Scholarly article on topic 'Institutional re(turns) and the strategic - relational approach'

Institutional re(turns) and the strategic - relational approach Academic research paper on "Law"

0
0
Share paper
Academic journal
Environ. Plann. A
OECD Field of science
Keywords
{""}

Academic research paper on topic "Institutional re(turns) and the strategic - relational approach"

00l:10.1068/a32183

Institutional re(turns) and the strategic-relational approach

Bob Jessop

Department of Sociology, University of Lancaster, Lancaster LAI 4YL, England; e-mail: r.jessop@lancaster.ac.uk

Received 22 October 1999; in revised form 25 November 2000

Abstract. The author distinguishes and comments on three different forms of the institutional turn: thematic, methodological, and ontological. He argues that there is a wide range of institutional turns that have been undertaken for quite different theoretical, empirical, and policy-related reasons; and suggests that the returns from any given institutional turn are by no means guaranteed to be positive. The different senses in which 'institutions matter' are explored and the need to contextualize the institutional turn, both at more macro and at more microlevels, is also emphasized. One way of undertaking this contextualization is through the 'strategic - relational approach' with its concern with both the structural and the strategic dimensions of a contextualized institutional analysis. As well as presenting the approach in general terms, the author also illustrates its relevance to the spatiotemporal dimensions of institutional analysis. Eight broad conclusions about the institutional turn are presented.

The institutional turn can be defined in very broad and loose terms as the more or less consistent elaboration of the intuition, hypothesis, or discovery that 'institutions matter' in one or more theoretical, empirical, or practical contexts where their existence and/ or relevance had previously been overlooked, denied, or deliberately ignored. This definition does not take us very far, however, because there can be many forms of institutional turn. There is wide variation in how institutions are defined, in the respects in which the are held to matter, and in the reasons for suggesting that they do. The institutional turn can also refer to the personal intellectual trajectories of individual scholars; to general developments within a particular approach regardless of individual proponents; and to changes in the relative weight of different approaches in a broader disciplinary field—or even in the social sciences more generally. Thus one can say that a scholar makes an institutional turn when he or she rejects an earlier essentialist account of patriarchy and examines the institutional specificities and dynamics of different gender regimes; that neoclassical economics made an institutional turn when it adopted a transaction costs approach to explain the problematic existence of the firm as an economic institution; that geography as a discipline initiated an institutional turn when it broke with the qualitative, behavioural geography versus quantitative, spatial science dichotomy to adopt a more middle-range approach concerned with the role of specific formal organizations in producing geographical patterns;® and that the social sciences as a whole have shown a renewed interest in institutions in the last two decades. Accordingly, describing and evaluating the 'institutional turn' requires one to address a complex polyvalent phenomenon with many different and contrary, if not actually contradictory, aspects.

Institutional turns can be seen as moments in the continuing self-organization of social scientific enquiries and also as moments in a new (or renewed) interest in institutional design in diverse policy fields. Thus they can be assessed in terms

(1) See the various contributions in Flowerdew (1982). These, in common with many other institu-tionalist analyses, typically define formal organizations as 'institutions'. For comments on this conflation of organization and institution, see below.

of the value they add in the social sciences and/or their effectiveness in making and implementing policy. With regard to the first of these, we can distinguish analytically three forms of turn. The first, and simplest, can be called the thematic turn, that is, the intuition, hypothesis, or discovery that various institutional aspects of social life should be included among the key themes of social enquiry. The second can be named a methodological turn, that is, the intuition, hypothesis, or discovery that the institutional aspects of social life provide a fruitful—or even, indeed, the most productive—entry point for exploring and explaining the social world even if the ensuing research is later extended to include other themes or explanatory factors. The third can be described as an ontological turn, that is, the intuition, hypothesis, or discovery that institutions constitute the essential foundations of social existence. Not all of those who thematize institutions make a methodological turn; nor, of course, do all of those who make such a turn affirm an ontological institutionalism.(2)

In this paper I offer an essentially theoretical discussion of the nature and limits of these three forms of institutional turn in the social sciences as a whole. Other papers in this theme issue of Environment and Planning A are more directly concerned with the institutional turn in geography. Nonetheless, it is worth noting that, to the extent that such institutional turns have originated within geography itself—as opposed to having been imported from outside—they are not so strongly identified with a specific intellectual movement nor are they so strongly differentiated from other approaches as they have been elsewhere in the social sciences.® This reflects the fact that such turns in geography are typically undertaken pragmatically for some particular thematic and/or methodological purpose, rather than being firmly based on ontological first principles. It is tempting to relate this to the eclectic nature of human geography as an academic enterprise organized around place, space, and scale and to its openness to theoretical and empirical developments in the social sciences and humanities/4) But this suggestion is contraindicated by the fact that the cultural turn in geography—not to be confused with cultural geography as a subfield of the discipline—is often explicitly linked to strong ontological claims of one kind or another (for example, Cook et al, 2000; Mitchell, 2000) and, unsurprisingly, has proved more controversial because of these claims (see, for example, Barnett, 1998; Castree, 1999). More generally, although I argue that thematic institutional turns in any discipline are often theoretically trivial (even where they generate important empirical insights and/or policy conclusions), I also propose that the methodological turn can prove very valuable heuristically. Both claims are highly relevant to geographical work on institutions. I then briefly affirm the importance of the ontological turn for some purposes but suggest a broader,

(2) The same threefold distinction could be made for other turns, such as the cultural turn.

(3) It is worth noting here that, although the regular reports on subfields and approaches in Progress in Human Geography sometimes refer to specific studies on the effectivity of institutions and/or discuss the relevance of developments in other disciplines (such as the new institutional economics), this journal has not commissioned any dedicated reviews on institutionalism or the institutional turn as such in geography. Likewise, the fourth edition of Johnston et al's Dictionary of Human Geography (2000) contains no explicit entries on 'institution' or 'institutionalism' apart from one concerned with the 'new institutional economics'. Peet's Modern Geographical Thought (1998), which surveys a wide range of developments, has no general discussion of how institutions matter. The introduction to a recent themed issue of Geoforum on institutional geographies identifies a pioneering anthology on the geography of institutions dating from 1982 (Flowerdew, 1982) but cites no other major text devoted to the topic (Philo and Parr, 2000). The articles included in this issue are quite heterogeneous and largely adopt a pragmatic approach to institutions, rather than one based on an ontological turn.

(4) In noting this one should not ignore the extent to which geography has contributed in turn to the development of other social science disciplines, especially in regard to the understanding of the spatiality of social life (see, for example, Martin, 1999).

more developed, 'strategic - relational' alternative and note its implications for the spatiotemporal moments of an institutional analysis. In addition, without affirming all the claims advanced by institutionalists across disciplines and for diverse purposes, I offer eight general conclusions about why institutions might matter.

Three types of institutional turn

Since thematic institutionalism is widespread in the social sciences outside economics, the most interesting question about thematic turns is why they should ever be deemed necessary. Answers can often be found in the methodologies and/or ontologies that informed the earlier approach(es) of the scholars, schools, or disciplines making such a turn. These generally involve methodological and/or ontological individualism, or their equivalent forms of holism. For individualists, the methodologically appropriate and/ or ontologically irreducible microfoundations of social life are located in the identities, interests, calculations, meaning systems, and actions of individual actors. In making a thematic turn, individualists seek to explain the emergence of institutions in terms of individualist microfoundations in order to show the latter's explanatory power and/or to study how such emergent institutions react back on individual behaviour. In contrast, holists assume the primacy of wider cultural or societal dynamics and seek to interpret and explain lower order phenomena in terms of macrolevel laws, logics, functional needs, or other macroproperties. Those holists who make a thematic institutional turn aim to show how such macroproperties affect institutions. In some cases, individualist and holistic theorists have been prompted to take a thematic turn when institutional crisis, change, or design become major concerns in the real world and stimulate awareness of earlier neglect of institutions. Recent examples are the crisis of Fordism, the end of the Cold War, problems of multilevel institutional design in the European Union, and calls for a new international financial architecture after the so-called 'Asian crisis'.

A well-known example of a thematic turn from a methodological (and often ontological) individualist position is the attempt by neoclassical economics to explain institutions, such as the firm, in terms of transaction costs. This approach became influential in economic and regional geography in the 1980s (see, for example, Scott, 1988; for a review and critique of the transaction costs approach, see Storper, 1997; and on institutionalism and cities more generally, see Lambooy and Moulaert, 1996). It takes institutions seriously by recognizing and problematizing their existence—but then argues that institutions can be fully explained within the neoclassical paradigm. The firm is held to be an economically rational institution because it serves in certain circumstances to lower transaction costs compared with operating in and through markets. A similar case has been made for the networked form of the industrial district and agglomeration economies in cities. Conversely, as firms face diminishing returns to organizational scale, the market will prove more efficient in other cases (see, classically, Coase, 1937; and, for a review, Williamson, 1994). However, although the transaction costs approach problematizes the existence of the firm, the latter is usually seen as a dependent or, at best, an intervening, variable. Indeed, some rational choice theorists suggest that the survival of institutions is explicable in terms of an equilibrium in transaction costs because it is rational for almost all individuals to adhere to institutional prescriptions most of the time, given that nearly all others do so too (for example, Calvert, 1995, page 60). In this sense, ''rational choice theory defines institutions as though they are subject to a continuing recall by their participants ... tantamount to contract renewal'' (Grafstein, 1992, pages 6-7).

