing how citizens select their representatives, rules governing the formation and operation of political parties, and campaign finance and lobbying regulations, which specify how citizens can influence formal state offices through means other than voting.

We can tentatively argue that constitutional rules enable self-organization within the state in a manner roughly analogous to how rules of exchange and property function in the market. Let us first consider jurisdictional constitutional rules. These can moderate the interdependencies within the legal system if they do a reasonably good job of internalizing externalities and thereby promote a basic threshold of stability and predictability. Further, when jurisdictional rules divide authority by function, they allow the state to enjoy the benefits of the Smithian division of labor. Indeed, one prominent theory behind the rise of administrative agencies is that legislative and judicial officials prefer to leave specialist tasks to experts with the requisite technical knowledge (Vermeule, 2016).

When jurisdictional constitutional rules divide authority geographically, such as via federalism, they promote a Hayekian discovery process within the policy-making process. Different jurisdictions can experiment with different policies, and successful policies can be adopted more broadly via processes of imitation without any comprehensive legal oversight. Consider the legalization of gay marriage in the US. This policy was first pioneered at the subnational level by states such as Massachusetts before being recognized by other states and eventually the Supreme Court. Likewise, Elinor and Vincent Ostrom have argued that competing municipal or metropolitan authorities can be surprisingly effective in providing various public goods, including policing and environmental management.¹⁸ One key reason for this is parallel learning—“information about what has worked well in one setting can be transmitted to others who may try it out in their settings” (Ostrom, 2005: 184).

Contestatory constitutional rules play a critical role in enabling the political process to incorporate dispersed information about the needs, preferences and values of myriad citizens. We saw that the Hayekian price mechanism performs this function for the market process. More specifically, prices aggregate and transmit information to market actors of which they have no direct awareness. The political process generally cannot utilize the Hayekian price mechanism to utilize dispersed information. This means it faces the following kind of problem. A policymaker might initiate a policy with the intention of benefiting a particular group, but that (unbeknownst to them) does not sufficiently consider the interests of another group. Perhaps they create a property right that empowers some citizens but has negative externalities for others, or they introduce a welfare policy that excludes citizens with hidden disabilities. The political process thus needs a mechanism equivalent to market prices that can transmit information to political agents of which they have no direct awareness.

Contestatory constitutional rules perform this function by empowering citizens to vote or otherwise influence the political process through civil society associations or political parties. Such rules enable citizens neglected by a political decision to make their voices heard. Thus, in a society with robust contestatory rules, the group whose interests our policymaker neglected can respond, either by how they vote or by campaigning directly, to have the policy modified to account for their interests. Of course, the modification might neglect the interests of a third group of citizens who will, in turn, organize to further their interests, and the process of iterative political adjustment will continue. However, in this way, the political process can incorporate diverse citizen interests in the absence of a more comprehensive oversight. I here draw on the democratic theory of Charles Lindblom who gives an example of how a state's distributive policy might emerge in this fashion.

… [N]o part of government attempts a comprehensive overview of policy on income distribution. A policy nevertheless evolves, and one responding to a wide variety of interests. A process of mutual adjustment among farm groups, labor unions, municipalities and school boards, tax authorities, and government agencies with responsibilities in the fields of housing, health, highways, national parks, fire, and police accomplishes a distribution of income in which particular income problems neglected at one point in the decision processes become central at another point (Lindblom, 1959: 85; emphasis added; c.f. Lindblom, 1965).

Indirect reform of the legal system

Now, as we saw, a complex account of the market does not rule out deliberate initiatives to improve systemic market outcomes. However, as Gaus and Hayek insist, we must focus our efforts on improving the procedural rules that enable self-organization in the market. I think we can make an analogous argument about the emergent legal system. We can aim to improve the legal system as a whole by modifying constitutional rules. For instance, we could improve contestatory constitutional rules to make them responsive to a wider variety of interests. Alternatively, we could amend jurisdictional constitutional rules to mitigate externalities between political authorities, thereby broadening the scope for policy experimentation and/or leveraging specialized knowledge.

