Abstract The paper presents the thought of the political philosopher Cornelius Castoriadis on institutions and argues that his insights contribute significantly to the development of a theory of institutions and institutional change which moves beyond the neoclassical framework. Stylised facts of historical experience such as strong path-dependence and lock-in effects and the diversity of observed institutional forms are explained naturally in his proposed framework. In addition, a schema for institutional change is suggested, which accommodates a number of other strands in the literature as special cases. This schema could be elaborated in various directions and facilitates structuring the further development of an alternative institutional theory. 1 Introduction The emphasis on institutions has been a common trend among both neoclassical and alternative theoretical approaches. The former strand of literature has produced a number of influential studies which interpret historical and institutional evolution through the lenses of the underpinning neoclassical concepts. The latter strand has been less unified and includes both ‘old-institutionalist’ approaches and modern alternative ones. Studies following alternative methodologies consist both of critiques of the limitations characterising the neoclassical approaches, exploring different directions such as the assumptions on human motivation, the effect of culture, and the nature and scope of and the human cognition process, and interpretations of historical test cases which depart from the neoclassical paradigm. Given that almost all alternative approaches represent attempts to encompass more complex conceptualisations of human behaviour than those which are modelled within the neoclassical microeconomic framework, there is always the concomitant challenge of formalising these approaches so that they can be further developed. This paper aims to present an outline of Cornelius Castoriadis’s thought on institutions and to propose a schema of institutional change based on his ideas. Despite the fact that Castoriadis worked as a highly ranked professional economist in the OECD, he is most well known for his work on political science and political philosophy. His has been a point of reference for significant contemporary philosophers, e.g. Habermas (1987) and Rorty (1991). The sociologist Edgar Morin characterised Castoriadis as a ‘titan of thought’. The paper argues that Castoriadis’ analyses represent also an original contribution to institutional economics and especially on the attempt to move beyond the neoclassical framework. Moreover, his conceptual framework is amenable to codification and flexible to accommodate other contributions, therefore useful as a starting point for structuring the discussion and conducive to further development. Castoriadis’s work presents insightful explanations of phenomena such as strong path-dependence, lock-in effects and the observed diversity of institutions in historical societies. The central concept of the ‘social individual’ is used in this direction. In addition, the latter concept provides a promising avenue to elaborate on human agency, striking a middle ground between the two extremes of methodological individualism and holist modes of explanation (see Toboso, 2001; Mayhew, 1987). Moreover, his decomposition of institutions into a functional and a symbolic (operational) component provides explanations on the actual outcomes and of failures of real-world institutions. The paper is structured as follows. Section 2 provides biographical information on Castoriadis. Section 3 presents Castoriadis’s views on institutions and institutional change. The subjects of the components of institutions and his views on human behaviour are elaborated. In addition, a schema of institutional change is proposed and its relation with modern institutional theories is discussed. Section 4 concludes the paper. 2 Biographical information Castoriadis (1922–97) was born in Istanbul. His father, a ‘Voltairean Francophile’ (Curtis, 1997, p. vii), decided to move the family to Athens given the uncertainty surrounding the political developments in Turkey in the midst of military confrontation between Greece and Turkey. In the context of pre-war Greece, governed by the authoritarian regime of Ioannis Metaxas, Castoriadis developed an interest in politics even as a teenager. He joined a leftist group of Trotskyist orientation after entering into opposition with the pro-Stalinist Communists. In 1945, he immigrated to Paris, along with a number of other Greek intellectuals, after receiving a scholarship for a doctorat d’État in philosophy at the Sorbonne in Paris. He collaborated with Claude Lefort (a student of the philosopher Maurice Merleau-Ponty) and in 1949 they published the inaugural volume of their journal Socialisme ou Barbarie. The journal criticized Stalin and the Soviet regime, arguing that the monopoly power of the party bureaucracy with respect to the means of production had rendered the position of workers even weaker compared to a capitalist economy (Linden, 2007, pp. 116 ff). The group by the same name had been one of the main influences of the May 1968 student revolt. Specifically, one of the leaders of the events, Daniel Cohn-Bendit, had claimed in 1968 that he had ‘plagiarised’ Castoriadis and the Socialisme ou Barbarie analyses (Curtis, 1997, p. vii). Castoriadis worked at the Organization for European Economic Cooperation (OEEC, the precursor of the OECD) from 1948 until 1970 and became a Director of Statistics, National Accounts and Growth Studies. This experience allowed him to get insights on the economies of developed and developing nations and informed his concurrent theoretical work. After he left the OECD, he taught at Paris’s École des Hautes Études en Sciences Sociales in 1979, where he taught until he died in 1997. His work covered political theory in a wide sense, including investigating the evolution of history and institutions and also the role of individual. In a recent article in the OECD Insights series, Dosse (2014) argues that Castoriadis’s analyses on the factors which are necessary to deepen the autonomy of a society, in the sense of understanding that individuals constitute their societies on the basis of a voluntary collective act, are very relevant for thinking about the historical evolution of the world at the outset of the twenty-first century. One of the very few instances where Castoriadis’s work is examined in an economic context is Patalano (2007). That paper examines Castoriadis’s analysis of institutional change as a complementary view and contrasts it to the cognitive institutionalism expressed by Mantzavinos et al. (2004). A broader examination of Castoriadis’s thinking with respect to the vast literature and different research strands on institutional economics has not been undertaken so far. 3 Castoriadis on institutions Castoriadis’s vision is clearly distinct from the neoclassical institutional approach, which is premised on methodological individualism and analyses institutions using the tools of microeconomics, namely utility maximisation under constraints. The latter approach, which is usually termed neo-institutionalism,1 conceptualises institutions as efficiency outcomes, performing well-defined functions and shaped by rational individuals who react to price signals: ‘[neo-institutionalism] employs price theory as an essential part of the analysis of institutions; and sees changes in relative prices as a major force inducing change in institutions’ (North, 1992, p. 4). However, the Castoriadic thinking is also distinct from the old institutionalist analyses which ascribe a central position of economic change into collective units such as groups, firms, states, political parties, etc. Old institutionalism tends to downplay the role of human agency and of individuals’ rationality:2‘[the] social whole is seen as influencing and conditioning individual behaviour’ (Rutherford, 1994, ch. 3). Castoriadis strikes a middle ground between the predominance of the individual in the new institutional approach and the focus to social groups characterising old institutionalism. Instead, his primary agent is the ‘social individual’. The social individual, a concept originating from Aristotle,3 refers to the coexistence within each individual of the capacity for self-determination, supported by the ability to rationally pursue her aims, and the socialisation of the individual within a given historical and social context which affects her worldview and behaviour. Schematically, the new and old institutionalism represent the two extremes of Castoriadis’s synthetic view, with the individual and social components being predominant, correspondingly. In Toboso (2001), a middle way is also sought between the new and old institutionalism, the first being characterized by the adoption of methodological individualism, and the latter by methodological holism. This proposed middle way is termed ‘institutional individualism’, based on Agassi (1975), and is defined by three propositions: 1) Only persons can pursue aims and promote interests; 2) Formal and informal sets of institutional rules affecting interactions among persons must be part of the explanatory variables; and 3) Marginal institutional changes always result from the independent or collective actions of some persons and always take place within wider institutional frameworks. Old institutionalists like Commons (1934) and Veblen (1898, pp. 390–91) also noted the dual role of institutional structure and individual behaviour. The contribution by Castoriadis addresses the same theoretical question and seems to be clearly elaborated, offering a framework to think about creative responses in specific cultural and institutional contexts. Broadly, the analyses of Castoriadis could contribute to address the current limitations of neo-institutionalist thinking in its various strands which ‘tend to reduce institutions to functionalist consequences of efficiency considerations or instrumental reflections of interest’ and specifically move beyond the current ‘thin’ view of institutions toward a ‘thick’ view as regards the factors which shape and affect the operation of institutions (Chang and Evans, 2005, pp. 99–100). In this direction, Castoriadis explicitly questions what he labels ‘the functional-economic point of view’ (Castoriadis,  1998, p. 72) of institutions. He clarifies that he is not disputing the functionalist view inasmuch as it draws our attention to this obvious, yet crucial fact that institutions fill vital functions without which the existence of society is inconceivable. We are challenging it, however, inasmuch as it holds that institutions can be reduced to this and they are perfectly comprehensible on the basis of this role […] [W]e observe in given societies functions that ‘are not filled’ (although they could be, at the given level of historical development), with consequences that may be minor or catastrophic for the society in question. (Castoriadis,  1998, p. 116) and adds that ‘internal historical collapses of particular societies—Rome, Byzantium, etc.