Such arguments leave explanatory power rooted in the neoclassical microfounda-tions. The neoclassical paradigm sees institutional emergence and retention as rooted

in the universality of economizing actions taken by preconstituted rational individuals oriented to the price mechanism or, at least, to measurable forms of utility maximization. It denies that economic identities, interests, and calculation are conditioned by how markets are inserted into institutional arrangements and socioeconomic processes that shape information and conduct. Its thematic institutional turn ultimately changes nothing. In this sense, the neoclassical turn is trivial—even if it also marks a significant extension of the paradigm into new fields of enquiry.

The second type of institutional turn is methodological. This can take several forms and is usually associated with the alleged mediating role of the institutional turn in regard to well-established and troublesome ontological antinomies, epistemological dualisms, and methodological dilemmas in the social sciences. Thus institutions have been endorsed as an excellent entry point for overcoming such ontological antinomies as:

(a) structural determination and social agency (for example, the structuration approach sees institutions as recursively reproduced sets of rules and resources that constrain and enable social action);

(b) holism and individualism (for example, as emergent mesolevel phenomena, institutions are said to provide a bridge between macrophenomena and microphenomena or between macrosocial logics and microsocial foundations);

(c) necessity and contingency (for example, although, because they always need to be interpreted and renegotiated, institutions are not fully determinative of action, they do not permit any action whatsoever in an historically and sociologically amorphous, purely wilful, contingency—hence they are sites of the necessarily contingent and the contingently necessary).

They are also recommended as entry points in resolving epistemological issues such as:

(a) abstract-concrete (for example, institutional analysis allows one to reveal the specificities of national capitalisms or stages of capitalism relative to the generic features of the capitalist mode of production before one moves to the analysis of particular crises, conjunctures, etc);

(b) simple-complex (for example, analyses of the institutional embeddedness of economic activities provide a bridge between simple economic and more complex societal analyses);

(c) empirical description or grand theory (for example, the claim that a series of middle-range institutional theories could be developed to make sense of fine-grained empirical data, and later be combined to generate a general theory or, conversely, that an institutional approach provides a 'middle way' between the search for generality of theory and the desire for relevance to specific applications);®

(d) idiographic versus nomothetic approaches (for example, arguments for the 'institutionally thick' description as a way to avoid simplistic empiricism and covering law models or, again, arguments for middle-range analyses that transcend the choice between idiographic studies of particular places and an overly abstract spatial science). And they have also been proposed as resolutions for methodological dilemmas such as:

(a) anascopic (bottom-up) and katascopic (top-down) approaches to power (for example, institutions are seen as the site for the strategic codification and mediation of power relations translating between the microphysics of power, and as attempts to impose a more general strategic line on 'street level' or 'grass roots' politics);

(b) global and local approaches to spatial or scalar phenomena (for example, institutions may be regarded as mediating and articulating spatial and/or scalar divisions of labour or as shaping structural and cultural changes in the milieus of personal experience or, again, as "intermediaries between the specific everyday moments of human interaction (localised in time and space) and the general distribution of economic, (5) The second example is drawn from Manion and Flowerdew (1982, page 3).

political, cultural, communicational, symbolic and other resources [reaching across time and space'' (Philo and Parr, 2000, page 516)].

Such methodological turns are particularly common in comparative and/or historical analyses, in studies of crises and crisis management, and in work on path dependency and path shaping. In these and other cases it is suggested that institutions matter in so far as they provide the best entry point for understanding social life, even if the search for understanding is subsequently moved down towards microfoundations or up to emergent macrostructural phenomena. In this context it is no surprise that a methodological institutional turn is often presented as one among several alternative entry points with the choice among them being pragmatically determined. Thus March and Olsen (1984; 1996) often present the 'exchange' and institutional models as alternative methodologies without committing themselves to one or other on ontological grounds. A similar methodological pragmatism can be found in the 1980s proposal to 'bring the state back in' (the locus classicus here is Evans et al, 1985). And a recent comprehensive survey of rational choice, economic, sociological, and historical 'new institutionalisms' suggests that, despite obvious differences in their treatments of institutions, the origins of institutions, the relationship of institutions to individual behaviour, and so forth, they could, nonetheless, be synthesized with productive results (see Hall and Taylor, 1996). This is echoed in DiMaggio's call for his new institutionalisms to be used to search for ''common ground around particular ideas and approaches to obdurate problems'' (1998, page 699).

The third, and the most radical, type of institutional turn is ontological. It rests on the belief that institutions and institutionalization are the primary axis of collective life and social order. In Grafstein's words, ''institutions are typically conditions of choice, not objects of choice'' (1992, page 3). This approach often presupposes the existence of an instituted encompassing social order (or, at least, the primacy of tendencies for such an order to emerge and be reproduced) with little concern shown for the empirical conditions of its emergence and survival (Wagner, 1994). Thus institutions matter because they are seen, inter alia, as the points of crystallization of social forms, as defining the rules and resources of social action, as defining opportunity structures and constraints on behaviour, as shaping the way things are to be done if they are to be done, as path-dependent path-defining complexes of social relations, as the macrostructural matrices of societies and social formations, and so on.

Emile Durkheim provides an early example of this ontological position in presenting his rules of sociological method and justifying them against individualistic or psychologistic perspectives. Thus he identifies the essence of social life in the externally constraining, collectively produced, 'institutions' that every single individual must confront fully formed, unable to evade or change them (1938, page lvi, cited by Wagner, 1994, page 270). Another good example is Karl Polanyi's analysis of the economy as an instituted process. This goes beyond the thematic claim that it is worthwhile studying economic institutions, and the methodological claim that they provide a useful entry point for studying economic activities: Polanyi argues that it is in the very nature of economic activities that they are instituted and that they cannot possibly be understood otherwise (Polanyi, 1957). He elaborates this position as follows:

"The instituting of the economic process vests that process with unity and stability; it produces a structure with a definite function in society; it shifts the place of the process in society, thus adding significance to its history; it centers interest on values, motives and policy. Unity and stability, structure and function, history and policy spell out operationally the content of our assertion that the human economy is an instituted process'' (page 34).

Other forms of institutional economics also adopt this ontological position. They insist that economic activities are irreducible to the actions of Homo economicus but rather are mediated through institutions that socially embed and socially regularize behaviour. This can be seen in geographers' appropriation of the regulation approach and its elaboration around issues of the production and regulation of space, and the spatial embedding of accumulation regimes and modes of regulation. Thus, rather than study economizing behaviour and the formally rational calculation of opportunities for profit on the market, regulationists explore the differential constitution of economic rationality, the historical emergence and generalization of specific norms of production and consumption, the embeddedness of structural forms and economic practices in specific and changing institutions in particular times and places, how the development of these forms and practices is coupled to that of environing, embedding institutions, and how these last assist in the 'reproduction - regularization' of the economy (for an introduction, see Boyer, 1990; for an encyclopedic survey, see Boyer and Saillard, 1995; and, on geographical appropriations, see MacLeod, 1998).

Similar ontological claims are advanced by the new institutionalism in political science. Thus March and Olsen have often gone beyond a simple methodological turn to make the stronger ontological claim that ''Intentional, calculative action is embedded in rules and institutions that are constituted, sustained, and interpreted in a political system. ... Political actors act and organize themselves in accordance with rules and practices which are socially constructed, publicly known, anticipated and accepted. Actions of individuals and collectivities occur within these shared meanings and practices, which can be called institutions and identities'' (1996, page 249; see also March and Olsen, 1984; 1989). Analogous arguments can be found in political geography.