However, the purpose of this indirect reform of the legal system would be relatively modest. Just as good rules of property and exchange do not guarantee the emergence of any fine-grained distributive outcome but make more likely the emergence of functional distributive outcomes, so do good constitutional rules make the emergence of functional (i.e. more responsive or innovative) legal systems more likely without guaranteeing the emergence of any fine-grained legal system. That would depend on how the contingent process of interdependent political activity unfolds.¹⁹

Emergent versus unitary state

The emergent account of the state I have just outlined differs markedly from the more prevalent unitary view. In this final section, I make the case explicitly for why we should accept the emergent view. I compare the emergent view to two different models of the unitary view.

The deliberative model of the unitary state

The deliberative model of the unitary state accepts that decision-making authority within the state is disaggregated. It does not posit a supreme decision-making authority to whom all other state agents are subservient. However, it does posit the existence of a decision-making procedure that enables the various offices of state to act as a unified group by producing “consistent, rational decisions (as an individual agent usually does)” (Lawford-Smith, 2019: 29).

Anna Stilz (2011) and Holly Lawford-Smith (Ibid.) model the state this way. Lawford-Smith (2019: 86) recognizes explicitly that the state will contain many departments and divisions capable of making independent decisions but suggests there is an “overarching coordination … sufficient to resolve coordination and cooperation problems between them.” By this, she does not mean the decentralized forms of self-organization I discussed in the previous section, but a more deliberate method of coordinating the behavior of state agencies. For example, she remarks that departments “often work together on projects, consulting with each other before producing final reports and recommendations” and concludes that this means “the group is unified in the way required by the Unitary Actor model of the state” (2019: 86–87).

Stilz holds that a collection of individuals is incorporated into a collective agent when a constitution governs them that “specifies binding decision procedures and allocates responsibilities to various roles” (2011: 192). She thinks this condition is met regarding the various offices of the state:

When we say “state” … we refer to a structure of institutions that governs a territory. These institutions define various offices—including the executive, legislative, and judicial branches; the police; and the bureaucratic administration—that make and enforce law on that territory. The institutional rules allocate responsibilities across these offices, and specify the procedures used in taking collective decisions. Like an incorporated group, then, the state is defined by its internal “constitution.” This constitution allows it to form intentions by means of standing decision procedures (2011: 195; emphasis added)

The unitary authority condition looks sound if the deliberative model is right. On Stilz's version of the story, the constitution provides a decision-making mechanism by which myriad political agents can consciously coordinate their individual responsibilities over parts of the legal system into a more comprehensive oversight. Pace the emergent view, the whole legal system need not be an emergent phenomenon that escapes the intentions of the plural political agents whose myriad rule-making initiatives are causally responsible for it. Instead, state actors can use the constitution to form group-level intentions to deliberately design the legal system.

Before I make the case for why we should accept the emergent model, I should clarify the difference between it and the deliberative model. The emergent view does not dispute the existence of collective agents as such. It recognizes that the state contains a variety of organizations that meet the conditions for collective agency, namely, they have decision procedures that enable the formation of common goals and allocate responsibility for achieving those goals. Such organizations include legislatures, courts, and administrative agencies. Furthermore, as part of the political process, these formal state offices will interact with myriad civil society associations that are also collective agents, such as political parties, business associations, campaign groups and so on. The difference between the emergent and deliberative models hinges on whether an overarching mode of coordination constitutes myriad state offices into one grand collective agent with the authority to design the legal system holistically.

I think Stilz is right that constitutional rules are the best candidate for such an overarching mode of coordination. However, I also think she mischaracterizes the nature of such rules. I have argued that constitutional rules are not categorically different from rules of property and exchange that govern the market process. The constitution regulates the interaction between political agents, many of whom will be collective agents, but does not constitute them as an over-arching collective agent. Rather, it conditions how political agents make otherwise independent decisions. To continue the market analogy, many would accept that Amazon and Walmart count as collective agents. Yet it is a mistake to claim that rules of property and exchange constitute them and myriad other economic agents into an over-arching, “economic” agent. That claim would misconstrue the function of property and exchange rules. For those rules fragment rather than unify economic decision-making.