—provide counterexamples to the functionalist view’ (Castoriadis,  1998, p. 72). In general, Castoriadis places emphasis on such historical examples of institutions which were proven to be dysfunctional or incompatible with ‘rational’ functionality. In this light, it is obvious that the formation and change of institutions is multifaceted and cannot be captured effectively with microeconomic approaches based on maximisation. His approach offers a contrasting view to the ‘teleological’ method, especially favoured by neoclassical economists (David, 1994), with its strong underlying assumptions regarding the supposedly intended functionality of social arrangements. Castoriadis develops a generic approach for analysing institutions, which can be flexibly adapted to different sets of institutions, such as economic ones. His theory encompasses institutions in a wide sense of the term, including law, religion, rituals, practices and customs. Regarding the reason for the existence of institutions, he notes that institutions deal with the problems arising ‘from the collective existence of mankind’ and address the ‘necessity for arrangements and procedures that will permit discussion and choice’ (Castoriadis, 1975 , p. 72). For example, regarding the institution of the capitalist business enterprise, he notes that The capitalist business enterprise is […] [a] specific […] institution. There can be no capitalism without that enterprise […]. That institution conveys a signification, that set of arrangements and rules brings together large numbers of people, forces them to use specific tools and machines, controls their work and organizes it hierarchically, and its goal is its own unlimited self-aggrandizement. (Castoriadis,  2007, p. 100) Fundamental in Castoriadis’s analysis is the thesis that the ‘needs’,4 which the institutions aim to serve, are not pre-defined but are ‘created’—this ‘creation’ of needs is constrained by rationality and for coherence, and the past. More broadly, society defines its needs based on the prevailing worldview (‘social imaginary’, i.e. the image that individuals share about the world). As Palatano notes in relation to Castoriadis’s thought: ‘meanings [i.e. Castoriadis’ imaginary significations] have a structuring power: … they contribute to determine goals’ (Patalano, 2007, p. 233). Given his rejection of the functionalist view and his emphasis on the interrelation of needs and institutions, with the needs being non-constant but humanly created, Castoriadis’s analysis of institutions is inherently dynamic: A society can exist only if a series of functions are constantly performed … but it is not reduced to this, nor are its ways of dealing with its problems dictated to it once and for all by its ‘nature’. It invents and defines for itself new ways of responding to its needs as well as it comes up with new needs. (Castoriadis,  1998, p. 73) Therefore, needs and institutions both evolve, with constant feedback among each other. Furthermore, Castoriadis emphasises that needs for humans include to a large extent ‘representational pleasures’, which exceed ‘any simple reflection of the natural or objective world’. Therefore, human can find pleasures in making an object, in talking with others, in hearing a story or a song, in looking at a painting, in demonstrating a theorem or in acquiring knowledge—and also in learning that others have a ‘good opinion’ on him [or her] and even in thinking that he [or she] has acted well. (Castoriadis,  1998, p. 315) These representations are formed through the socialisation process, as will be elaborated in Section 3.1 below (see also Smith, 2014A, pp. 76–81). In contrast to the conceptualisation of institutions as a set of rules (formal or informal), which is typical of the neo-institutional school of thought, Castoriadis does not restrict his analysis to the ‘content’ of the institutions, but also considers the form with which the institution is expressed, what he calls the ‘symbolic’ component of institutions. His view is that the symbolic layer is not trivial and can have autonomous effects, also economic ones. Specifically, institutions contain a functional and a symbolic component, and therefore, each institution should be construed as a symbolic network in which a functional component and a symbolic component are combined. The mix of the two components could be in variable proportions and relations (Castoriadis,  1998, p. 83). Furthermore, Castoriadis argues that the functional view underestimates the importance of the symbolic component which is of fundamental importance for the operationalisation of institutions.5 The symbolic is not merely a neutral surface covering a ‘real’, ‘rational’ content. In the following characteristic excerpt, Castoriadis provides a vivid example on the dual nature, functional and symbolic, of a number of fundamental market institutions such as the wage earner’s salary, the banknotes and the salaried work: A paycheque is the symbol of the wage earner’s right to demand a given number of banknotes, which, in turn, are the symbol of their possessor’s right to perform a variety of acts of purchasing, each of which will be symbolic in its turn. The work itself which is the basis for the paycheque, although it is eminently real both for its subject and in its results, is, of course, constantly bound up with symbolic operations (in the mind of the person working, in the instructions he receives, etc.). And it becomes a symbol itself when, after being reduced to hours and minutes multiplied by given coefficients, it enters into the accounting office’s calculations of the paycheque or the company’s ‘operations account’, or when, in the event of disputes, it fills the empty squares in the premises and conclusions of the legal syllogism that will settle matters.’ (Castoriadis,  1998, p 73) According to Castoriadis, narrowing the scope of institutional analysis only to the (stated or purported) ‘aim’ of an institution (i.e. its ‘content’) potentially misses the impact of the operational dimension, which in turn is affected by the ‘form’ with which the institution is expressed, on the ensuing outcomes. The operational dimension in this view is not trivial as most institutional analyses implicitly assume, and ‘frictions’ or unintended consequences may feature because of the intermediation of the symbolic component on the expression of an institution’s content (e.g. because of the complexity of the ‘content’ which is aimed to be expressed). For example, in many instances institutions are ‘incomplete’ in the sense that their content is not unequivocal, i.e. functionality is not as concrete as the standard microeconomic framework usually assumes. The institution of religion is an obvious example of this type, but Castoriadis argues that even institutions much more directly grounded to the economic reality such as the law can be likewise characterised (see his analysis of Roman law in Section 3.2). These cases show that the symbolic component may determine to a large extent the economic outcomes of institutions by filling the space left by the functional component. Furthermore, it is sometimes non-trivial for the designers of an institution to master the symbolic component sufficiently in order to express effectively a desired social outcome, e.g. simply because of the complexity of the issue at hand. For example, the regulatory attempts in the last decades to limit banks’ risk-taking, which did not prove sufficient to prevent the recent global financial crisis, are a case in point. Therefore, the ‘intermediation’ of the symbolic component may have an effect on the functional dimension of the institution, i.e. the symbolic component may lead to suboptimal functionality by introducing a ‘friction’ inhibiting the desired functioning of the institution. Institutional change is triggered because of the discrepancy between the desired functions of the institutions and the given set of institutions at a given moment in time. In Castoriadis’s own words: There will always be a distance between society as instituting and what is, at every moment, instituted—and this distance is not something negative or deficient; it is one of the expressions of the creative nature of history. (Castoriadis,  1998, p. 71) Castoriadis uses the term ‘social-historical’ to refer to the intersection of (and tension between) ‘materialised’ institutions and the instituting society. The ‘anonymous collective whole’ is synonymous with the social-historical (Castoriadis, 1975, p. 108), as it refers to both of these elements (see e.g. Smith, 2010, pp. 130 ff). Castoriadis’s approach seems to be a suitable starting point for the formulation of a generic theory of institutional change, given the evolving nature of needs in his analysis and the emphasis on mutual feedback between institutions and rules. In addition, the existence of a symbolic component, besides the functional component, reinforces the potential for a discrepancy between stated and achieved aims of an institution, which in turn is a force behind institutional change. It will be argued that the salient features of Castoriadis’s approach to institutional theory, namely his identification of the social individual as the basic unit of analysis, his decomposition of institutions into a functional and a symbolic component (and the related emphasis on the ‘incomplete’ nature of institution) and, finally, his focus on the endogeneity of needs (which also may mean the ‘creation’ of needs) which the institutions serve, enable his framework to provide a theoretically sound alternative to the crucial methodological dilemmas separating old and new institutionalism (see also Toboso, 2001), specifically regarding the level of analysis (e.g. individual vs. social groups), the rationality assumption and the intended/designed versus unintended/evolutionary conceptualisation of institutional change. In the next section, Castoriadis’s understanding of human behaviour will be discussed and his critique of the methodology proposed by Max Weber for the social sciences. Subsequently, Section 3.2 will elaborate on the effects which originate from the presence of the symbolic component. These two sections will prepare the ground for the formulation of a schema for institutional change in Section 3.3. 