The polymorphism and polyvalence of institutional turns

Any of the three kinds of institutional turn can be undertaken from quite different initial positions or developmental paths. Because a turn is always relative to a previous position or path, it is unlikely that institutional turns will automatically promote convergence in and across specific theoretical approaches and general disciplinary trends, or among individual scholars. Indeed, their impact could just as easily produce divergence. This is quite evident from Hall and Taylor's recent review (1996) of the marked theoretical differences among three major versions of the 'new institutional-ism'—rational choice, sociological, and historical—in political science. Peters goes even further in identifying, admittedly in a rather ad hoc unsystematic fashion, no less than seven forms of institutionalism within the same discipline (1999).(6) There are similar divergences in the forms assumed by the so-called institutional turn in other social sciences considered individually (see, for example, Brinton and Nee, 1998 on sociology; Powell and DiMaggio, 1991 on organizational analysis; Ensminger, 1998 on anthropology). Conversely, DiMaggio identifies three forms of new institutionalism within the social sciences as a whole—rational action, social constructionist, and an approach concerned with how institutions mediate conflict—but argues that each is rooted in a different discipline: economics, sociology, and political science, respectively (1998, pages 696-697). Likewise Goodin suggests that what distinguishes the institutionalisms in different disciplines is the key variables that they 'own', the distinctive

(6) These seven forms of institutionalism comprise: rational choice, historical (concerned with the path-dependent impact of institutional innovation), empirical (pragmatic studies of how institutions matter), sociological (concerned with social embedding, population ecology, etc), interest representation (concerned with whether social movements are institutions like political parties), international (focusing on international regimes), and normative (that is, how institutions define logics of appropriateness in specific situations).

problems that institutions are said to resolve, and the suggestions they make regarding how institutions shape social life (1996, page 2). Thus he argues that the old and new historical institutionalisms are concerned with the past (and how it shapes the present and the future); that sociological institutionalism is concerned with 'the collective' —the old institutionalism focusing on how collective entities subsume and subordinate individuals and the new one examining the impact of different forms of social embedding on individuals; that economic institutionalism deals with institutional constraints on individual choice; and, finally, that political science is concerned with organizational and institutional constraints on the exercise of power and its outcomes (Goodin, 1996, pages 2 -16). By analogy, one might be tempted to claim that institutionalism in geography 'owns' space as its key variable, 'regards' areal differentiation and uneven development as its key problems, and 'studies' how institutions make a difference to landscapes, regions, places, and spaces.

The meaning of the term 'turn' also varies with the context in which it occurs. Thus the phrase 'institutional turn' can be applied to conversions, ruptures, and reversals as well as to progressive or regressive modifications in a given approach. My own view is that there should be a significant element of continuity in discontinuity if 'turn' is to be used rather than 'paradigm shift', 'epistemological break', 'methodological breakthrough', and so forth. This applies particularly to developments within a given school or approach, where a turn would involve a reorientation that preserves the initial framework. In such contexts it could even involve a major thematic extension in an otherwise-unchanged paradigm. It could involve a simple zig or zag, twist or turn, on a tacking trajectory as scholars seek to fill out relatively neglected areas within a shared paradigm—bending the stick this way or that as circumstances or fashion dictate. Or, again, it could involve a minor but permanent methodological adjustment in an established trajectory.

Given this broad, but by no means exhaustive, set of possibilities, the evolutionary mechanisms and personal reasons, if any, behind a given institutional turn are highly varied. They range from the relatively autonomous logic of social inquiry (for example, apparent anomalies), through secular shifts in the real world (for example, growing post-Cold-War interest in comparative capitalisms), to the practical demands of policy-making (for example, problems of institutional adjustment in the face of globalization). Accordingly, such turns can have little meaning outside specific contexts and conjunctures. No institutional turn is good or bad in itself: its significance depends on where those who undertake it are coming from, currently situated, and ultimately headed. All turns are also polyvalent—both materially and symbolically—as their significance can be modified by later turns.

Whether or not there is value added in an institutional turn depends on one's trajectory or location beforehand and the context in which institutions are said to matter. Thus, a methodologically individualist rational choice theorist who goes beyond a thematic turn to an ontological one and considers how institutions help to shape modes of individuality and forms of calculation might well be thought to have progressed (Carver, 1992). The same judgment might apply to an Althusserian Marxist who went beyond a structuralist view of capitalist reproduction to study the institutional mediations and accumulation strategies involved in regularizing capital as a social relation (Lipietz, 1993). Progress might also have occurred were a Foucauldian theorist to link the microphysics of power to the state's role in the strategic codification of power relations (Foucault, 1975, pages 92-96; 1980, pages 101, 122, 199-200). But what would one think of a 'plain Marxist' who previously moved in a critical realist fashion from abstract to concrete, and from simple to complex, in order to understand the contradictory nature of labour markets but now engages in a more modest

comparative institutional analysis of wage formation because he or she has become convinced that only institutions matter? One might view this turn as regressive because it ignored the constraints on labour-market regulation imposed by the basic contradictions in the wage and money forms, as well as the difference that forms of class struggle and class alliances could make within these institutions. Or what would one think of an institutionalist theorist who felt that, where uncertainty, complexity, and risk were predominant, institutions would be underdetermining—with the result that argumentation and rhetoric would be crucial to the social construction and determination of policy? One might rejoice that he or she had seen the limitations of the institutional turn and had embarked on an argumentative, narrative, or rhetorical turn (see Fischer and Forester, 1993; Roe, 1994; Throgmorton, 1996; Walsh, 2000). Or, again, what would one think of an institutionalist theorist who realized that it was no longer plausible to treat institutions as fixed, coherent, irreversible, and singular objects and took an actor-network theory and/or an ethnomethodological turn in order to highlight their fluidity, heterogeneity, reversibility, and polyvalence? One might celebrate the fact that they had begun to see institutions as ''fragile achievements, as filamental and reversible accomplishments'' and was thereby contributing to ''a non-essentialist, non-reductionist, and relational human geography'' (Philo and Parr, 2000, page 518; see also Murdoch, 1995).

Why institutions matter

There is a wide variation in how institutions are defined and in the respects in which they are held to matter. In part, this reflects different disciplinary traditions and ''the inclination to opt for a discipline-based, theory-impregnated internalist-style definition of the term'' that makes most sense from a given disciplinary perspective (Goodin, 1996, page 21). Thus the likely returns from the institutional turn will depend largely on the proposed definition of institutions and the respects in which they are held to matter. There are real problems here because definitions are often vague, diffuse, and mutually inconsistent and may even naturalize and reify institutions in the same way that neoclassical or rational choice theorists tend to reify human motivations. The conventional social scientific literature, for example, tends to regard institutions as social practices that are regularly and continuously repeated, that are linked to defined roles and social relations, that are sanctioned and maintained by social norms, and that have a major significance in the social structure (Abercrombie et al, 1994, page 216; Eisenstadt, 1968, page 409; Goodin, 1996, page 19; Peters, 1999, pages 18 -19; Wallis, 1985, pages 399-401). Examples of institutions in this sense include the family, religion, property, markets, the state, education, sport, and medicine. Structuralists and regulation theorists sometimes use the concept of 'structural forms' to describe such institutions; other theorists have called them apparatuses or dispositifs. Whatever the preferred nomenclature, however, institutions thus defined should certainly not be mistaken for their instantiation in particular cases, nor confused with the existence of organizations. Thus, to take the list above, individual families, church congregations, commodities, economic transactions, cabinets, schools, athletic competitions, or hospitals would not count as institutions.® An important alternative view regards institutions as organizations or social bodies that have major significance for the wider society and act in a quasi-corporate manner. Examples are the executive, legislative, and judicial branches of government; transnational firms, banks, or the peak organizations of capital and labour; established religious faiths; and organizations more generally. This last approach owes much to the economics and/or sociology of organizations

(7) Goffman (1961) applied the concept of 'total institution' to mental institutions, prisons, etc; but his analysis actually identified features of a distinctive class of organizations.

(see, for example, Aranson, 1998; Williamson, 1994). The concept of 'institutional thickness' introduced by Amin and Thrift rests on such a conflation of institution and organization because it involves a plethora of organizations that interact intensively, generating shared understandings, socializing costs, and developing mutual awareness of being involved in a common project to promote and sustain local or regional economic development (Amin and Thrift, 1995). There are also disagreements within and across new institutionalist approaches about the roles of informal and formal rules, norms, procedures, etc; and about the significance of the cognitive, as opposed to the normative, properties of institutions (for surveys, see Brinton and Nee, 1998; DiMaggio, 1998; Hall and Taylor, 1996; Powell and DiMaggio, 1991).

Of course, all these approaches imply that institutions matter ex definitione— otherwise they would not have major social structural significance. Thus, if a scholar, school, or discipline undertakes more than a thematic institutional turn, it is because differences in individual institutional forms, interinstitutional configurations, institutional histories, or other properties of institutions make a difference in one or more other respects to the particular issue at hand. If so, this must be because such institutional variations function as independent or intervening variables in one or another causal chain. It follows that institutional turns must always be specified in relation to specific institutional properties, specific issues, and specific alternative ways of explaining the phenomenon in question. However, as Hall and Taylor (1996) note, whereas rational choice neoinstitutionalists offer robust accounts of causality based on a thin conception of rationality, sociological and historical neoinstitutionalists are often rather imprecise when identifying the causal connections between institutions and individual behaviour. This same point is reflected self-reflexively in Amin and Thrift's recognition that, although concepts such as 'institutional thickness' are very general, vague, and liminal, they nonetheless ''hold the key to understanding the workings of the global economy'' (1995, pages 101 -102).