Similarly, jurisdictional constitutional rules fragment legal decision-making. The separation of powers fragments authority geographically between national and local governments and functionally between the executive, judicial, legislative and (increasingly) administrative branches of the state. Stilz is right to say that such rules “allocate responsibilities across these offices” but I think she is mistaken to suggest that they specify “procedures used in taking collective decisions” across the whole state. Rather, jurisdictional constitutional rules disaggregate decision-making within the state. Moreover, we saw that contestatory constitutional rules enable a process of iterative adjustment between independent decision-makers rather than enjoining them to act on consciously-decided common goals.

Furthermore, it is very difficult to envisage a standing procedure that allows the numerous and differentiated offices of the state to make decisions as a unified group. The epistemic division of labor within the state is considerable, as it is in the market setting. Just the administrative branch of the contemporary state has many specialized branches, regulating diverse domains ranging from nuclear to monetary policy. If a decision-making procedure is to empower these myriad actors to act as a group, it would need to gather and synthesize qualitatively different information that is dispersed among them. It is very difficult to envision a decision-making procedure that could accomplish this task given the diversity of agents involved and the often specialized nature of the information that the procedure would need to aggregate.

The centralized model of the unitary state

The centralized model of the unitary state rests on the familiar claim that the state must possess a single, supreme source of authority. Proponents of this view might accept my claim that constitutional rules disperse decision-making power within the state rather than enabling state agencies to act as one unified agent. However, they could argue that such dispersal merely reflects a choice by the sovereign authority to delegate its powers. On this view, the supreme authority could, if it chose, re-appropriate decision-making authority and holistically oversee the legal system. Thus, the emergent account of the state mistakes a contingent and alterable feature of the state for an essential one.

In response, let us first consider the claim that the state must possess such a supreme authority. It dates back, of course, to the social contract tradition (Hobbes in particular but also Bodin, Rousseau and Kant). It can be understood in two ways. The first is conceptual and was advanced by Hobbes. The state must have a final, supreme authority to prevent logical regress. If a state actor's authority is constrained in some way, this implies there must be a higher authority—“for he that prescribes limits, must have a greater power than he who is confined by them” (Hobbes, 1984). So, we just follow the chain of command, as it were, until we find the agency with supreme authority.

However, as Morris (2002: chap. 7) observes, this conceptual argument is flawed because it assumes that power relations between agents are the same across all dimensions—that is, A has power over B in one dimension iff A has power over B in all dimensions. But, Morris observes, this assumption seems straightforwardly false. The Pope may have more power over Catholic doctrine than the King, while the King has a bigger army. If power relations between actors can vary across dimensions, the fragmentation of authority creates no conceptual difficulties. In the United States, we can say the President has more power over the Supreme Court on the matter of declaring war, while the latter has more power over constitutional interpretation—without creating any obvious regress problem.

The second version of the argument for the necessity of a supreme authority is more practical. This version need not deny that a state with an effective separation of decision-making powers is a conceptual possibility. However, it would insist that such a state could not long endure. A stable state requires a supreme authority to act as the final arbiter of disputes between other political agents. Otherwise, disputes could escalate into constitutional crises or even civil war.²⁰ My initial response is that this version of the argument seems refuted by empirical experience. Liberal democratic states divide power across a variety of dimensions yet have proven to be stable polities.

Perhaps this initial response looks question-begging. Maybe the very stability of these states implies that their dispersed authority is only an appearance; that there really is a supreme arbiter keeping the peace. But it is unclear who or what would have such authority. Candidates for such a supreme authority do abound—the people, the legislature, the executive, the supreme court, the governing party, the bourgeois, the managerial elites and so on. Yet the authority of all these groups or agencies is constrained in one way or another in existing liberal democracies.

Still, suppose we set these concerns aside and granted the existence of a supreme authority within the state—wherever it may reside. Several considerations suggest that the sovereign could not reclaim its delegated authority even if it technically had the authority. The de jure existence of a supreme authority would not prevent the legal system from being the emergent result of many independent rule-making decisions de facto.