3.1 Castoriadis on the individual The underpinning of Castoriadis’s analysis is that it is always the ‘social individual’ who lies at the basis of any theoretical explanation of social phenomena. The socialised individual has internalised norms and values which represent creations of the given society, i.e. the ‘anonymous collective’; however, she is also able to transcend the behaviour implied by them through her creative imagination, also utilising in this direction the capability of logical thinking. Thus the conception of the individual in Castoriadis’s work is distinctive and the theoretical framework is parsimonious while allowing at the same time for potentially rich individual behaviour. In more detail, Castoriadis’s individual has undergone a process of socialisation. This process begins as soon as the individual comes to the world and is mediated both by identifiable individuals (e.g. the mother) and by social institutions (e.g. the language). Socialisation introduces the individual to social significations, i.e. the worldview of the society she is living in. Two important questions arise if we attempt to incorporate the role of social significations into institutional analyses, specifically who is the creator of the social significations and whether social significations can be interpreted solely within the rationality framework. Castoriadis posits that the social significations are irreducible to ‘individual activity’ (Castoriadis,  1991, p. 54). The latter observation is especially important and is a distinctive insight of Castoriadis’s analysis: social significations are not created by individuals but only by socialised individuals. This point distances Castoriadis both from neo-institutionalists, who would explain social significations as a creation of rational individuals, and old institutionalists, who would probably tend to downplay the importance of human agency. For example, Castoriadis noted that: It would … be less false to say that homo economicus is a product of capitalist culture than to say that capitalist culture is a product of homo economicus. But we should say neither one nor the other. In each case there is a deep homology and correspondence between the structure of the personality and the content of the culture, and there is no sense in predetermining one by the other. (Castoriadis,  1998, p. 16) This interpretation of social significations provides a strong amplification mechanism, in the sense that the existence of an institution has feedback effects on the socialisation of ‘new’ individuals and their internalisation of the existing social significations, given that the new socialised individuals are the agents of institutional change. The existence of such an amplification mechanism is consistent with the wide diversity of institutional forms that are observed throughout history because of the strong lock-in effects that it implies. The use of the word ‘creation’ above, when referring to the origins of social significations and when characterising the individuals’ behaviour, is a salient feature of Castoriadis’s framework which merits to be mentioned here, because it is connected with the discussion about the rationality of individuals’ behaviour but also with understanding institutional change which follows later. Castoriadis notes that social significations ‘do not correspond to, or are not exhausted by references to “rational” or ““real” elements, and … are posed by a creation’ (Castoriadis, 1997, p. 8). The origins of ‘creation’ in Castoriadis do not lie exclusively to an internal impulse, as is the case in economists like Schumpeter (see e.g. Vouldis et al., 2011), but it arises as a result of the ‘encounter’ of the individual with the social environment, i.e. it is driven by this encounter specifically and not solely by the individual. Smith (2014A, p. 83) clarifies that this encounter ‘is neither a simple external imposition, nor merely an autopoietic expression of an innate organisation’. One could qualify the characterisation of ‘creation’ to the class of cases where modifications to existing social significations occur or new ones are created. On the contrary, in most cases the individuals’ encounter with the already instituted social significations would lead via a process of internalisation to idiosyncratic adaptations of existing instituted significations for each individual, without pronounced effects at the societal level. Two implications stem from this view on ‘creation’. First, the process of change does not originate solely from the individual, nor it is a result of impersonal forces at the societal level. It is a product of two mutually irreducible processes, namely the encounter of the individual with the meanings and significations of the society and the ‘reaction’ of the individual. For Castoriadis it is precisely the individual’s psyche which ‘must institute meanings for itself with reference to … social institutions’ (Smith, 2014A, p. 83; see also pp. 78–84). Second, regarding rationality, this view on creation implies that rationality itself is created via this ‘encounter’, therefore ‘the criteria of functional utility depend on cultural interpretations’ (Arnason, 2014, p. 49). For Castoriadis, humans are characterised by this capacity of creating new meanings, forms, norms through their imagination. ‘Imaginary’, Castoriadis explains (Castoriadis,  1998, pp. 127–31), is something ‘invented’, either in the sense of an ‘invention’ or the investment of a symbol with other significations than their canonical ones. Examples of the former case are the conception of ‘God’ as the creator and founder of the Law in monotheistic religions and the idea of the ‘nation’, binding together people following the same way of life, speaking a common language and living under specific institutions. An example of the latter case is the national flag, which has a clear rational function as a sign of recognition in certain settings but which can also become what one can die for. Imaginary social significations animate social and economic life and institute ‘reality’ as faced by the individual. Gold as money and machines as capital are examples of this animation of objects through social significations: To say that gold in itself is not money may appear superficially as trite but it immediately leads to the question of the institution of society and of this institution as essentially historical. For gold to become money, it is not enough that it possesses the ‘natural’ qualities listed in the manuals of political economy which ‘predestine’ it for this role; it requires that particular social-historical development which, starting from embryonic forms of exchange, leads to the institution of a ‘general equivalent’ (this, anyway, is Marx’s conception, which we shall not discuss here for itself). For a machine to become capital, it has to be placed within the network of socio-economic relations which capitalism institutes. (Castoriadis,  1998, p. 223) This imaginary capacity is expressed both through the symbolic component and the functional component of the institutions (see also Section 3.2). In addition, according to Castoriadis, the social significations cannot be reduced to ‘rationality’, whatever breadth one grants to the meaning of that term. The irreducibility of social significations to rationality can be validated, e.g., if one considers the plethora of civilizations which existed at the same time, e.g. in the fourteenth and fifteenth centuries: Aztecs, Incas, Chinese, Japanese, Mongols, Hindus, Persians, Arabs, Byzantines and Western Europeans plus a number of other cultures on the African, Australian Asian, and American continents (Castoriadis,  1991, p. 54). Castoriadis argues that this diversity of civilisations cannot be construed as reflecting different ‘figures of rationality’ and clearly proves the theoretical inadequacy of rationalist theories. Consequently, the analysis of Castoriadis avoids both the reduction to methodological individualism and collectivism. The individual is shaped by the socialisation process but also attains the potential for behaviour which transcends the type of behaviour which could be attributed to the influences from the anonymous collective, through values and norms. This is possible by the utilisation of the individual’s creative imagination (see Arnason, 2014). The existing social imagery serves as the raw material for the individuals’ creative behaviour. Rational calculation, e.g. pertaining to some form of utility maximisation, is only a part of the individual’s creative capacity and presupposes the adoption (or creation) of a value system which supports specific form of ‘rationality’. As a result, the neoclassical microeconomic behavioural assumption is a special case in the Castoriadic framework, corresponding to an individual who, e.g., is placed in a setting with minimal institutional features (usually only property rights and existence of a market), facing a fully transparent optimisation problem, where she is trying to maximise her (usually monetary) gains, disregarding usually other considerations, e.g. reputation, social group pressure, uncertainty of various types, complexity, etc. However, as Hodgson (1998) notes, theories of pricing which are cross-fertilised by disciplines such as psychology and marketing, consistent with the tradition of old institutionalism, could be more promising than research directions which follow the micro-economic neoclassical framework. When one embarks on an analysis of the formation and change of economic or political institutions, the suitability of the standard neoclassical assumptions may be a fortiori questioned. Castoriadis distinguishes two different modes of the individual’s relationship with the ‘input’ she receives from the social environment. Specifically, he draws a sharp distinction between the ‘autonomous’ and the ‘heteronomous’ individual. These two modes of being, which can also refer to whole societies and the collectively established relationships with their institutions, are distinguished by the existence of self-reflectivity between the individual/society and the institutions. Specifically, autonomy presupposes the possibility to interrogate the logical underpinnings of social significations and institutions, their effect on the individuals’ behaviour and the explicit realisation that institutions and their supporting significations are open to assessment and could be subject to alterations. For example, this type of relationship with the institutions emerged, according to Castoriadis, from the European Enlightenment (Smith, 2014B, p. 185) and in ancient Athens (Straume, 2014, pp. 191–202). See also Castoriadis ( 1997) for an elaboration on the preconditions of autonomy with reference to three distinct different spheres of social life, namely, private sphere (oikos), the public/private sphere (agora) and the public/public sphere (ekklesia in a democracy). Smith (2010, pp. 122 ff) and Adams (2014, pp. 10–11) question the sharp polarisation between autonomy and heteronomy, which Castoriadis stresses. Smith (op. cit.) favours the idea that a continuum between these two extremes is both more common historically (e.g. the citizens of a