Discursive institutionalism

Discursive institutionalism is the term I use for the fourth and newest of the ‘new institutionalisms’ (Schmidt 2002; see also Campbell 2001), although other terms such as ideational institutionalism (Hay 2001), constructivist institutionalism (Hay 2006), and economic constructivism (Abdelal, Blyth and Parson 2005) would also be appropriate.

Among rational choice institutionalists, the foray into the realm of ideas has been relatively short-lived. In international relations, an early move to ideas was made by Judith Goldstein (1993), who suggested that under conditions of uncertainty, ideas behave like switches (or ‘road maps’) that funnel interests down specific policy directions, serving as filters, focal points, or lenses that provide policy-makers with strategies (see also Bates etal.

In the historical institutionalist tradition, the move into ideas has been more lasting. Here, the question is really where the tipping point is between historical institutionalists who continue to see institutions as constitutive of ideas, determining which ideas are acceptable, and those who might better be called discursive institutionalists within a historical institutionalist tradi­tion because they see ideas as constitutive of institutions even if shaped by them. Thus, Peter Hall, whose earliest work was squarely within historical institutionalism, focusing on the institutional stability of institutions over time (1986), and whose latest work combined historical institutional struc­tures of capitalism with a rational choice institutionalist focus on strategic firm co-ordination (2001), in between focused on the role of economic ideas to explain change.

policy, although King also retains a strong historical institutionalist compo­nent with his emphasis on how institutional context makes it easier for the British government to take up ideas and impose reform than in the US.

Within the historical institutionalist tradition, in fact, much recent work focused on ideas tips into what I call discursive institutionalism - although the dividing line is admittedly fuzzy. What defines these is the focus on ideas as explanatory of change, often with a demonstration that such ideas do not fit predictable ‘rationalist’ interests, are underdetermined by struc­tural factors, and/or represent a break with historical paths (see discussion in Blyth 2003). Examples include Sheri Berman’s (1998) historical contrast between the German Social Democrats’ capitulation before Nazism, in large measure because they could not think beyond their long-held Marxist ideas, and the Swedish Social Democrats’ success in not only fighting fascism but also in creating a social democratic state because they were free of any such ideational legacy and able to reinvent socialism; Kate McNamara’s (1998) account of European monetary union which posited a three-step learning process of, first, policy failure, second, the search for new ideas that led to a neo-liberal consensus on monetarism and, third, the adoption of the German exemplar; and finally Mark Blyth’s (2002) analysis of the role of foundational economic ideas at moments of economic crisis first in ‘embedding’ liberalism in the 1930s and then ‘disembedding’ it beginning in the 1970s in Sweden and the United States.

It is interesting to note that most of the ideational approaches that follow from the historical institutionalist tradition sit closer to the positivist end of the positivist-constructivist continuum, and are found for the most part in comparative politics. In these approaches, ideas are seen as representing the necessary conditions for collective action within the state, by serving to redefine economic interest and to reconfigure interest-based coalitions. The focus tends to be on the cognitive aspects of ideas, that is, on how new ideas get accepted, how to determine kinds and degrees or ‘orders’ of change, generally following Kuhn (Hall 1993; Hay 2001; Schmidt 2002: Ch. 5), and what criteria for success can be applied in terms of relevance, applicability, and coherence or consistency (Schmidt 2002: Ch. 5).

Most ideational approaches that are within the sociological institutionalist tradition, by contrast, sit closer to the constructivist end, and are mostly found in international relations. In these approaches, ideas constitute the norms, narratives, discourses, and frames of reference which serve to (re)construct actors’ understandings of interests and redirect their actions within the institutions of the state. The focus here is on the normative aspects of discourse, that is, how and why new ideas ‘reso­nate’ with national values, and how they may ‘revaluate’ values, all within a logic of appropriateness (March and Olsen 1989; Rein and Schon 1991; Schmidt 2000).

In the sociological institutionalist tradition, one cannot talk about a move into ideas as such, since ideas have always been at the basis of the approach - as norms, cognitive frames, and meaning systems.

Some scholars don’t fit neatly into one or the other camp. In comparative politics in particular, those who focus on the role of economic ideas may look positivist because they consider the cognitive usages of those ideas in legitimizing policy change. But they are often also constructivists in their critique of the normative spin of those ideas.