First, senior state authorities often delegate power to make credible commitments to third parties. This is typically the rationale for delegating authority to independent administrative agencies (Vibert, 2007). For instance, elected officials often delegate authority over monetary policy to independent central banks. Such a policy aims to credibly signal to key economic actors, such as owners of government bonds, that elected officials will not abuse interest rate policy for short-run electoral purposes. Thus, in many cases, the re-appropriation of delegated authority by the supreme authority would undermine such commitments. One might argue that this trade-off is acceptable, but this is doubtful because it would come at a heavy expense to the state's capacity to pursue important social goals. For instance, a national government that lacks favorable standing in global capital markets would have a limited ability to redistribute and finance public goods.

The second consideration is based on the epistemic division of labor I mentioned just above. Many areas of governance require specialist knowledge that the supreme authority would lack. Consequently, the supreme authority would have to delegate de facto decision-making to specialists. Relatedly, there are only so many hours in a day. Supposing ordinary human beings controlled the supreme authority, they would be temporally and cognitively constrained to oversee only parts of the legal system at any one time.

The third consideration is based on an insight of Carl Schmitt (2005: 148). He argues that implementing a rule has an unavoidably “decisionistic and personalistic” element to it. His point is not merely the familiar one that law is indeterminate and requires interpretation. Schmitt says who interprets the rule matters. That one interpretation rather than another will be selected is a function of the particular official's personality, circumstances and so forth.²¹ This was precisely what Michael Lipsky (2016) found in his famous study of “street-level bureaucrats.” Central state policies list several criteria or values for subordinates to uphold without specifying how to make trade-offs between them. These and other such decisions are left to street-level bureaucrats who interact with citizens on the ground. Lipsky concludes that such bureaucrats have significant discretion in substantively deciding how to implement such schematic policies. We can generalize his insight. The more the supreme power centralizes, the more it will be forced to legislate schematically, such that the substantive content of rules will be left to the discretion of subordinates lower down the hierarchy.²²

Conclusion—wider implications

The emergent view of the state has important implications for several debates within political philosophy. I highlight two of these in this concluding section. Though, unfortunately, I will have to postpone a full elaboration of both to other occasions.

First, the emergent state view challenges the dichotomy between an emergent civil society and an intentional state common among political theorists, who often subscribe to the following kind of view (e.g. Nagel, 1991; Rawls, 1996: chap. 7; Valentini, 2017). Civil society actors cannot shoulder the burden of realizing important social goals because coordination problems beset them. Interaction between private individuals, charities, commercial organizations, religious associations and so on leads to emergent outcomes that none of the constituent actors intended. On the other hand, the unitary state can take a comprehensive overview of society, issuing regulations as necessary to guide the interaction of civil society and market actors towards desirable social outcomes. We thus require an institutional division of labor between the state and private actors in civil society. The task of private citizens, firms and civil society associations is to pay their taxes, support just political institutions and comply with the law. Otherwise, they do not have broader social obligations and can use their (post-tax) resources on their private concerns. The broader responsibility for ensuring justice and arranging the basic structure of society falls on the unitary state.

However, if my account is correct, the state is not insulated from the complexity that affects civil society. Both state and civil society are self-organized, emergent orders. We need to account for this sociological fact when considering which institutional mechanism to rely upon to realize important social goals. In other words, our institutional evaluation of whether we should rely on the state or civil society to realize important social goals should be conducted on the symmetrical assumption that complexity is present in both domains. I believe, all other things being equal, this will show civil society to be a more attractive means of promoting social change than political philosophers have tended to assume.

Second, the emergent view has important ramifications for the debate on the distribution of responsibility for wrongdoing by state actors. Participants in this debate generally take the unitary authority condition for granted. They generally grant that the state—qua unitary collective agent—can be held holistically responsible for the wrongful acts of its individual members and that this responsibility is transmitted at least to all state employees who voluntarily took up their employment (Lawford-Smith, 2019: chap. 6; Pasternak, 2013: 361–362; Stilz, 2011: 196). The debate has focused on the extent to which the wider citizenry can also be held responsible and the form such responsibility might take.