democratic polis did not emerge through a rupture) and potentially more desirable (in the sense that new ideas and institutions built upon existing social significations can more easily flourish compared to ex nihilo creations which are totally disconnected from the existing significations). The two aforementioned points, i.e. the identification of social individuals (rather than rational individuals) as the designers of institutions and the irreducibility of social significations to rationality, seem to be pertinent theoretical contributions in the understanding of institutional change, consistent with the diversity of institutional forms which are observed throughout history and the persistence of institutional forms. Further, on the issue of rationality, it is important to present Castoriadis’s critique on the methodological individualism of Max Weber, as it is exposed in the article ‘Individual, society, rationality, history’ ( 1991), where Castoriadis reviews critically the theses supported by Max Weber as regards the rationality of humans in the first methodological chapter (‘Methodological foundations’) of Economy and Society (1922). Castoriadis stresses in this instance that the limits of the instrumental rationality assumption are not only due to ‘bounded’ rationality, i.e. the inability of humans to possess all the available information, but also, and equally importantly, on the fact that the rational elements of human behaviour are unavoidably immersed in a universe of social significations outside which they would make no sense. Weber builds his methodological approach on two fundamental premises: On the one hand, he defines as the task of social scientists to understand, through a process of ‘sympathetic reliving’ (sympathisches Nacherleben), the behaviour and motivations of individuals within a given social context. On the other hand, Weber rejects any inference from his methodological rationalism about the predominance of rational motives in human action or even a positive evaluation of rationalism.6 Castoriadis argues that while developing his thought, Weber falls into an antinomy. This is important because it does not concern only Weber but also a number of other social scientists who feel obliged to ‘close’ their theoretical systems by positing clear causal structures. The antinomy is expressed by the gradual shift of Weber’s position into a reduction of the understanding, aimed at by the social scientist, to the understanding of instrumentally rational action. Weber acknowledges that instrumentally rational behaviour is a marginal case in human history and behaviour has been driven historically primarily by impulses or habit. Despite that, Weber emphasizes that social sciences should aim at furnishing a causal interpretation of human actions which, in turn, require an actual or ‘effective’ (tatsächlich) meaning underpinning them (i.e. ‘meaningful’ behaviour: sinnhaftes Handeln). Then one can argue that ‘insofar as the action was rigorously rational in an instrumental way, it would have had to (müsste, in the sense of necessity and not obligation) occur in this way and no other’ (Castoriadis,  1991, p. 43, italics in original). Therefore, Weber connects consecutively ‘understanding’ with ‘causality’ and consequently ‘causality’ with a rationally intelligible representation, thus attributing a de facto dominant position to the instrumental rational dimension, despite the lack of empirical evidence in support of this methodological choice. Castoriadis interprets this connection of understanding with the rational dimension of human action to Weber’s Kantian and neo-Kantian roots. Castoriadis considers Weber’s approach unsatisfactory as it allows an unresolved question to remain, i.e. why human behaviour is driven to such an extent by habits (i.e. cases where there is a shortage of meaning: Fehlen an Sinnadäquanz), and what does this tell us about the fundamental determinants of human behaviour? Castoriadis argues that in a large number of crucial cases humans behave in a way that is not fully intelligible to them. For example, humans may ‘understand’ an institution (in the sense of being able to contribute to its development or simply participate in it) but they are not able to ‘redo’ (reproduce) it; in the sense of being able to reproduce, mentally, the generic form of the institution or to really comprehend its functioning within society. For example, a person may be able to invent a particular law but not the idea of a social law (the idea of the institution). Therefore, there are certain limits to the intelligibility and the understanding that humans possess over the institutions of the society in which they live in. This argument is related to the ‘bounded rationality’ conception of human behaviour, however, Castoriadis transfuses this concept with a new perspective: boundedness of human cognition does not refer solely to how humans perceive external information (e.g. prices and other data relevant for taking economic decisions) but also on the degree to which they can comprehend their institutional environment, e.g. the concept of a law is understood to different degrees among different cultures and this represents also an expression of the bounded nature of human cognitive capabilities. Another example offered by Castoriadis concerns language, a fundamental institution with critical economic consequences. The significance of language as an institution with economic consequences is especially relevant for a theory which stresses the symbolic/operational component of institutions, such as that proposed by Castoriadis. Specifically, Castoriadis argues that language was not consciously developed by specific humans pursuing their ‘interests’ or ‘ideas’. In addition, the cognitive process of participating in this institution, i.e. the process of learning the language, is not structured in a way which would be compatible with the notion of an individual characterised by instrumental rationality but rather with that of individuals following rules of thumb and adapting their behaviour. Therefore, both the creation of and the participation of humans in the fundamental institution of language evade the instrumental rationality framework and can only be explained within the framework of a rules-of-thumb adaptive behaviour. In addition, language is, similarly to law, an example of an institution which cannot be ‘reproduced’ logically by the individuals in the sense of being able to fully ‘understand’ the institution and create it anew. In this vein, Weber (1978, p. 105) has pointed out that there are rules that are followed ‘without any subjective formulation in thought of the rule’. Consequently, it is impossible to defend the conception of a ‘rational’ and ‘conscious’ individual having a logical control over its created institutions. Hodgson (2015, p. 61) also notes that ‘few of us could specify fully the grammatical rules of the language that we use regularly or completely specify in detail some of our practical skills. We rely on learned habits’. See also the discussion in Gordon (1989) regarding the origins of language with respect to Hayek’s theory and the notion of ‘unintended consequence of human action’. Such examples show that the theoretical edifice which is based on the postulate of the rational approach of individuals proves to be inadequate to explain fundamental facts about human societies. Castoriadis is certainly in line with Weber when the latter explicitly refuses to ask questions of the kind: ‘Is it the individual or the society that comes first?’ What is also common is their stated aim, to understand the behaviour of the actual (or ideal-typical, as defined by Weber) individual. Castoriadis distances himself from Weber’s postulation of instrumental rationality as the horizon of intelligibility for the socio-historical change. For him, the rational elements of ‘instrumental rationality’ are always immersed in the universe of the prevailing social significations, the latter being the elements without which the rational elements would have no meaning. In contrast, Weber considers that the empathic reliving (einfühlend Nacherleben) which forms, according to him, the basis for understanding, is very difficult for the cases when the ultimate ‘ends’ or ‘values’ of a given society depart from our own ultimate values. Therefore, he narrows the range of analysis to the instrumentally rational component7—a solution which Castoriadis considers unsatisfactory as it leaves out the most fundamental drivers of human behaviour, specifically, the instituted social significations. 3.2 Symbolic component of institutions Castoriadis’s insight about the existence of an additional component of institutions, besides the functional one, could be utilised constructively in the field of Economics because it focuses the researcher’s attention on the need for the institution to be expressed, using pre-existing concepts and forms, and operationalised. In addition, it stresses the ‘incomplete’ nature of many institutions, an aspect which has not been sufficiently stressed in institutional analysis. The symbolic component presupposes the imaginary, i.e. the capability to posit a permanent connection between two elements. Castoriadis explained as follows how he arrived at the idea that the radical imaginary is critical to understand human behaviour and concomitantly socio-historical change: The discovery that the human psychism cannot be ‘explained’ by biological factors or considered as a logical automaton of no matter what richness and complexity and, also and especially, that society cannot be reduced to any rational-functional determinanations whatsoever (for example, economic/productive, or ‘sexual’, in a narrow view of the ‘sexual’) indicated that one had to think something else and to think otherwise in order to be able to comprehend the nature and specific mode of being of these two domains, the psychical on the one hand, the social-historical on the other. (Castoriadis,  1997, pp. 290–91) Specifically, the symbolic component of an institution is a network of symbols: [E]ach institution constitutes a particular symbolic network. A given economic organization, a system of law, an instituted power structure, a religion—all exist socially as sanctioned symbolic systems. These systems consist in relating symbols (signifiers) to signified (representations, orders, commands or inducements to do or not to do something, consequences for actions—significations in the loosest sense of the term) …. A property title, a bill of sale is a symbol of the socially approved ‘right’ of the owner to undertake an unlimited number of operations with respect to the object of his ownership. (Castoriadis,  1998, p. 73) This symbolic network cannot, in many cases, be reduced to corresponding functions. The symbolic component is on the one hand the means by which the institution