Even in these relatively simple terms, taking an institutional turn requires that institutions be put in their place. There are several steps involved in this. The first step is to define, locate, and thematize institutions so that they become less vague and liminal. The next is to understand how institutions operate and are reproduced through routine actions that 'do' or perform institutions. This is illustrated from a temporalized perspective in Bourdieu's arguments about the potential convergence of objective history and embodied history, habitat and habitus, position and disposition to enact history (Bourdieu, 1981, page 306).(8) Storper's work on the connection between institutions and conventions also emphasizes this point, especially in his remarks on the effects of disjunction or incongruence in these respects (1997, pages 268-270). One might then look behind the naturalization of institutions to examine institutional emergence as a complex evolutionary phenomenon that depends on specific mechanisms of variation, selection, and retention in specific spatiotemporal contexts. Further steps on a research agenda might include questions about institutional embeddedness or about institutional governance, that is, the governing of institutions and interinstitutional relations and their systemic environments. Finally, one might examine issues of institutional design and implementation—issues that would also require attention to the reflexive skills and capacities of actors as well as to the inevitable disjunctions between the intentions of institutional designers and actual institutional outcomes (see, for example, Goodin, 1996; Grafstein, 1992). As one turns from single institutions to examine institutional ensembles, institutional interfaces, institutional design, intersystemic relations, etc, more attention needs to be paid to the structural coupling and coevolution of institutions as well as to the attendant problems of their strategic coordination or (8) This quotation, as well as the later one cited from Bourdieu (1988) occur in Parker (2000).

guidance. This is where issues of path dependency, path shaping, and metagovernance are important (on the dialectic of path-dependent path shaping, see Hausner et al, 1995; and, on metagovernance, see Jessop, 1998).

A more interesting account of institutions seems to be offered by Giddens's structuration theory (for an excellent critical introduction, see Parker, 2000; for a suggestive account of the appeal of structuration theory, see Clegg, 1992). Basically, Giddens rejects the dualism that treats structure and agency as logically exclusive, and argues instead that they are mutually constitutive—and hence, in some sense, identical. In this context, he treats institutions as sets of chronically reproduced, deeply sedimented, rules and resources which constrain and facilitate social actions and which also bind social actions in time and space so that more or less systematic action patterns come to be generated and reproduced (Giddens, 1984, pages 17 - 25). A similar approach can be found in Polanyi's view of the economy as an instituted process (see also Weber, 1978). Yet Giddens makes two key innovations. First, he explicitly introduces time and space into analyses of institutions—an innovation that made structuration theory particularly appealing to geographers. And second, albeit less often noted in secondary commentaries on his overall approach, he stresses that institutions are connected to specific forms of power and domination. But Giddens does not develop the full critical potential of these innovations because he is actually more interested in the existential situation of individuals, rather than in the nature and effectivity of structures of domination, and also grounds his general analysis of structures on the atypical case of language as a set of rules and resources.(9)

An alternative perspective, that avoids the pitfalls of structuration theory whilst maintaining its emphasis on the spatiotemporal dimensions of institutions and their connection with power is the strategic - relational approach (or SRA). This was first developed to overcome the common division in Marxist analyses between capital-versus-class-theoretical accounts of the capitalist state (Jessop, 1990). As this is just a special case of the structure - agency dualism,(10) the same approach should be relevant to the latter (Jessop, 1996). I will first present it in general terms, then link it more directly to institutional analysis; and, in a final step, I will respecify the model to illustrate how it can take account of the spatiotemporal aspects of institutional analysis.

Giddens insists on the duality of structure and agency and then brackets (that is, temporarily ignores) one or other when examining its complementary moment in the duality (1984). But this is to treat structure at any given time in isolation from action, and so implies that a given structure is equally constraining and/or enabling for all actors and all actions—simply serving (no more, but no less) as a set of rules and resources for action. Similarly, action at any given time is isolated from structure, as actors are seen to choose a course of action more or less freely and skilfully within

(9) Thus Giddens treats agency and structure asymmetrically—privileging agency and displacing much of what is conventionally considered as 'structure' into a thin account of 'systems'; he analyzes agency more deeply and extensively than structure—having more ontological strata and more concepts for discussing agency than structures. Giddens gives an ontological primacy to the knowledgeable social actor; he shows an exaggerated concern for individual agents/identities at the cost of collective agents and organizational identities and learning; he regards language as the archetypal structure even though, like knowledge as an intellectual commons, its rules and resources are not inherently scarce; and, even when he addresses scarce resources, he has a narrow view of power and domination that tends to imply that they operate evenhandedly rather than in a systematically asymmetrical manner (Archer, 1990; Boyne, 1991; Jessop, 1996; Parker, 2000).

(10) Specifically, the capital - theoretical approach derives the changing forms and functions of the state from the immanent logic of capital accumulation, and the class - theoretical approach regards class struggles as the real motor force of history.

these rules and resources.(11) This bracketing approach essentially involves adopting alternating perspectives on the structure - agency duality and thereby tends to relate structure and agency in a rather mechanical fashion. Moreover, despite its ritual reference to recursivity, it remains largely atemporal. The mutual theoretical isolation of these complementary moments at any given time (as expressed in the bracketing of one or other term) is resolved theoretically over time by claiming that specific structures get modified in and through the intended and unintended effects of action and inaction, thereby creating new sets of constraints and opportunities for action. However, even allowing for reflexive transformation of structure by agency (as proposed in Giddens's more recent work), there is little, if any, recognition (let alone adequate explanation) of the differential capacities of actors and their actions to change different structures.

One way to go beyond the duality of structuration theory is to examine structure in relation to action and action in relation to structure, rather than bracketing one of them. Structures are thereby treated analytically as strategic in their form, content, and operation; and actions are thereby treated analytically as structured, more or less context sensitive, and structuring. Applying this approach involves examining how a given structure may privilege some actors, some identities, some strategies, some spatial and temporal horizons, some actions over others; and the ways, if any, in which actors (individual and/or collective) take account of this differential privileging through 'strategic-context analysis' when choosing a course of action.(12) In other words, it involves the study of structures in terms of their structurally inscribed strategic selectivities and actions in terms of (differentially reflexive) structurally oriented strategic calculation.

The basic differences between Giddens's solution to the structure - agency dichotomy and that proposed in the SRA are indicated in figure 1 (see over). The various arrows in this figure represent the dialectical logic that underpins the SRA and its claim to transcend structuration theory. The figure depicts the logic of conceptual development and different degrees of theoretical sophistication, rather than a necessary historical sequence of institutional development or a mandatory order of presentation for empirical arguments.

The top row of the figure presents the inadmissible dichotomy between (absolute) external constraint and (unconditional) free-willed action—the two terms that serve as the initial thesis and antithesis of the theoretical movement leading to the SRA analysis of the potential for an emergent structured coherence in institutional orders. The second row then presents Giddens's structurationist analysis of the structure - agency duality, which sublates both thesis and antithesis by treating structure (defined in his terminology, however, as 'system') as an emergent effect of action, and agency as a structurally constrained and enabled mode of skilful action. But this retains a dualistic form because, at any given point in the analysis, it brackets one or other aspect of the resulting duality. The core themes of the SRA occupy the next two rows of the figure and disclose its radical 'methodological relationalism', that is, its insistence on treating social phenomena in terms of social relations. The concepts presented in the third row refer to the strategic - relational aspects of particular conjunctures; the concepts presented in the fourth row refer to the strategic-relational aspects of successive conjunctures. And the fifth row indicates a possible outcome of the recursive interaction between the strategic selectivities of institutions and the reflexive behaviour of agents in producing a structurally coherent, apparently self-reproducing, social configuration—marked in some cases by systematic contradictions or patterned incoherence.

(11) From a strategic - relational viewpoint, this 'freedom' exists only in relation to a given structure. It does not mean that actors have free will—their choices within the range of freedom permitted by a given structure are typically constrained by other factors.

(12) On strategic-context analysis, see Stones (1991).

Dichotomy

Structure

External constraints

Agency

Free-willed subjects

Dualism masquerading as duality

Emergent structures Socialized agents

Genuine dialectical duality

Structurally inscribed strategic selectivity

Structurally oriented strategic calculation

Reflexive - recursive unity of opposites

Reflexively reorganized structural configurations

Recursively selected strategies and tactics

Second-order limits to transforming selectivity

Strategies to transform selectivities

Structured coherence (or patterned incoherence)

Figure 1. A strategic-relational approach to structure and agency.