The emergent view questions the shared premise of this debate. If the state is not a unitary actor, then the whole state cannot be held responsible for wrongdoing by its parts. The procedural nature of constitutional rules cannot establish shared responsibility. Allow me one last market-based analogy. Procedural rules of property and exchange govern commercial entities, yet we do not think wrongdoing by one corporate actor implicates others simply by virtue of their sharing that legal framework. Thus, we might agree with Stilz (2011: 193) that all shareholders and employees of Air New Zealand are responsible (albeit to varying extents) for one of their planes crashing if the crash can be traced to Air New Zealand's faulty corporate procedures. Still, neither other airlines nor their stakeholders share this responsibility.

Similarly, suppose Congress and the President authorize an unjust war. Congress as a corporate body, the President as an individual, and arguably the US military are jointly responsible for the injustice. However, the governor of the Federal Reserve (say) does not share any responsibility because they were not involved in the decision to go to war, and the Federal Reserve is a distinctive organization within the US state that is not concerned with military policy. This account will also have to answer the complicated question of how and if citizens can bear responsibility for the wrongdoing of particular state actors.

Footnotes

Acknowledgments

I am grateful to Jacob Barrett, Brookes Brown, Tom Christiano, Aliya Dewey, Christopher Freiman, Otto Lehto, Pablo Paniagua, Alex Schaefer, Sam Thomas, and Kevin Vallier for helpful feedback and discussion.

Declaration of conflicting interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Kaveh Pourvand

Notes

References

Arthur

(2021) Foundations of complexity economics. Nature Reviews Physics 3: 136–145.

Bache

Flinders

(eds) (2004) Multi-Level Governance. Oxford: Oxford University Press.

Blake

(2001) Distributive justice, state coercion, and autonomy. Philosophy & Public Affairs 30(3): 257–296.

Brown

(2023) The romantic lie in the brain: collective agency, moral responsibility, and the state. Oxford Studies in Political Philosophy 10: 57–85.

Dworkin

(2002) Sovereign Virtue: The Theory and Practice of Equality, 4. printing Cambridge, Mass.: Harvard University Press.

Ferguson

(1995) An Essay on the History of Civil Society. Cambridge Texts in the History of Political Thought. Cambridge; New York: Cambridge University Press.

Freeman

(2007) Rawls. Routledge Philosophers. London; New York: Routledge.

Gaus

(1998) Why all welfare states (including laissez-faire ones) are unreasonable. Social Philosophy and Policy 15(2): 1–33.

Gaus

(2012) The Order of Public Reason: A Theory of Freedom and Morality in a Diverse and Bounded World. Cambridge: Cambridge University Press. http://dx.doi.org/10.1017/CBO9780511780844.

10.

Gaus

(2016) The Tyranny of the Ideal: Justice in a Diverse Society. Princeton: Princeton University Press.

11.

Gaus

(2021) The Open Society and Its Complexities. Philosophy, Politics, and Economics. New York, NY: Oxford University Press.

12.

Hayek

(1945) The use of knowledge in society. The American Economic Review 35(4): 519–530.

13.

Hayek

(1967a) Degrees of explanation. In: Studies in Philosophy, Politics and Economics. Chicago: University of Chicago Press, 3–21.

14.

Hayek

(1967b) The theory of complex phenomena. In: Studies in Philosophy, Politics and Economics. Chicago: University of Chicago Press, 22–42.

15.

Hayek

(1982) Law, Legislation and Liberty: Vol. 1 Rules and Order. Abingdon: Routledge.

16.

Hayek

(2002) Competition as a discovery procedure. Translated by Snow Marcellus. The Quarterly Journal of Austrian Economics 5(3): 9–23.

17.

Hayek

(2006) The Constitution of Liberty. London: Routledge.

18.