can be operationalised and perform its function, but on the other hand, and in many cases, the symbolic layer is not a trivial, intermediate layer and can even determine the actual functionality of the institutions. In general, the symbolic component covers the space which goes beyond pure functionality and may override to some extent the functional component. E.g. the institution of a seven-day week has religious origins but one could doubt whether this symbolic expression is the most functional arrangement from an economic viewpoint in the context of the capitalist system within which output maximisation would be prioritized over following a religious tradition (e.g. why not change the week into a ten-day week?). In addition, the effective use of the symbolic layer in order to achieve the desired functionality is not straightforward in many cases. In the latter case, the handling of the symbolic component as a medium of expression of the institutions’ functionality is non-trivial. E.g. Castoriadis discusses the symbolic component in the context of religion and notes the impossibility of reducing all religious symbols into corresponding functions: ‘the “choice” of a symbol is never either absolutely inevitable, or merely haphazard. A symbol never imposes itself with a natural necessity, but neither does it ever lack all reference to reality’ (Castoriadis,  1998, p. 74). His aim was not to reject attempts, e.g. by anthropologists, to explain rituals (e.g. see the discussion on rituals in Mayhew, 1987, p. 592) but to state that there cannot be a full determination of all components from functionality consideration. Admittedly religion is a very specific institution. However, Castoriadis extends this type of analysis for the institution of law, which is tightly knit with economic relations and one would expect that the symbolic component is of little significance. Law is ‘directly related to the “substance” of society, which is … the economy and the real and concrete social relations that are expressed in property, transactions and contracts. In law, one should be able to show that symbolism is in the service of content and can be otherwise only to the extent that rationality forces it’ (Castoriadis,  1998, p. 75). According to the functionalist view of institutions, one would expect that the mapping from the symbolic component to the supposed functionality would be straightforward: ‘Thus in the institution that remains an eternal monument to rationality, to economy and to functionality, the institutional equivalent of Euclidian geometry, I mean Roman law, over a period of ten centuries extending from the Lex Duodecim Tabularum to the Justinian code, this veritable, though well-ordered and well-trimmed forest of rules serving property, transactions and contracts was developed’ (Castoriadis,  1998, p. 75). However, Castoriadis notes that in contrast to what a functionalist view would predict, the symbolic layer had a significant impact on the operation of the Roman law. In particular, Castoriadis argues that its formalism seems to have been independent from the functionality and that the process of improving this functionality was very slow: ‘…the dominant characteristic of its evolution over ten centuries … makes it a fascinating example of the type of relations between the institution and the “underlying social reality”: this evolution has been a long attempt, precisely, to attain this functionality starting from a state that was far from possessing any such thing’ (Castoriadis,  1998, p. 75). In the case of the Roman law, the symbolic component was not in harmony with functionality and the process of attaining an effective expression of a given functionality extended throughout centuries: At the start Roman law is a rough set of rigid rules, in which the form overwhelms the content to a degree that far exceeds whatever would be justified by the requirements of any set of laws as a formal system…To cite only one example … the will and the intentions of the parties entering into an agreement, which is the functional core of any transaction, plays for a long time only a minor role with respect to the law; what predominates is the ritual of the transaction, the fact that certain words were uttered, certain gestures made. Only gradually is it admitted that the ritual can have legal effects only inasmuch as these are intended by the actual will of the parties concerned. However, the symmetrical corollary to this proposition, namely that the will of the parties can constitute obligations independently of the form its expression may take, the principle that is the very foundation of modern law of obligations and that truly expresses its functional character: pacta sunt servanda, this will never be acknowledged. The lesson of Roman law, considered in its real historical evolution, is not the functional character of the law but the relative independence of formalism or of symbolism with respect to functionality at the outset, followed by the slow and never complete conquest of symbolism by functionality. (Castoriadis,  1998, p. 75) This is in contrast to the case of religion, where the symbolic component is present but without being reduced to any specific functionality; in the case of law, mastering the symbolic component to achieve the functionality of the institution proved to be a formidable task and the progress in this respect was gradual. One corollary from such examples is that the symbolic component of institutions, given its degree of independence, can have important implications with respect to institutional change. Therefore, social evolution is bound by symbolism. Symbolism ‘can draw society into one of several different directions left undetermined by functionality, or even create effects that have a rebound effect on the latter (the stock market represents, in relation to industrial capitalism, essentially this sort of case)’ (Castoriadis,  1998, p. 77). The insight of Castoriadis regarding the challenges of mastering the symbolic component of law in order to attain functionality is related to subsequent research. For example, Benson (1989) analyses in detail how the dysfunctional character of the common law in England was to a large extent the cause for the development of the Law Merchant as a separate legal system with adjudication procedures and penalties enforced by the merchants themselves—rather than by the royal courts. According to this interpretation, the inadequate mastering of formalism was a trigger for institutional change in the form of a parallel legal system. In addition, Castoriadis’s insight is relevant for assessing the research strand which, in the context of institutional theories of growth, links legal origins and growth, through the scope offered to the individuals for innovation, as summarized e.g. in La Porta et al. (2008). Specifically, Castoriadis offers an interpretive framework which can be used to challenge the claim of a causal link from the legal origin to economic efficiency. Specifically, one could argue that the handling of the symbolic component of the law is not trivial and affects crucially the operation of the institution in practice. This effect may dominate the impact originating from the ‘content’ of the law. For example, Berkowitz et al. (2003) has argued, in a similar vein, that the implementation of the law in each country is what matters for the economic results, rather than the legal origin, i.e. the ‘content’, of the law. The potential for a wide discrepancy between the formalism of an institution and the objective reality because of the difficulty in mastering the symbolic component of institution possibility seems to be very relevant for e.g. developing countries whereby legal frameworks may not function effectively not only because of lack of enforcement, as is usually emphasized by neo-institutionalists, but also because the mastering of the formalism is inadequate. The use of the concept of symbolic with reference to institutions is relatively rare among economic theorists.8 The need for the operationalisation and implementation of institutions has not been explicitly incorporated in institutionalist theories. However, effects determining the outcomes of institutional set-ups can be attributed to this component. Castoriadis points to the need for a given institutional content to be expressed using the existing symbolic toolkit of a society and emphasizes that this layer may determine the institutional outcome. Overall, Castoriadis proposes a conceptualisation of the symbolic component which can be integrated effectively into a theory of institutions and which possess explanatory power over historical experiences. 3.3 A schema of institutional change Castoriadis envisages a schema for institutional change based on his conception of the individual and the institutions presented above. This schema could be represented as consisting of the following steps: Initial state. Let us assume an initial constellation of a given society/individuals/social actors and existing preferences/worldviews (‘significations’) corresponding to a definition/understanding of needs and wants. Existing institutions are directed towards the satisfaction of these needs and are expressed through the symbolic component, which possess elements irreducible to functionality. The initial state plus all subsequent steps lean on the ‘first natural stratum’ which encompasses geography, climate, biological needs and technological capacity but are not determined by it. Institutional change is triggered. Two cases can be distinguished here: a. Institutional change triggered by functionality gap. When there is a gap between needs (of social groups which have the power to alter the institutional setting) and the functionality of existing institutions, there is institutional change. This case would lead then to Step 3. b. Institutional change triggered by new social significations. Institutional change can also occur due to the creation of new social significations. Such a development would lead back to Step 1 and a new definition of needs which could trigger a gap in functionality of the existing institutions. Institutional change takes place aiming to restore functionality. Institutions are created attempting to fulfill the defined needs (of the initiating social groups), drawing on the existing symbolic toolkit and constrained by the set of current social significations and institutions. Outcome may or may not close the gap between institutions and intended functionality. There is no guarantee that the resulting institutional framework will correspond to what is ‘optimal’ given society’s needs. This may be the case e.g. due to inefficient handling of the symbolic component or ensuing unanticipated consequences of the resulting institutions. Feedback from institutions to significations. The resulting institutions, in turn, are constitutive