The theory of spontaneous order has been the object of a large number of studies in the history of political thought. Authors such as Bernard de Mandeville ( 1732), David Hume ( 1975,  1978), Adam Smith ( 1976), and Carl Menger (1963) were among the first to develop a fully fledged account of how institutions spontaneously develop as a result of the unintended design arising out of intentional human actions. Similarly, authors such as Robert Nozick (1974) and Anthony de Jasay (1989) used the spontaneous order as a tool to show whether public goods can be produced by simply relying on individuals' self interest. However, the popularity of the spontaneous order has to be ascribed to Friedrich Hayek, who dedicated a large part of his work to developing the concept.
Although such tradition encompasses many important authors, political theorists have largely neglected it in the past twenty years. Yet authors such as Gerald Gaus (2011) and John Tomasi (2012) have recently insisted on the relevance that the spontaneous order has or should have within the classical liberal tradition.
In this paper, we aim to revive the research project on the spontaneous order by examining it critically. We aim to show that normative formulations of the spontaneous order suffer from one main flaw: they focus on the origin of orders rather than on how orders actually perform.
In particular, we argue that such normative formulations tend to qualify orders as spontaneous according to two main requirements: unintendedness and negative liberty. The first requirement prescribes that to be considered spontaneous an order must not be the result of human design but the unintended consequence of human actions. The second requirement, in contrast, prescribes that an order is spontaneous when it arises out of free individuals' interactions.
Our main goal is to show that both requirements tell us very little about how orders actually perform and in fact justify a large variety of institutional arrangements that many classical liberal theorists would qualify as unacceptable.
Our second aim is to propose a new formulation of the spontaneous order that focuses on orders' actual performance (actual spontaneity). Such formulation has been inspired by Hayek's insistence on the importance that social orders be able to adapt to new circumstances--in particular, on their ability to find efficient solutions to coordination problems through the use of knowledge dispersed among individuals.
Cooperation and coordination depend on an order's ability to adapt to changes in individuals' preferences and beliefs, and it is this ability, to our mind, that should determine the spontaneity of an order.
Although we do not indicate the institutional arrangement that may satisfy this condition, we argue that designing such an institutional framework is not a prohibitive enterprise. Hayek himself seemed aware of this possibility when he claimed that "[w]e can 'plan' a system of general rules, equally applicable to all people and intended to be permanent (even if subject to revision with the growth of knowledge), which provides an institutional framework within which the decisions as to what to do and how to earn a living are left to the individuals. In other words, we can plan a system in which individual initiative is given the widest possible scope and the best opportunity to bring about effective coordination of individual effort" ( 1997, 194).
Orders, we argue, do not need to be the result of unintended design in order to be spontaneous. The difference between a spontaneous and a nonspontaneous order should not depend on how an order has originated but on its actual ability to adapt to new circumstances.
Furthermore, we argue that even "designed" spontaneous orders may evolve into nonspontaneous orders if individuals who inhabit them lack certain personal qualities. Diversity, or our capacity to empathize with diversity, is one of these qualities insofar as a homogeneous society w'ould naturally tend to manipulate its institutional arrangements in order to jeopardize diverse perspectives.
In the first half of this paper, we analyze two formulations of the spontaneous order that have emerged from Hayek's work. Here, we argue that both accounts fail to provide a genuinely normative account of the spontaneous order. In the second half, we sketch a third formulation that focuses on what we have been defining as the "spontaneity of orders." Specifically, this formulation measures orders' ability to adapt to new circumstances. We also briefly discuss the role that diversity plays in preserving orders' spontaneity.
Spontaneous Order and Evolution
The first formulation of the spontaneous order we wish to explore focuses on a descriptive account of the evolution of orders. In particular, Hayek seems to insist on the "twin ideas of spontaneous order and evolution," according to which rules and institutions that shape social orders are not the result of human design but rather of a largely unintended evolutionary mechanism.
According to Gaus (2006), Hayek identifies three main ways in which norms evolve: group survival, group growth, and an endogenous mechanism. The first two mechanisms capture the idea of an intergroup selection of rules:
Although the existence and preservation of the order of actions of a group can be accounted for only from the rules of conduct which individuals obey, these rules of conduct have developed because the individuals have been living in groups whose structures have gradually changed. In other words, the properties of the individuals which are significant for the existence and preservation of the group, and through this also for the existence and preservation of the individuals themselves, have been shaped by the selection of those individuals from the individuals living in groups which at each stage of evolution of the group tended to act according to such rules as made the group more efficient. (Hayek 1967, 72)
According to those two mechanisms, a process of imitation of the more successful groups shapes the evolution of norms. If, for instance, Peter and Kate, under the set of norms k, managed to survive or to achieve better outcomes than Alf and Betty did under set z, the process of imitation will lead Alf and Betty to shift toward k. It is important to point out that such evolution may not be the result of Alf and Betty's explicit will or awareness but the result of a partially unintended process of imitation. Furthermore, at the intergroup level, the process of imitation normally concerns the whole set of norms instead of particular rules owing to the impossibility, in Hayek's perspective, to isolate the effects of particular norms on the general outcomes. The process of imitation of successful groups is, in Hayek's mind, a sort of Darwinian device that embeds locally dispersed knowledge and that is supposed to lead to the diffusion of efficient sets of norms and institutions.
The endogenous mechanism, in contrast, works at the intragroup level and is shaped by the "competition between individuals." According to Gaus, "This stress on individual competition and the evolution of rules suggests that, instead of a competition between social orders, Hayek has in mind a competition between individuals within a social order that leads to the selection and evolution of rules" (2006, 244).
Specifically, such intragroup evolution is shaped by nonrandom deviations from the existing set of rules. According to the endogenous mechanism, rules that do not...