Hobbes

(1984) De Cive. Oxford: Clarendon Press.

19.

Hooghe

Marks

(2003) Unraveling the central state, but how? Types of multi-level governance. American Political Science Review 97(2): 233–243.

20.

Ismael

(2016) How Physics Makes Us Free. New York: Oxford University Press.

21.

James

(2005) Power in social organization as the subject of justice. Pacific Philosophical Quarterly 86(1): 25–49.

22.

Lawford-Smith

(2019) Not In Their Name: Are Citizens Culpable For Their States’ Actions? 1st ed. Oxford: Oxford University Press.

23.

Lewis

(2012) Emergent properties in the work of Friedrich Hayek. Journal of Economic Behavior & Organization 82(2–3): 368–378.

24.

Lewis

(2016) The emergence of ‘emergence’ in the work of F. A. Hayek: a historical analysis. History of Political Economy 48(1): 111–150.

25.

Lindblom

(1959) The science of ‘muddling through’. Public Administration Review 19(2): 79.

26.

Lindblom

(1965) The Intelligence of Democracy: Decision Making through Mutual Adjustment. New York: Free Press.

27.

Lipsky

(2016) Street-Level Bureaucracy: Dilemmas of the Individual in Public Service. Vol. 1. Oxford: Oxford University Press. https://doi.org/10.1093/oxfordhb/9780199646135.013.19.

28.

List

Pettit

(2011) Group Agency. Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199591565.001.0001.

29.

McGinnis

(ed) (1999) Polycentricity and Local Public Economies: Readings from the Workshop in Political Theory and Policy Analysis. Ann Arbor, MI: University of Michigan Press. https://doi.org/10.3998/mpub.16061.

30.

Meckled-Garcia

(2008) On the very idea of cosmopolitan justice: constructivism and international agency. Journal of Political Philosophy 16(3): 245–271.

31.

Miller

(2003) Principles of Social Justice. Cambridge, Mass.: Harvard University Press.

32.

Morris

(2002) An Essay on the Modern State. 1. paperback ed. Cambridge: Cambridge University Press.

33.

Nagel

(1991) Equality and Partiality. New York: Oxford University Press.

34.

Nagel

(2005) The problem of global justice. Philosophy & Public Affairs 33(2): 113–147.

35.

North

Wallis

Weingast

(2009) Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History. Cambridge and New York: Cambridge University Press.

36.

Ostrom

(2005) Understanding Institutional Diversity. Princeton Paperbacks. Princeton: Princeton University Press.

37.

Pasternak

(2013) Limiting states’ corporate responsibility*. Journal of Political Philosophy 21(4): 361–381.

38.

Pettit

(2014) How to tell if a group is an agent. In: Lackey

(eds) Essays in Collective Epistemology. Oxford: Oxford University Press, 97–121.

39.

Pogge

(1992) Cosmopolitanism and sovereignty. Ethics 103(1): 48–75.

40.

Rawls

(1996) Political Liberalism. 2nd ed. New York: Columbia University Press.

41.

Rawls

(1999) A Theory of Justice: Revised Edition. 2nd ed. Cambridge: The Belnap Press of Harvard University Press.

42.

Schmitt

(2005) Political Theology: Four Chapters on the Concept of Sovereignty. University of Chicago Press ed. Chicago: University of Chicago Press.

43.

Smith

(2009) An Inquiry into the Nature and Causes of the Wealth of Nations. Edited by R. H. Campbell. 11. print. Indianapolis, Ind: Liberty Fund.

44.

Stilz

(2011) Collective responsibility and the state. Journal of Political Philosophy 19(2): 190–208.

45.

Valentini

(2017) The natural duty of justice in non-ideal circumstances: on the moral demands of institution building and reform. European Journal of Political Theory 20(1): 45–66.

46.

Vermeule

(2016) Law’s Abnegation: From Law’s Empire to the Administrative State. Cambridge, Massachusetts: Harvard University Press.

47.

Vibert

(2007) The Rise of the Unelected: Democracy and the New Separation of Powers. Leiden: Cambridge University Press.