of the preferences/worldviews of the individuals (i.e. affect their ‘significations’). Go to Step 1. The above schema of institutional change9 is rich and can be considered as containing a number of current institutional theoretical strands as special cases. For example, most of the work under the heading of neo-institutional economics (as represented e.g. by Douglass North and Oliver Williamson10) is based on a simpler version of Steps 1–3 (whereby the role of worldviews is more trivially represented by purely economic interests). In the neo-institutionalist approach, the possibility of non-functional institutions, i.e. Step 4, is not very well grounded theoretically—although it is admitted that it can occur in reality—while Step 5 is not considered. The emphasis on creation, as a result of the socialisation process and the encounter between the socialised individual and the instituted society, underlines the fundamental fact that ‘institutions … are patterns of action and thought, imposed on the individuals but also subject to changes in the course of historical events’ (Arnason, 2014, p. 103), i.e. that change results from the encounter of socialised individuals with the existing institutional order. Therefore, Castoriadis posits a framework for understanding change which takes into account the relationship individual-society rather than focusing on a single pole of this relationship. Therefore, regarding Step 1, an important feature is that the needs of society do not correspond to some objective economic calculus, but are conditioned on the prevailing worldviews, which in turn depend on past choices, institutions, symbolic and imaginary layers and history. In this manner, an explicit amplification and path-dependence mechanism is introduced into the schema of institutional development. Castoriadis stresses that economic considerations are not the only dimension guiding the individuals in their decisions. Economic analysis usually presupposes an invariable type of basic motivation for all individuals, broadly speaking, an economic motivation: from all time, human societies are held to have aimed (whether consciously or unconsciously, little matter) first and foremost to increase their production and their consumption. But this idea is not simply false in a material sense; it overlooks the fact that the types of motivation (and the corresponding values that polarize and direct human lives) are social creations, that each culture establishes its own values and rears individuals in relation to these. (Castoriadis,  1998, p. 14) Humans’ needs can refer directly to imaginary social significations, as explained in Section 3.1, e.g. when the Tiv of Southeastern Nigeria reject any actions or institution-building that could lead to tsav, i.e. concentration of power by any groups or individuals given their signification of tsav as related to death, even if such institution-building could improve economic prospects (Acemoglu and Robinson, 2016). The fact that we assume a given environment of significations and institutions corresponds to the ‘substantive’ approach. An aspect of this step, which is in need of further elaboration, is the specification of the agents of institutional change. Castoriadis does not offer a formalisation on this issue, although his work is very much concerned with the analysis of antitheses of social groups, especially in the context of modern capitalism (see Castoriadis,  1998, pp. 91 ff, for the problem of the emergence of classes in human societies). Step 2, case (a), is common also in the functionalist view: institutional change takes place in order to address a discrepancy between needs and existing institutions. Other strands of institutional research also stress this point. For example, in the context of the ‘varieties of capitalism’ approach, Hall and Thelen (2009, p. 15) note that because political economies are full of entrepreneurial actors interested in improving their positions, existing institutions are bound to come under pressure. Institutional equilibria change as developments shift the material situation, power and self-understanding of the actors. In reality, institutional change is polymorphous and may manifest itself either as gradual institutional change, brought about by political bodies or in an ad hoc way, or as ruptures, brought about by political authorities or contesting parts of the population. An example of this case is the development of the Amsterdam securities market around 1600, triggered by the need to finance long-distance trade in Asia. Expeditions to Asia changed the demand for capital in three key respects compared to the previous trading routes e.g. in the Baltic Sea: i) bigger scale of operations requiring higher amounts; ii) increased risk; and iii) longer time horizon to break even. These new challenges triggered institutional developments in different directions. First, business organisation forms were adapted to accommodate the need to mobilise more resources e.g. newly created associations of merchants run by a committee of directors (bewinthebbers) instead of simple associations of fewer merchants which was the case before. In addition, a secondary market for shares was created so that participants could liquidate their claims in case they faced liquidity needs (see Gelderblom and Jonker, 2004). In this case, a gap in functionality which arose due to the need to finance long-distance enterprises to Asia triggered institutional developments which covered the missing functionality. In this case, the functional component was clearly defined and ‘leaned on’ but not determined by existing significations like the enterprise as an anonymous entity, which mobilises people and capital to generate profit flows, and the homo economicus mentality of rational calculation prevalent in the Dutch Republic, which excluded extra-economic, e.g. religious, considerations from the economic sphere (see Greenfeld, 2001, pp. 102–4). In addition, a central theme for Castoriadis is the potential for the creation of new social significations11 and concomitant aims and needs by the individuals; these aims may not be purely economic, and this is captured by Step 2, case (b). The intellectual revolution of the ‘Industrial Enlightenment’ represents an example of how the creation of new social significations triggered institutional change. The notion of understanding nature and controlling her to attain material progress, a pivotal proponent of which was Bacon, was spread from Enlightenment intellectuals, institutionalised into organisations such as the Royal Society, and then spread into society, facilitated by technological features of the era, such as printing, mail, railways and other means of transportation, which reduced access costs. These developments were pivotal behind the industrial revolution. This is a case where the creation of a new social signification, that of material progress achieved by the systematic application of ‘useful knowledge’, guided individuals’ behaviour and spurred institutional change: the first free public library in Britain was founded in 1653, scientific and scholarly societies were founded in the course of the eighteenth century, international markets for industrial instruments emerged, work on universal terms and standards was undertaken and a number of institutions of formal education like universities and mechanics institutes were established (Mokyr, 2005). In that case, the new signification was aiming directly at improving material conditions. In addition, the advances in natural philosophy and the emerging social signification of an open-ended progress achieved through the use of analytic-synthetic methods were also critical in influencing the thinking of ‘moral philosophers’ like Adam Smith (see e.g. Montes, 2008, with respect to the Newtonian influence) and were also underpinning changes in institutions like the extension of education and political rights. Furthermore, ‘gods, or God with a capital G, or the polis for the ancient Greeks, or citizenship, or the nation’ (Castoriadis,  1982, p. 47) are imaginary significations that played a substantial role in institutional change and had sometimes far-reaching economic effects without being directly connected with economic aims (see e.g. Greenfeld, 2001). Step 3 represents the creation of new institutions. The inputs used for the creation of the new institution are defined in this step. A distinctive element of Castoriadis’s thought is his emphasis in the scope for institutional change that is perceived as possible by different societies. First, the new institution is determined by the current institutions, leading to a path-dependence effect, a point which is already made in the evolutionary economics literature. Other institutional theories usually assume implicitly that the scope for institutional change is only limited by past institutional choices. Castoriadis posits that the prevalent social significations, i.e. the worldviews of society, affect the feasibility set of institutions and therefore institutional change is further constrained, besides the constraints imposed by institutional path dependence.12 The effect of the latter path-dependence mechanism further reinforces the overall lock-in effects.13 Castoriadis involves the symbolic component in the set of inputs which are used to bring about a change in institutions. This is a further mechanism which constrains the feasibility set of the new institutions. Castoriadis rejects strict causality and determinism as regards institutional change. Causality is substituted with the idea of Anlehnung, i.e. leaning on, which does not imply the existence of ‘laws’. E.g. ‘the institutions of capitalism may be thought of as leaning on the economic and political—and also religious, as the analyses of Weber, Durkheim, Mauss and others suggest—institutions of preceding societies’ (Klooger, 2014, p. 132) rather than being determined by them. Step 4 differs from traditional functional analyses by positing that the ensuing institutions could prove not to be functional and this is not a rare exception. If an institution is opposite to the functional aspect, it could collapse immediately; Castoriadis mentions Law’s paper money as an example of a non-functional institution destined to collapse. However, it is common in history that institutions which fail the test of optimality have been devised and persisted. Instances of financial crises can be used to provide examples where the rational element of institutions could be adverse to the functional component. This seems to be a challenge to the rational approach to institutional change and an indication that this approach is not corroborated by the historical experience. The view that ensuing institutions may not be functional also distances Castoriadis from the ‘cultural efficiency’ approach which posits that given culture and preferences the