The concepts introduced after the second row preserve the admissible elements of the preceding row(s). Thus, the concepts that appear under the agency column in the third row draw attention to the possibility of reflection on the part of individual and collective actors about the strategic selectivities inscribed within structures so that they come to orient their strategies and tactics in the light of their understanding of the current conjuncture and their 'feel for the game'. This can (but need not) extend to self-reflection about the identities and interests that orient their strategies. For individuals and organizations can be reflexive, can reformulate within limits their own identities, can engage in strategic calculation about the 'objective' interests that flow from these alternative identities in particular conjunctures. Likewise, the concepts in the structure column highlight the tendency for specific structures and structural configurations to reinforce selectively specific forms of action, tactics, or strategies and to discourage others. Together these concepts indicate that the scope for the reflexive reorganization of structural configurations is subject to structurally inscribed strategic selectivity (and thus has path-dependent as well as path-shaping aspects); and that the recursive selection of strategies and tactics depends on individual, collective, or organizational learning capacities and on the 'experiences' resulting from the pursuit of different strategies and tactics in different conjunctures. This is indicated in row four and reflects what Bourdieu, rather too deterministically to my mind, refers to as individual agents' practical anticipation of the immanent necessity of their social world with the result that they reproduce their subjection to conditions similar to those in which they are placed (1988, page 783).(13)

(13) Storper likewise refers to ''a circular relation between conventions and institutions'' (1998, page 269).

Moreover, insofar as reflexively reorganized structural configurations and recursively selected strategies and tactics coevolve over time to produce a relatively stable order out of a potentially unstructured complexity, we can talk of the structured coherence of this coevolving, self-organizing, order (see the fifth row). This can be understood in terms of the continuing interaction between the reflexive reorganization of strategic selectivities and the recursive selection and retention (or evolutionary stabilization) of specific strategies and tactics oriented to those selectivities. Such a 'structured coherence' (or stability) in a given institutional complex can be ascribed to accumulation regimes, modes of regulation, innovative milieus, industrial districts, worlds of production, and so forth (on the concept of structured coherence, see Harvey, 1982). It involves a structurally inscribed strategic selectivity that rewards actions that are compatible with the recursive reproduction of the structure(s) in question. Storper expresses this circular relation as follows:

"Institutions have a strong effect, by generating regularity and precedent, in the formation of conventions that people employ to cope with the persistent and pervasive uncertainty of their interactions with other people in the economy. But by the same token, formal organized institutions can only function successfully if the rules, procedures, incentives, and sanctions they establish are integrated into the conventions that guide people's behaviour. ... Successful formal institutions, then, have a hard organizational side, and a 'soft' conventional foundation'' (1997, page 269).

Nonetheless, from the viewpoint of the SRA, this coherence is always multiply tendential. For, first, because the reproduction of structures is only ever tendential, so too are their strategic selectivities; second, because structures are strategically rather than structurally selective (see above), there is always scope for actions to overflow or circumvent structural constraints; third, because subjects are never unitary, never fully aware of the conditions of strategic action, never fully equipped to realize their preferred strategies, and always face possible opposition from actors pursuing other strategies or tactics, failure is an ever-present possibility; and, fourth, institutions often embody structural contradictions and create strategic dilemmas (see, from a strategic -relational perspective, Jessop, 1990; from an Anglo-Foucauldian perspective, Malpas and Wickham, 1995; and, from a discourse - analytical viewpoint, Scherrer, 1995).

Some accounts of discourse adopt a similar approach to the ways in which discursive paradigms privilege some interlocutors, some discursive identities/positionings, some discursive strategies and tactics, and some discursive statements over others (for example, Hay, 1996; Jenson, 1995). Combining structural and discursive foci in a more inclusive SRA would help develop a reflexive analysis (concerned with extradiscursive and discursive structures, transformative and self-transformative capacities, and individual and collective learning) well suited to the study of structurally inscribed selectivities in different fields of action. Similar points could be made in relation to the linguistic, rhetorical, and argumentative turns as potentially fruitful complements to the institutional turn (see below).

The strategic - relational approach proposed here is similar in certain respects to the 'methodological relationalism' advocated by Bourdieu (see Wacquant, 1996) and to the 'methodological situationism' of the 'new French institutionalism' (Boltanski and Thevenot, 1991). But it is closer in more respects to Archer's 'morphogenetic' approach (1995). This rejects the identity of structure and agency (implying that it is possible to bracket one or other perspective while focusing on the other) in favour of their nonidentity (thereby making it possible to explore the changing forms and effects of the relations between structure and agency). In other words, for Archer, whereas structure and agency are necessarily related ontologically, they must be distinguished

analytically in order to establish the changing nature of this relation (Archer, 1996; see also Parker, 2000, page 72). Only in this way can one pose questions about when actors can change things and when they cannot, about variations in the strength of constraints, or concerning what gives people more or less freedom (Archer, 1990, pages 78-79).

Adopting a strategic -relational approach (or an equivalent) has several implications for how to take or make an institutional turn. First, institutions never exist outside of specific action contexts. They do not matter as such, but in terms of their structurally inscribed strategic selectivity: institutions select behaviours. This institutional framing role is now recognized, for example, within rational choice theories; but it is also conceded that institutions do not fully and precisely determine the course of action (see the discussion in March and Olsen, 1996, pages 251-255). Instead, actors have some freedom of manoeuvre to choose a path of action more or less skilfully and reflexively. Second, actors not only engage in action within a given institutional matrix but, in certain circumstances, can reflexively reconstitute institutions and their resulting matrix. Their capacity to do so depends both on the changing selectivities of given institutions and on their own changing opportunities to engage in strategic action.

Thus the spaces in figure 1 could easily be renamed to take account of the different forms of institutional turn. For, apart from the crudest neoclassical or rational choice institutionalists (for whom institutions are subject to a continuing voluntaristic 'contract renewal'), those who have made the institutional turn also reject the dichotomy of external constraints and universal modes of rational action. On the structural side of the dichotomy, for example, they argue at the least for analyses of emergent 'rules of the game', including laws, inherited organizational structures, and formal and informal norms and sanctions. Likewise, on its action side, they have noted at the least the role of bounded rationality, context-bound forms of rationality, cognitive habits, selective attention, the logic(s) of appropriateness, how atomized individuals are transformed into molecular groups through their embeddedness in an inherently social world with socially determined preferences and ideologies, and so forth. Some institutionalists have also moved to the third level by emphasizing asset specificity, rigidities in transaction costs, the differential dynamics of organizational ecology, path dependency, differentiated and competitive institutional environments, 'structural holes', and other forms of structural selectivity. And/or by noting the scope for deliberative rather than automatic cognition, the key role of strategic choices, agenda control, gate-keeping, sequencing, strategic interaction, coalition formation, and various forms of entrepreneurship. Yet others have approached the fourth level by examining how institutions come to be reproduced and regularized through their coevolution with distinctive forms of appropriate conduct so that temporary equilibria are achieved in an otherwise turbulent environment. In short, in proposing a strategic -relational (or equivalent) approach, I am not rejecting the returns from institutional turns. Instead, I am offering a model which is more general, and essentially heuristic, that may serve to locate different types of institutional turn and to highlight the limitations of approaches that are one-sided and/or fail to move from dualisms or dualities to genuine recursive-reflexive dialectical analyses (for further discussion, see Jessop, 1996).

The temporality and spatiality of institutions

A strategic - relational approach implies that institutions are inherently spatiotemporal. It rejects the neoclassical account of general equilibrium that discounts the role of time (among other reasons on the grounds that individuals have perfect knowledge about the future) and that regards any temporal development as essentially reversible

(permitting return to any status quo ante). It goes beyond the trite claim that institutions exist in time and space, or that attempts to transform them must be coordinated over time and space. It goes beyond the rational choice argument that institutional variations emerge, get selected, and are retained because they are efficient in a given environment. It goes beyond the institutionalist argument that selection and retention are not quick, precise, frictionless, and path independent (March and Olsen, 1996, page 255). And it even goes beyond the neoinstitutionalist claim that institutions and their environments coevolve as environments and are modified by institutions, as well as vice versa. For the SRA implies that the structurally inscribed strategic selectivities of institutions are always and inevitably spatiotemporal and this in turn creates the space for both a geography of institutions and for a recognition of institutions as being, inter alia, geographical accomplishments (Philo and Parr, 2000, pages 517-520).

This is so for three reasons. First, all structures (and, a fortiori, all institutions) have a definite spatiotemporal extension. They emerge in specific places and at specific times, operate on one or more particular scales and with specific temporal horizons of action, have their own specific ways of articulating and interweaving their various spatial and temporal horizons of action, develop their own specific capacities to stretch social relations and to compress events in space and time,(14) in consequence, they have their own specific spatial and temporal rhythms. These spatiotemporal features should not be seen as accidental or secondary features of institutions, but as constitutive properties that help to distinguish one organization, institution, or institutional order from another. It is these features that define the power geometries or 'envelopes of space - time' associated with different ways of organizing and institutionalizing social interaction (Massey, 1993; 1995). Second, all structures (and, a fortiori, institutions) privilege the adoption, as a condition for success, of certain spatial and temporal horizons of action by those seeking to control it, resist it, or transform it. Thus the spatiotemporal selectivity of an organization, institution, or institutional ensemble involves the diverse modalities in and through which spatial and temporal horizons of action in different fields are produced, spatial and temporal rhythms are created, and some practices and strategies are privileged and others made more difficult to realize according to how they 'match' the temporal and spatial patterns inscribed in the structures in question. Spatiotemporal matrices are always differentially distantiated and differentially compressed; and strategies and tactics can be oriented to the most appropriate spatiotemporal horizons, to changing the forms of chronotopic governance, the reflexive narration of past and present to change the future, and so on. And, third, a short-term constraint for a given agent or set of agents could become a conjunctural opportunity over a longer time horizon if there were a shift in strategy. This in turn implies that agents may be able to pursue different types of alliance strategy and so modify the selective impact upon themselves and others of social structural (including, a fortiori, institutional) constraints and opportunities. Likewise, with respect to the spatial dimension of strategic contexts, this approach implies that, because agents may be able to operate across variable spatial scales as well as across several time horizons, spatial structural constraints and conjunctural opportunities are also determined in a strategic-relational manner.

A possible outcome of the reflexive reorganization of spatiotemporal matrices and the recursive selection of strategies and tactics are specific 'spatiotemporal fixes'. These

(14) Implied here is a distinction between space - time distantiation (the stretching of social relations over time and space) and space-time compression (the conquest of space by time through increased velocity of movement and the social 'production' of more events within a given time period). They provide different bases for the exercise of power and should not be confused. For further discussion, see Jessop (1999).

establish spatial and temporal boundaries within which the relative structural coherence of an institutional complex is secured; they externalize the material and social costs of securing this coherence beyond the spatial, temporal, and social boundaries of this fix by displacing them elsewhere and/or deferring them into the future. Even within these boundaries we typically find that some classes, class fractions, social categories, or other social forces located within these spatiotemporal boundaries are marginalized, excluded, or subject to coercion. The primary scales and temporal horizons around which these fixes are constructed, and the extent of their coherence, vary considerably over time (for an extended discussion of spatiotemporal fixes in Atlantic Fordism, see Jessop, 2000).

The basic theoretical logic of this strategic - relational approach to the spatiotemporal features of interaction, organization, and institutions is depicted in figure 2, which has the same basic structure as figure 1 and should be interpreted in the same way. Limitations of space preclude the type of detailed discussion presented in relation to figure 1. But a few brief conclusions are in order. First, the SRA does not posit abstract, unlocated, and atemporal structures or wholly routinized activities performed by 'cultural dupes' or habituated actors. Structures are irredeemably concrete, spatial-ized, and temporalized; and they have no meaning outside the context of specific agents pursuing specific strategies —even if these last are expressed at the level of practical consciousness rather than in an explicit, reflexive manner. It is impossible to conceptualize specific structural (including, a fortiori, institutional) constraints out-

Simple dichotomy

Structure

Newtonian space - time (external, absolute)

Agency

Kantian a prioris (ideal, universal)

Heterogeneous dualities

Embedded spatialities and embedded temporalities (emergent, regularized)

'Social spaces' and 'social times' (constructed measures)

Genuine dialectical dualities

Structurally inscribed spatiotemporal selectivities (differential spatiotemporal constraints)

Spatiotemporally oriented strategic calculations (differential horizons)

Reflexive - recursive unity of opposites

Reflexively reorganized spatiotemporal matrices (differentially distantiated, differentially compressed)

Recursively selected strategies and tactics (chronotopic governance, 'historicity', etc)

Structured coherence (or incoherence) Figure 2. A strategic - relational approach to spatiotemporal selectivities.

side specific time horizons and spatial scales of action, as any constraint could be rendered inoperable through competent actors' choice of longer term and/or spatially more appropriate strategies that are concerned to disrupt or reconfigure the existing hierarchies of structures (including institutions) and the selective patterns of constraint and opportunity with which they are associated.

Nonetheless, it is in this context that one can study the spatiotemporal dialectics involved in strategy and tactics and the spatiotemporal dialectics of path dependency and path shaping. On the former, for example, de Certeau suggests that, whereas ''strategies pin their hopes on the resistance that the establishment of a place offers to the erosion of time; tactics on a clever utilization of time, of the opportunities it presents and also of the play that it introduces into the foundations of power. ... the two ways of acting can be distinguished according to whether they bet on place or on time'' (1985, pages 38-39). This can be linked to the dialectic of path dependency insofar as this is the product of reflexive path shaping. Path dependency implies that the prior development of an institution shapes current and future trajectories. It suggests that institutional legacies limit current possibilities or options in institutional innovation. History make a difference. But this need not imply fatalism: social forces could intervene in current conjunctures and actively rearticulate them so that new trajectories become possible. Reflexivity involves second-order observation of one's situation, actions, and its repercussions on one's own identity and interests. Applied in a strategic - relational context, it involves reflexively reorganized spatiotemporal matrices and recursively selected strategies and tactics.

Historicity is a particular case of reflexivity. It refers to the reflexive use of history to make history—or, more precisely, the reflexive use of 'history-as-account' to make 'history-as-event' (Gosden, 1994). Thus a well-developed sense of history or temporality would involve strategically reflexive temporally oriented actors who manage relationships between different time horizons and who take account of the ways in which specific social forms (for example, specific organizations or institutions) privilege actions oriented to certain temporal horizons over actions oriented to others.(15) Historicity is especially relevant during crises or other exceptional periods that disrupt established routines and habits (Debray, 1973).

Concluding remarks

Eight main lessons can be drawn from this review of the institutional turn in the social sciences. First, and most obviously, there is actually no such thing as the institutional turn in the sense of a generic turn that is made wherever and whenever a scholar, school, or discipline adopts a particular version of the new institutionalism or discovers one or another version of the old institutionalism. There are only specific institutional turns in particular contexts, made for specific purposes, and, perhaps, the chaotic sum of all such institutional turns. Thus the actual meaning and significance of a given institutional turn depends on the nature of the turn (thematic, methodological, or ontological), the position or path from which the turn is made (micro -macro, idiographic-nomothetic, anascopic-katascopic, etc), the particular theoretical or disciplinary framework within which it is made, and the extent to which institutions are reified and naturalized or, alternatively, analyzed in strategic - relational (or equivalent) terms. This implies that an institutional turn has no value in itself. Its descriptive and explanatory returns, if any, depend on how it is integrated into a continuing research programme—and on the willingness of its originator(s) or follower(s) to

(15) An example of the temporal selectivity of social forms is the contrast between the 'short-termism' favoured by UK financial institutions and the 'long termism' favoured by the German industrial - financial system.

make further turns or, even, ruptures where necessary or appropriate. In short, institutional turns (and new institutionalisms in general) should be seen as particular moments in a more complex process of scientific enquiry which itself proceeds in a dialectical rather than linear fashion.

Second, a major problem in many early institutional turns is that institutions were taken for granted, reified, or naturalized. A strategic-relational approach suggests that they should be analyzed as complex emergent phenomena, whose reproduction is incomplete, provisional, and unstable, and which coevolve with a range of other complex emergent phenomena. Institutions must be deconstructed rather than reified. In particular, they have histories. They are path-dependent, emergent phenomena, recursively reproduced through specific forms of action. Institutionalization involves not only the conduct of agents and their conditions of action, but also the very constitution of agents, identities, interests, and strategies. Institutionalization cocon-stitutes institutions as action contexts and actors as their institutional supports. This coconstitution is always deeply problematic. Thus neoinstitutionalists should examine the many and varied struggles over the constitution of institutions, competing strategies, tactics, and techniques of institutionalization, and the contingently necessary incompleteness, provisional nature, and instability of attempts to govern or guide them. Precisely because institutions are never fully constituted, space exists for competing institutional projects and designs. This ensures that the future remains pregnant with a surplus of possibilities. The new institutionalism in economics and rational choice institutionalism do not do justice to this set of issues. But many versions of sociological and historical new institutionalism do take explicit account of them.

Third, institutions cannot be meaningfully or productively analyzed without locating actors, identities, interests, strategies, or tactics in a wider strategic-relational context. At any given point in time, institutional analysis occurs prior to action—even if the action subsequently transforms institutions and institutional contexts (Grafstein, 1992). Interrelated constraints matter because actors cannot change all the conditions of action at once. In this sense, "explanation of the rules of the game and the focal points that attract [strategic] actors rests on the sort of institutional analysis provided by sociology" (Nee and Strang, 1998, pages 713-714). Social scientific explanations must be formally adequate in the sense that they explain all the effects included within the explanandum (which will not, of course, include every conceivable aspect of the phenomenon in question); and they must also be socially adequate insofar as they explain the discursive (intentional, meaningful, subjective, interpretative, etc) features involved in the social mediation of the chain of events which produce the explanandum. Social constructionist forms of institutionalism are particularly useful in this regard.

Fourth, a strategic - relational analysis would examine reflexivity as well as recur-sivity. In other words, it would address agents' capacity to engage in learning and to reflect on institutional context, institutional design, etc. This suggests the importance of adding a reflexive turn to the institutional turn in order to take account of actors' capacity to monitor their own actions; to integrate social science knowledge into their activities; and to programme their own development (producing evolution in the modes of evolution). Indeed, if the literature on the increased importance of the learning economy, the knowledge-driven economy, informational capitalism, reflexive cities, and so on, is to be believed, such a focus on reflexivity is even more important now than before.

Fifth, although time and space are important dimensions of institutions at micro and macrolevels, they are often neglected in institutional analyses. These analyses must go beyond reference to time and space as external parameters of institutions and/or action. They should pay careful attention both to (a) the temporalities and spatialities

inscribed in (and reproduced through) specific institutional forms, and to (b) the differential temporal and spatial horizons of various actors and their capacities to shift horizons, modify temporalities and spatialities, jump scales, and so forth. Institutions provide a framework in which relevant actors can reach and consolidate agreements over (albeit possibly differential) spatial and temporal horizons of action vis-a-vis their environment. They may also stabilize the cognitive and normative expectations of these actors by shaping and promoting a common worldview as well as developing adequate solutions to sequencing problems, that is, the predictable ordering of various actions, policies, or processes over time, especially where they have different temporal logics.(16)

Sixth, institutions have both microfoundations and macrocontexts. They are sustained and instantiated in individual, organizational, and interorganizational activities but they are also embedded in functionally differentiated institutional orders (especially those that can be interpreted as autopoietic subsystems)(17) in a complex, decentred societal formation (lessop, 1997). This is where the historical-institutionalist approach has some real merit. Nonetheless, analyses that do not attempt to locate institutions within broader contexts have problems addressing the limitations of institutional design or institutional change. A useful illustration of this is found in the contradictions of capitalism and the role of institutions in transforming the forms in and through which these contradictions appear. Institutional analyses certainly permit distinctions to be made among different forms or stages of capitalism, and facilitate historical and comparative studies of capitalist societies. But such analyses cannot explain the generic features of capitalism, and ignore the generic constraints imposed by the self-organizing dynamic and 'ecological dominance'(18) of capitalism in favour of more middle-range analyses. This is a potential weakness in, for example, the regulation approach. Although its interest in structural forms (institutions) enabled it to develop historically specific analyses of accumulation regimes and the ways in which modes of regulation embodied specific institutionalized class compromises, more recent regulationist analyses have tended to ignore the inherent limitations, contradictions, and dilemmas of any and all accumulation regimes and their modes of regulation. This is reflected in problems seen in some recent regulationist analyses of the neoliberal forms of globalization and post-Fordism (see lessop, 1999).

Seventh, if institutions do matter, then the institutional turn is justified. But we must also ask what else matters? There are other turns that can—and perhaps should—be made. This is especially necessary where institutions have been reified and naturalized. Further shifts could include discursive turns (or subspecies such as the rhetorical, argumentative, and narrative turns), and these too could be thematic, methodological, or ontological. For example, Somers notes that ''it is through narra-tivity that we come to know, understand, and make sense of the social world, and it is through narratives and narrativity that we constitute our social identities. ... all of us become to be who we are (however ephemeral, multiple, and changing) by being located or locating ourselves (usually unconsciously) in social narratives rarely of our own making '' (Somers, 1994, page 606, italics in original). The argumentative turn

(16) A good example is the speed and sequencing of reforms in and across different social domains in postcommunist transition (Hausner et al, 1995).

(17) Autopoiesis is a condition of radical autonomy secured through self-organization. It emerges when a system defines its own boundaries relative to its environment, develops its own operational code, implements its own programmes, reproduces its own elements in a closed circuit, and obeys its own laws of motion.

(18) Ecological dominance exists to the extent that other systems are obliged to adapt more to a given system than this is obliged to adapt to them: in this sense they assume a dominant role in shaping the coevolution of the ecological system as a whole. It has been argued that the economy is the ecologically dominant system in the modern world.

could prove useful where uncertainty, risk, social polarization, or contradictions among institutions mean that these last underdetermine behaviour and thus open a major space for argumentation, rhetoric, etc. (Fischer and Forester, 1993; Throgmorton, 1996; Walsh, 2000). The institutional turn could also be supplemented by a pragmatic turn. This would imply that "No further efforts are made to isolate individual actions analytically and to ask for the intentions and rationalities in them, on the one hand, nor for the accepted norms and applied rules, on the other. In the centre of interest, instead, we find the situation in its temporality, the individual's uncertainty about the identification of the situation and the interpretative effort that is required to determine, together with others, the situation as a shared and common one'' (Wagner, 1994, page 274, italics in original). Whether or not these turns occur will depend, of course, on the development of particular research agendas, paradigms, or disciplines.

Eighth, and finally, it would be useful to make a self-reflexive turn. This means that social theorists must be reflexive about the nature of their work and its implications, including its repercussions on what is studied. A common problem with the institutional turn is the lack of reflexivity on the part of its theorists on its nature—and thus of its possible limits. Applying the strategic - relational approach to the institutional turn would suggest that it is itself path independent as well as path shaping—that it is not only generated by specific problems but is also shaped by the resources available to resolve such problems. This could explain the too-easy rapprochement between neoclassical and rational choice paradigms and the general search to show how and why institutions matter. But the differential openness of disciplines to the new institutionalism (with the new economics of organization and US political science being more open, for example, than anthropology)(19) suggests that the institutional turn itself needs to be studied in institutional terms.

Acknowledgements. In writing this paper I benefitted from discussions with Ash Amin, Bjorn Asheim, Henrik Bang, Terrell Carver, Bulent Diken, Kan-Lin Hsu, Gordon MacLeod, Markus Perkmann, Andrew Sayer, and Ngai-Ling Sum. The usual disclaimers apply.

References

Abercrombie A, Hill S, Turner В S, 1994 The Penguin Dictionary of Sociology (Penguin,

Harmondsworth, Middx) Amin A, Thrift N, 1995, "Globalisation, institutional 'thickness' and the global economy'', in Managing Cities Eds P Healey, S Cameron, S Davoudi, S Graham, A Madani-Pour (John Wiley, New York) pp 91 - 108 Aranson P H, 1998, ''The new institutional analysis of politics'' Journal of Institutional and

Theoretical Economics 154 744 - 753 Archer M S, 1990, ''Human agency and social structure: a critique of Giddens'', in Anthony Giddens: Consensus and Controversy Eds J Clark, C Modgil, S Modgil (Falmer Press, Lewes, Sussex) pp 73 - 84 Archer M S, 1995 Realist Social Theory: A Morphogenetic Approach (Cambridge University Press, Cambridge)

Archer M S, 1996, ''Social integration and system integration: developing the distinction'' Sociology 30 679-699

Barnett C, 1998, ''The cultural worm turns: fashion or progress in human geography?'' Antipode 30 379-394

Boltanski L, Thevenot L, 1991 De la justification: Les economies de la grandeur [On justification:

economies of grandeur] (Gallimard, Paris) Bourdieu P, 1981, ''Men and machines'', in Advances in Social Theory and Methodology

Eds K Knorr-Cetina, A V Cicourel (Routledge and Kegan Paul, London) pp 304-317 Bourdieu P, 1988, ''Vive la crise! For heterodoxy in social science'' Theory and Society 17 773 - 787 Boyer R, 1990 The Regulation School: A Critical Introduction (Columbia University Press, New York)

(19) Compare Hall and Taylor (1996) with Ensminger (1998).

Boyer R, Saillard Y (Eds), 1995 Theorie de la regulation. L'Etat des savoirs [Regulation theory:

the state of knowledge] (Editions la Decouverte, Paris) Boyne R, 1991, "Power-knowledge and social theory: the systematic misrepresentation of

contemporary French social theory in the work of Anthony Giddens", in Giddens' Theory of Structuration: A Critical Appreciation Eds C G A Bryant, D Jary (Routledge, London) pp 52-73

Brinton M C, Nee V (Eds), 1998 The New Institutionalism in Sociology (Russell Sage

Foundation, New York) Calvert R L, 1995, "Rational actors, equilibrium, and social institutions", in Explaining Social

Institutions Eds J Knight, I Sened (University of Michigan Press, Ann Arbor, MI) pp 57 - 93 Carver T, 1992, "Putting your money where your mouth is: the social construction of individuality

in Marx's Capital" Studies in Political Thought 1(1) 19-41 Castree N, 1999, "Situating cultural twists and turns'' Environment and Planning D: Society and Space 17 257 - 260

Clegg S, 1992, "How to become an internationally famous British social theorist'' Sociological

Review 40 576-598 Coase R H, 1937, "The nature of the firm'' Economica 4 386 - 405 Cook I, Crouch D, Naylor S, Ryan J R (Eds), 2000 Cultural Turns, Geographical Turns

(Prentice-Hall, Englewood Cliffs, NJ) Debray R, 1973 Prison Writings (Allen Lane, London)

de Certeau M, 1985 The Practice of Everyday Life (University of California Press, Berkeley, CA) DiMaggio P, 1998, "The new institutionalisms: avenues of collaboration'' Journal of Institutional

and Theoretical Economics 154 696 - 705 Durkheim E, 1938 Rules of Sociological Method (Collier-Macmillan, Basingstoke, Hants) Eisenstadt S N, 1968, "Social institutions'', in International Encyclopedia of the Social Sciences, Volume 4 Ed. D Sills (Macmillan, Basingstoke, Hants; Free Press, New York) pp 409-429 Ensminger J, 1998, "Anthropology and the new institutionalism'' Journal of Institutional and

Theoretical Economics 154 774 - 789 Evans P B, Rueschemeyer D, Skocpol T (Eds), 1985 Bringing the State Back In (Cambridge

University Press, Cambridge) Fischer F, Forester J (Eds), 1993 The Argumentative Turn in Policy Analysis and Planning (Duke

University Press, Durham, NC) Flowerdew R (Ed.), 1982 Institutions and Geographical Patterns (Croom Helm, London) Foucault M, 1975 The History of Sexuality. Volume One: An Introduction (Penguin,

Harmondsworth, Middx) Foucault M, 1980 Power/Knowledge (Harvester Wheatsheaf, Hemel Hempstead, Herts) Giddens A, 1984 The Constitution of Society: Outline of The Theory of Structuration (Polity Press, Cambridge)

Goffman E, 1961 Asylums: Essays on the Social Situation of Mental Patients and Other Inmates

(Anchor Books, Garden City, NY) Goodin R E, 1996, "Institutions and their Design'', in The Theory of Institutional Design

Ed. R E Goodin (Cambridge University Press, Cambridge) pp 1 - 53 Gosden C, 1994 Social Being and Time (Blackwell, Oxford)

Grafstein R, 1992 Institutional Realism: Social and Political Constraints on Rational Actors (Yale

University Press, New Haven, CT) Hall P A, Taylor R C R, 1996, "Political science and the three new institutionalisms'' Political

Studies 44 936-957 Harvey D, 1982 The Limits to Capital (Basil Blackwell, Oxford)

Hausner J, Jessop B, Nielsen K, 1995, "Institutional change in post-socialism'', in Strategic

Choice and Path-dependency in Post-Socialism Eds J Hausner, B Jessop, K Nielsen (Edward Elgar, Cheltenham, Glos) pp 3-45 Hay C, 1996 Re-stating Social and Political Change (Open University Press, Milton Keynes, Bucks)

Jenson J, 1995, "Mapping, naming and remembering: globalisation at the end of the twentieth

century'' Review of International Political Economy 2 96-116 Jessop B, 1989, "The nation-state, surveillance, and violence'', in Social Theory of Modern Societies: Anthony Giddens and his Critics Eds D Held, J B Thompson (Cambridge University Press, Cambridge) pp 103-128 Jessop B, 1990 State Theory: Putting Capitalist States in Their Place (Polity Press, Cambridge)

Jessop B, 1996, "Interpretative sociology and the dialectic of structure and agency: reflections on Holmwood and Stewart's Explanation and Social Theory'' Theory, Culture and Society 13 119-128

Jessop B, 1997, "The governance of complexity and the complexity of governance: preliminary remarks on some problems and limits of economic guidance'', in Beyond Markets and Hierarchy: Interactive Governance and Social Complexity Eds A Amin, J Hausner (Edward Elgar, Cheltenham, Glos) pp 111 - 147 Jessop B, 1998, "The rise of governance and the risks of failure: the case of economic development''

International Social Science Journal 155 29-46 Jessop B, 1999, "Reflections on globalization and its (ll)logics'', in Globalization and the Asia Pacific: Contested Territories Eds K Olds, P Dicken, P F Kelly, L Kong, H W Yeung (Routledge, London) pp 19-38 Jessop B, 2000, "The crisis of the national spatio-temporal fix and the ecological dominance of

globalizing capitalism'' International Journal of Urban and Regional Studies 24 273 - 310 Johnston R J, Gregory D, Pratt G, Watts M (Eds), 2000 The Dictionary of Human Geography (Blackwell, Oxford)

Lambooy J G, Moulaert F, 1996, "The economic organization of cities: an institutional perspective''

International Journal of Urban and Regional Studies 20 217 - 237 Lipietz A, 1993, "From Althusserianism to 'regulation theory''', in The Althusserian Legacy

Eds E A Kaplan, M Sprinker (Verso, London) pp 99 -138 MacLeod G, 1998, ''Globalising Parisian thought-waves: recent advances in the study of social

regulation, politics, discourse and space'' Progress in Human Geography 21 530-553 Malpas J, Wickham G, 1995, "Governance and failure: on the limits of sociology'' Australia and

New Zealand Journal of Sociology 31 (3) 37 - 50 Manion T, Flowerdew R, 1982, "Introduction: institutional approaches in geography'', in

Institutions and Geographical Patterns Ed. R Flowerdew (Croom Helm, London) pp 1 - 50 March J G, Olsen J P, 1984, ''The new institutionalism: organizational factors in political life''

American Political Science Review 78 734 - 749 March J G, Olsen J P, 1989 Rediscovering Institutions (Free Press, New York) March J G, Olsen J P, 1996, ''Institutional perspectives on political institutions'' Governance 9(3) 247-264

Martin R, 1999, ''The new 'geographical turn' in economics: some critical reflections'' Cambridge

Journal of Economics 23 65 - 91 Massey D, 1993, ''Power geometries and a progressive sense of place'', in Mapping the Future: Local Cultures, Global Change Eds J Bird, B Curtis, T Putnam, G Robertson, LTickner (Routledge, London) pp 59 - 69 Massey D, 1995, ''Places and their pasts'' History Workshop Journal 39 182 -192 Mitchell D, 2000 Cultural Geography: A Critical Introduction (Blackwell, Oxford) Murdoch J, 1995, ''Actor-networks and the evolution of economic forms: combining description and explanation in theories of regulation, flexible specialisation, and networks'' Environment and Planning A 27 731 - 757 Nee V, Strang D, 1998, ''The emergence and diffusion of institutional forms'' Journal of Institutional

and Theoretical Economics 154 707 - 715 Parker J, 2000 Structuration (Open University Press, Milton Keynes, Bucks) Peet R, 1998 Modern Geographical Thought (Blackwell, Oxford)

Peters B G, 1999 Institutional Theory in Political Science: The "New Institutionalism" (Pinter, London)

Philo C, Parr H, 2000, ''Institutional geographies: introductory remarks'' Geoforum 31 513 - 521 Polanyi K, 1957, ''The economy as instituted process'', in Trade and Market in the Early Empire. Economies in History and Theory Eds K Polanyi, C M Arensberg, H W Pearson (Free Press, New York) pp 243 - 270 Powell W W, DiMaggio P J (Eds), 1991 The New Institutionalism in Organizational Analysis

(University of Chicago Press, Chicago, IL) Roe E, 1994 Narrative Policy Analysis: Theory and Practice (Duke University Press, Durham, NC) Scherrer C, 1995, ''Eine diskursanalytische Kritik der Regulationstheorie'' [A discourse-analytical

critique of regulation theory] Prokla 100 457 - 484 Scott A J, 1988 Metropolis: From the Division of Labour to Urban Form (University of California Press, Berkeley, CA)

Somers M R, 1994, ''The narrative construction of identity: a relational and network approach'' Theory and Society 23 605 - 649

Stones R A, 1991, ''Strategic context analysis: a new research strategy for structuration theory'' Sociology 25 673 - 695

Storper M, 1997 The Regional World: Territorial Development in a Global Economy (Guilford Press, New York)

Throgmorton J A, 1996 Planning as Persuasive Story Telling: The Rhetorical Construction of

Chicago's Electric Future (University of Chicago Press, Chicago, IL) Wacquant L J D, 1996, ''Toward a social praxeology: the structure and logic of Bourdieu's sociology'', in An Invitation to Reflexive Sociology Eds P Bourdieu, L J D Wacquant (University of Chicago Press, Chicago, IL) pp 1-59 Wagner P, 1994, ''Dispute, uncertainty and institution in recent French debates'' Journal of Political Philosophy 2 270-289

Wallis R, 1985, ''Institutions'', in The Social Science Encyclopedia Eds A Kuper, J Kuper

(Routledge, London) pp 399 - 401 Walsh J I, 2000, ''When do ideas matter? Explaining the successes and failures of Thatcherite

ideas'' Comparative Political Studies 33 483 - 516 Weber M, 1978 Economy and Society (University of California Press, Berkeley, CA) Williamson O E, 1994, ''Transaction cost economics and organization theory'', in The Handbook ofEconomic Sociology Eds N J Smelser, R Swedberg (Princeton University Press, Princeton, NJ; Russell Sage Foundation, New York) pp 77 - 107

© 2001 a Pion publication printed in Great Britain