Most of the discursive institutionalists just discussed - whether in the historical or sociological institutionalist tradition or straddling the two - tend to deal mainly with ideas, leaving the interactive processes of discourse implicit as they discuss the ideas generated, accepted, and legiti­mized by the various actors. Some scholars, however, have gone farther to formalize the interactive processes of idea generation, acceptance, and legit­imization, and to clarify how they are structured. They tend to see discourse not only as a set of ideas bringing new rules, values and practices but also as a resource used by entrepreneurial actors to produce and legitimate those ideas. Their approaches can be divided into those focused on the ‘coordinative’ discourse among policy actors and those more interested in the ‘communicative’ discourse between political actors and the public (see Schmidt 2002: Ch. 5)

In the co-ordinative sphere, discursive institutionalists tends to emphasize primarily the individuals and groups at the centre of policy construction who generate the ideas that form the bases for collective action and identity. Some of these scholars focus on the loosely connected individuals united by a common set of ideas in ‘epistemic communities’ in the international arena (Haas 1992). Others target more closely connected individuals united by the attempt to put those ideas into action through ‘advocacy coalitions’ in localized policy contexts (Sabatier and Jenkins-Smith l993) or through ‘advocacy networks’ of activists in international politics (Keck and Sikkink 1998). Yet others single out the individuals who, as ‘entrepreneurs’ (Fligstein and Mara-Drita 1996; Finnemore and Sikkink 1998) or ‘mediators’ (Jobert 1992; Muller 1995) draw on and articulate the ideas of discursive communities and coalitions in particular policy domains in domestic or international arenas.

In the communicative sphere, discursive institutionalists emphasize the use of ideas in the mass process of public persuasion in the political sphere. Some of these scholars focus on electoral politics and mass opinion (Mutz, Sniderman and Brody l996), when politicians translate the ideas developed by policy elites into the political platforms that are put to the test through voting and elections; others are more concerned with the ‘communicative action’ (Habermas l996) that frames national political understandings; yet others, on the more specific deliberations in the ‘policy forums’ of informed publics (Rein and Schon l991) about the on-going policy initiatives of governments.

Discursive institutionalism works best at explaining the dynamics of change (but also continuity) through its attention to ideas and discursive interac­tions, new or continuing. As such, it largely avoids the economic, historical, or cultural determinism of the other three ‘new institutionalisms’. By the same token, however, it risks appearing highly voluntaristic unless the structural constraints derived from the three newer institutionalisms are included. This is not so much an issue for the discursive institutionalist scholars discussed above, given that their approaches already follow from one or a combination of institutionalist traditions - but where ‘text’ appears without context, as in postmodernist approaches, the risks are significant.

There are other problems, however. In discursive approaches that follow in the sociological institutionalist tradition, there is always the danger that social construction goes too far, and that material interests qua material interests are ignored in favour of seeing everything as socially constructed within a given culture (see the critique of Sikkink 1991 by Jacobsen 1995). This leads one to question whether there is anything ‘out there’ at all, mutu­ally recognizable across cultures. But while discursive approaches in the socio­logical institutionalist tradition may suffer from too much constructivism, those in the historical institutionalist tradition may suffer from too much positivism, with political action assumed to be motivated by instrumental rationality alone (even if contextualized by history and culture), such that cognitive ideas about interests overdetermine the choice of ideas, crowding out the normative values which also colour any conceptualization of interest.

Finally, establishing causality can be a problem. Discourse, just as any other factor, sometimes matters, sometimes does not in the explanation of change. The question is when does discourse matter, that is, when does it exert a causal influence on policy change, say, by redefining interests as opposed to merely reflecting them in rationalist calculations (see Schmidt 2002), and when are other factors more significant, say, where the creation of new institutional paths or cultural norms may be better captured by historical or sociological institutionalist analysis, because actors don’t have any clear idea about what they are doing when they are doing it. Part of the reason many political scientists avoid explanations related to discourse is that it is difficult to separate it from other variables, to identify it as the independent variable. But instead of ignoring discourse because of the diffi­culties, because it cannot be the cause, it is much better to ask when is discourse a cause, that is, when does discourse serve to reconceptualize interests rather than just reflect them, to chart new institutional paths instead of simply following old ones, and to reframe cultural norms rather than only reify them.