optimal institutional set-up will be reached (see Ogilvie, 2007). Castoriadis’s conceptualisation of human behaviour recognises the cognitive limitations of humans but should not be identified with the ‘bounded’ rationality assumption (see Section 3.1). Fernández-Huerga (2008, p. 720) elaborates on the differences between the ‘bounded rationality’ view and non-neoclassical conceptions of human behaviour. In contrast to the latter approach, Castoriadis stresses that the rational elements of behaviour are inextricably immersed in the universe of social significations outside of which they have no meaning. Given these cognitive limitations, the economic non-optimality of institutional set-ups become much more natural than one is led to believe if the neoclassical assumptions are adopted. The structure of institutions, e.g. the interplay between the functional and the symbolic components, may be another reason for failed or non-optimal institutions. For example, the inefficient handling of the symbolic component may prevent the attainment of the desired functionality, as is the case with the Roman law presented before. In such a case, the discrepancy between the institution and the functional aspect can be partially resolved, and the desired functionality be attained, only after a long period of time. In addition, an institutional change may trigger unanticipated dynamics, which may turn against the social groups which pressed for the change (the latter effect is analysed by Hall and Thelen, 2009, pp. 15–16). Step 5 corresponds to a recent strand in economic research, e.g. in the ‘endogenous preferences’ literature (e.g. Bowles, 1998). Bowles notes that ‘most writers have implicitly invoked a kind of functionalist correspondence between economic structures on the one hand and values, customs, and tastes on the other, without explaining the mechanisms by which the former might affect the latter’ (Bowles, 1998, p 76) and reviews a number of transmission channels from institutions to preferences.14 Castoriadis adds to the ‘endogenous preferences’ research strand the idea that it is important for a society to reflect on the effects of its institutions to the preferences of the individuals, otherwise an ‘autonomisation’ of the institution occurs and the society “lives its relations with its institutions in the mode of the imaginary, in other words … it does not recognize in the imaginary of institutions something that is its own product” (Castoriadis,  1998, p. 82). The strand of economic literature which is more closely related to Castoriadis’s ‘thick’ view of institutions is the type of institutional analysis which incorporates cultural and symbolic aspects. This approach does not represent a unified body of thought and there are a number of authors working in the fields of Economics, Sociology and Political Science who independently from each other develop forms of institutional analysis which aim to go beyond the neo-institutionalism and the latter’s adoption of neoclassical framework and scope of analysis. E.g. Chang and Evans (2005) note that a theory of culture is a tall order, but ‘introducing even a primitive theoretical consideration of ideology and worldview into the discussion of institutional change allows for qualitative improvement over simple efficiency and interest-based theories’.15 They analyse two cases in which the enlargement of the scope of factors driving institutional changes enhances the interpretative power of institutional analysis, namely the process which led to the creation of the Korean developmental state and the role of US-trained Korean economists in its dismantling, and the interaction between the ideologies lying behind the institution of the World Trade Organization and the scope of its work, both on the side of the ‘elites’ and the workers. In addition, a similar direction is proposed in Mantzavinos et al. (2004), whereby a distinction is made between the external and internal aspect of institutions, the former corresponding to the typical rational choice approach and the latter to what they term a ‘cognitive approach to institutions’. In contrast to the rational choice approach, cognitive institutionalism does not judge the rationality of the individuals’ objectives, given the influence of beliefs in defining aims. Mantzavinos et al. suggest that the process of societal change evolves according to the schema ‘reality’ > beliefs > institutions > specific policies > outcomes (and, thus, altered ‘reality’); therefore, the interpretation of ‘reality’ both individually and collectively is a crucial step which should not be considered as an objective invariable function. Consequently, they posit that cognitive path dependence causes path-dependent institutional change (‘increasing returns of an institutional framework’) and consequently economic path dependence. This framework of analysing institutional change is close to Castoriadis’s schema presented above given the endogenous character of the aims.16 Overall, the schema presented above seems to represent a starting point for further elaboration and for the development of a theory of institutions and institutional development which would be both formal but also richer and closer to the complexity of reality than the mainstream current of the neo-institutional approach.17 4 Conclusions Castoriadis’s thinking on institutions contributes significant insights that could enhance the development of an institutional body of theory located beyond neo-institutionalism. The paper defends the thesis that Castoriadis’s contribution elaborates constructively on issues which have long been central themes of non-neoclassical approaches to institutions, such as the cognitive limitations of humans and the role of the social environment. Castoriadis’s theory explicitly accounts for the observed worldwide diversity of institutions, the year-long prevalence of inefficient institutions and the collapse of historical societies. The emerging literatures on ‘endogenous preferences’, ‘varieties of capitalism’ and those examining the role of ideology and culture could be smoothly incorporated in his proposed framework. Strong amplification and path-dependence effects arise naturally in Castoriadis’s theory in a consistent manner. In addition, this paper has proposed a tentative formal schema for institutional change, based on Castoriadis’s approach, which could be extended in various directions and could structure the discussion of such alternative theoretical approaches. Specifically, Castoriadis proposes a decomposition of institutions into a ‘functional’ and ‘symbolic’ component, whereby the latter is the form with which the institution is expressed and enables its operationalisation. Castoriadis emphasises the importance of the symbolic component and notes the critical impact it has on the actual functioning of an institution and on its economic outcomes. This view is related to his conception of institutions as ‘incomplete’ with respect to their content and seems to offer an interpretative framework for many real-world phenomena. The implication of his analysis is that the symbolic component may either represent a ‘friction’ in the optimal functioning of the institution (e.g. in cases where achieving the aim of an institution is inhibited by the complexity of the issue at hand, like e.g. in the case of regulating the banking sector) or it could also dominate the functional component, determining the outcomes (e.g. the seven-day week, which is not optimal from a capitalistic profit-maximisation perspective). Moreover, Castoriadis offers a description of human behaviour which is based on the concept of the ‘social individual’, thus distancing himself both from purely individualistic theories and from those which ascribe autonomous importance to collective entities. For Castoriadis, the individual is socialised from the first instance that she comes into the world and is exposed to the prevailing universe of social significations, i.e. the worldviews of the society. These social significations animate the rational elements of her behaviour; it is impossible to separate the latter from the former. The critical presentation of Castoriadis on Max Weber’s methodological section in the latter’s Economy and Society (1922) exemplifies how Weber is led to exalt the importance of instrumental rationality as the interpretative horizon of the social sciences by downplaying the inextricable link between rationality and social significations. Finally, a schema of institutional change is presented in this paper based on Castoriadis’s analyses. Given that alternatives to neo-institutionalist theory posit a much more complex view of human agency and institutional development, it is important to structure the discussion by codifying the alternative theoretical approaches. Existing theoretical approaches could be considered to represent subsets of the proposed schema of institutional change. The schema could be fleshed out in all of its discrete stages and could be extended in different directions. For example, the version proposed here does not include the effect of internal conflicts e.g. between different groups within a social formation. Further elaboration would be a subject of great interest for the development of institutional theories aiming to transcend neo-institutionalism. Bibliography Acemoglu, D. and Robinson, J. 2016. Paths to inclusive political institutions, in Eloranta, J., Golson, E., Markevich, A. and Wolf, N. (eds.), Economic History of Warfare and State Formation , Berlin and Heidelberg, Springer Google Scholar CrossRef Search ADS Adams, S. 2014. Autonomy , in Adams, S. (ed.), Cornelius Castoriadis: Key Concepts, London and New York, Bloomsbury Academic Agassi, J. 1975. Institutional individualism, British Journal of Sociology , vol. 26, 144– 55 Google Scholar CrossRef Search ADS Aristotle. Politics , edited and translated by Barker, E. (ed.), Oxford, Oxford University Press CrossRef Search ADS Arnason, J. 2014. Creative imagination , in Adams, S. (ed.), CorneliusCastoriadis: Key Concepts, London and New York, Bloomsbury Academic Benson, B. L. 1989. The spontaneous evolution of commercial law, Southern Economic Journal , vol. 55, no. 3, 644– 61 Google Scholar CrossRef Search ADS Berkowitz, D., Pistor, K. and Richard, J. F. 2003. The transplant effect, American Journal of Comparative Law , vol. 51, 163– 204 Google Scholar CrossRef Search ADS Bowles, S. 1998. Endogenous preferences: the cultural consequences of markets and other economic institutions, Journal of Economic Literature , vol. 36, no. 1, 75– 111 Castoriadis, C.[ 1975] 1998. The Imaginary Institution of Society , Cambridge, MA, MIT Press Castoriadis, C. 1982. Imaginary significations, in Escobar, E., Gondicas, M. and Vernay, P. (eds.), Cornelius Castoriadis: A Society Adrift , New York, Fordham University Press Castoriadis, C.  1997. The logic of magmas and the question of autonomy, in Curtis, D. (ed.), The Castoriadis Reader , Oxford and Malden, Blackwell Google Scholar CrossRef Search ADS Castoriadis, C.  1991. Individual, society, rationality, history, in Curtis, D. (ed.), Cornelius Castoriadis: Philosophy, Politics, Autonomy , Oxford, Oxford University Press Google Scholar CrossRef Search ADS Castoriadis, C.  1997. Done and to be done, in Curtis, D. (ed.), The Castoriadis Reader , Oxford and Malden, Blackwell Castoriadis, C. 1997. World in Fragments: Writings on Politics, Society, Psychoanalysis, and the Imagination , Curtis, D. A. (ed.), Stanford, CA, Stanford University Press Castoriadis, C.  2007. Figures of the Thinkable , Stanford, CA, Stanford University Press Chang, H. and Evans, P. 2005. The role of institutions in economic change, in De Paula, S. and Dymski, G. (eds.), Reimagining Growth, London, Zed Books Commons, J. 1931. Institutional economics, American Economic Review , vol. 21, no. 4, 648– 57 Commons, J. 1934. Institutional Economics: Its Place in Political Economy , New York, Macmillan Curtis, D. 1997. Foreword, in Curtis, D. (ed.), The Castoriadis Reader , Oxford, Blackwell David, P. A. 1994. Why are institutions the ‘carriers of history’? Path dependence and the evolution of conventions, organizations and institutions, Structural Change and Economic Dynamics , vol. 5, no. 2, 205– 19 Google Scholar CrossRef Search ADS Dosse, F. 2014. Castoriadis: Democracy’s Advocate , OECD Insights, 5 October 2014, http://oecdinsights.org/2014/10/05/castoriadis-democracys-advocate (date last accessed 7 December 2017) Fernández-Huerga, E. 2008. The economic behavior of human beings: the institutional/Post-Keynesian model , Journal of Economic Issues , vol. 42, no. 3, 709– 26 Gelderblom, O. and Jonker, J. 2004. Completing a financial revolution: the finance of the Dutch East India trade and the rise of the Amsterdam capital market, 1595–1612, Journal of Economic History , vol. 64, no. 3, 641– 72 Gordon, D. 1989. The origins of language: a review, Review of Austrian Economics , vol. 7, 245– 51 Google Scholar CrossRef Search ADS Greenfeld, L. 2001. The Spirit of Capitalism: Nationalism and Economic Growth , Cambridge, MA, Harvard University Press Habermas, J. 1987. The Philosophical Discourse of Modernity: Twelve Lectures , Cambridge, MA, MIT Press Hall, P. A. and Thelen, K. 2009. Institutional change in varieties of capitalism, Socio-Economic Review , vol. 7, 7– 34 Google Scholar CrossRef Search ADS Hodgson, G. 1998. The approach of institutional economics, Journal of Economic Literature , vol. 36, 166– 92 Hodgson, G. 2015. Conceptualizing Capitalism: Institutions, Evolution, Future , Chicago, University of Chicago Press Google Scholar CrossRef Search ADS Klooger, J. 2014. Anlehnung (Leaning On) , in Adams, S. (ed.), Cornelius Castoriadis: Key Concepts, London and New York, Bloomsbury Academic La Porta, R., Lopez-de-Silanes, F. and Shleifer, A. 2008. The economic consequences of legal origin, Journal of Economic Literature , vol. 46, 285– 332 Google Scholar CrossRef Search ADS Linden, M. 2007. Western Marxism and the Soviet Union: A Survey of Critical Theories and Debates since 1917 , London, Brill Mantzavinos, C., North, D. C. and Sharinq, S. 2004. Learning, institutions, and economic performance, Perspectives on Politics , vol. 2, 75– 84 Google Scholar CrossRef Search ADS Mayhew, A. 1987. Culture: core concept under attack, Journal of Economic Issues , vol. 21, no. 2, 586– 603 Google Scholar CrossRef Search ADS Mokyr, J. 2005. The intellectual origins of modern economic growth, Journal of Economic History , vol. 65, 285– 351 Google Scholar CrossRef Search ADS Montes, L. 2008. Newton’s real influence on Adam Smith and its context, Cambridge Journal of Economics , vol. 32, 555– 76 Google Scholar CrossRef Search ADS North, D. C. 1992. Institutions and economic theory, American Economist, vol. 36, no. 1, 3–6 North, D. C. 1994. Economic performance through time, American Economic Review , vol. 84, 359– 68 Ogilvie, S. 2007. ‘Whatever is, is right’? Economic institutions in pre-industrial Europe, Economic History Review , vol. 60, 649– 84 Google Scholar CrossRef Search ADS Patalano, R. 2007. Imagination and society: the affective side of institutions, Constitutional Political Economy , vol. 18, 223– 41 Google Scholar CrossRef Search ADS Rorty, R. 1991. Essays on Heidegger and Others: Philosophical Papers, Volume II , Cambridge, Cambridge University Press Google Scholar CrossRef Search ADS Rutherford, M. 1994. Institutions in Economics , Cambridge, Cambridge University Press Google Scholar CrossRef Search ADS Screpanti, E. and Zamagni, S. 2005. An Outline of the History of Economic Thought , Oxford, Oxford University Press Google Scholar CrossRef Search ADS Smith, K. 2010. Meaning, Subjectivity, Society: Making Sense of Modernity , Boston, Brill Google Scholar CrossRef Search ADS Smith, K. 2014A. Psyche , in Adams, S. (ed.), Cornelius Castoriadis: Key Concepts, London and New York, Bloomsbury Academic Smith, K. 2014B. Modernity , in Adams, S. (ed.), Cornelius Castoriadis: Key Concepts, London and New York, Bloomsbury Academic Straume, I. 2014. Democracy , in Adams, S. (ed.), Cornelius Castoriadis: Key Concepts, London and New York, Bloomsbury Academic Toboso, F. 2001. Institutional individualism and institutional change: the search for a middle way mode of explanation, Cambridge Journal of Economics , vol. 25, 765– 83 Google Scholar CrossRef Search ADS Veblen, T. 1898. Why is Economics not an evolutionary science?, The Quarterly, Journal of Economics , vol. 12, no. 4, 373– 97 Vouldis, A., Michaelides, P. and Milios, J. 2011. Emil Lederer and the Schumpeter, Hilferding, Tugan-Baranowsky nexus, Review of Political Economy , vol. 23, no. 3, 439– 60 Google Scholar CrossRef Search ADS Weber, M. 1978. Max Weber: Selections in Translation , edited and with an introduction by Runciman, W. G., Cambridge, Cambridge University Press Google Scholar CrossRef Search ADS Footnotes 1 Screpanti and Zamagni (2005) use the term ‘utilitarian neo-institutionalism’ to distinguish this current of thought from the ‘contractarian neo-institutionalism’ which examines the connection between institutions and justice. 2 For a review of the basic presuppositions of old institutionalism, see Hodgson (1998) and Toboso (2001). Commons argued that institutions like the corporation both control and liberate individual action (Commons, 1931, p. 651) 3 According to Aristotle, ‘man is by nature a political animal’ (Politics I, 1253a2-3). 4 The distinction between ‘needs’ and ‘wants’ (as e.g. in Fernández-Huerga, 2008) is consistent also with Castoriadis’s work; this distinction could be based on the degree to which a need/want rests on social significations rather than on biological necessities. 5 Given that the symbolic component is foremost linked to language, Castoriadis referred to the former component of institutions as a second-order symbolism. 6 In fact, Weber called the aforementioned inference, a ‘monstrous misunderstanding’ (ungeheuer Missverständnis). 7 Castoriadis considers the Weberian term, Zweckrationalität, not fortunate in the sense that it focuses on the ‘ends’, although instrumental rationality is defined with respect to the means used, and proposes that Mittelnrationalität would express Weber’s though more concisely. 8 Douglass North and co-authors emphasize the effects of ‘symbolic storage systems’, such as writing and the modern computer, on the development of human societies (Mantzavinos et al., 2004). In particular, they argue that symbolic systems make possible complex forms of human organization and posit a flow of causation from the continuous development of symbolic systems to the complexity of human organizations. 9 This schema admittedly leaves out of consideration the internal conflicts within social groups but could be enhanced in this direction. 10 Mantzavinos et al. (2004) represent a richer version of New Institutional Economics, especially with respect to the cognitive aspect of institutional change, which is discussed below. 11 This is a specific instance of a non-ergodic environment but here the cause of the non-ergodicity is the human behaviour itself. 12 Furthermore, Castoriadis analyses at length the perceptions prevailing within historical societies on the source of institutions, distinguishing between ‘heteronomy’ (societies which take institutions as defined by exogenous sources, e.g. God) and ‘autonomy’ (the society recognises that institutions are its own creation and organises the change of institutions as a regular political act). 13 Of course the persistence of institutions is time varying and is different between incremental institutional changes and ruptures. 14 The assumption of the individual as ‘given’, i.e. that his/her preference function should be taken as given, is considered as a primary limitation of new institutional theories by Hodgson (1998, e.g., p. 176). 15 Chang and Evans (2005) propose that ‘neither a functionalist view in which what is must be “efficient” since otherwise it would not exist, nor an instrumentalist view in which institutions are created and changed to reflect the exogenously defined interests of the powerful are adequate. Instead, we argue for a more “culturalist” … perspective in which institutional change depends on a combination of interest-based and cultural/ideological projects (in which worldview may shape interests as well as vice versa). Simply put, changing institutions requires changing the worldviews that inevitably underlie institutional frames’. 16 However, in this approach there is no mention of a concept corresponding to the symbolic component of institutions characterizing Castoriadis’s analysis. 17 An anonymous reviewer argued against the use of the word ‘evolutionary’ for the above-presented schema, which was used in a previous draft of this paper. Given the emphasis that Castoriadis places on human agency in the form of the creation of meanings and imaginary significations, as opposed to the emergence of forms that survive evolutionary criteria as it is usually stressed in evolutionary theories, Castoriadis’s approach can be considered as distinct from other evolutionary theories of institutions. An element which is usually absent in evolutionary approaches is that institutional change brought about by novel social significations can persist not because it is more ‘evolutionarily fitting’ but because once it is adopted it becomes part of the prevailing social imaginary and represents a raw material for further institutional change, i.e. a lock-in effect may be operating. © The Author(s) 2018. Published by Oxford University Press on behalf of the Cambridge Political Economy Society. All rights reserved.
Cambridge Journal of Economics – Oxford University Press
Published: Jan 4, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera