Hall’s paradigm approach

Hall’s (Reference Hall1993) seminal paper on policy paradigms and social learning had a great impact on ideational analysis. It argued that politicians and officials use interpretive frameworks – paradigms – that specify ‘not only the goals of policy and the kind of instruments that can be used to attain them, but also the very nature of the problems they are meant to be addressing’ (p. 279). In his description of how policy paradigms work, Hall draws an analogy with Thomas Kuhn’s (Reference Kuhn1970) science paradigms. Inspired by Kuhn, first- and second-order change can be viewed as ‘normal policy making’. During this period, policy is adjusted incrementally and routinely without challenging the structure of a policy paradigm. In contrast, third-order change is marked by a radical break with the usual terms of policymaking, which leads to a rupture in the policy field. Analogous to Kuhn’s (Reference Kuhn1970) vision of paradigm change within science, Hall (Reference Hall1993) argues that a shift in a policy paradigm is most likely preceded by policy experimentation and the accumulation of anomalies – that is, ‘developments that are not fully comprehensible, even as puzzles, within the terms of the paradigm’ (p. 280). The attempts to stretch the terms of the paradigm to explain these anomalies eventually undermine the latter’s authority. Policymaking thus follows a specific kind of trajectory of punctuated equilibriums: long periods of stability occasionally ruptured by sudden changes following a paradigm shift (see Hall, Reference Hall1993: 291, n63).

Since Hall uncritically embraces a Kuhnian understanding of paradigms,Footnote 1 he also imports Kuhn’s over-systemic understanding of actors (Schmidt, Reference Schmidtn.a.). Much of the trouble starts with Hall’s argument that paradigms are incommensurable, which in effect grants paradigms monopoly over the minds of actors. Another problem with Hall’s theory is that actors do not possess a critical sense of the ideas they hold. This is seen most clearly in his argument that the taken-for-grantedness of paradigms makes them ‘unamenable to scrutiny as a whole’. This understanding of actors as essentially passive followers becomes most problematic when we turn to Hall’s theory of ideational change. Actors are unable to change their ideational structure, both during ‘normal policymaking’ and ‘paradigm shifts’. Ideas are placed at the core of the paradigm where they structure both goals, instruments, and settings, so actors cannot challenge the ideas of the paradigm without challenging its overall structure, which is impossible considering that the paradigm is ‘unamenable to scrutiny as a whole’. Thus, the only way ideational change can occur is through paradigm change. In this case, it is not the actors who function inside a paradigm who bring about the new paradigm. When enough anomalies have revealed that the paradigm is no longer able to provide solutions to the problems of the policy field, actors with an altogether different paradigm will seek to delegitimize the existing (failing) paradigm with their own view of the world, often precipitated by a change in power. In effect, we end up with a model of ideational structures that actors are unable to change, because they cannot think outside the structure. Hall’s structuralist conception of ideas would obviously be less problematic had scholars inspired by Hall (Reference Hall1993) only used paradigms metaphorically or labelled historical periods. However, we see numerous studies using Hall’s paradigm approach in explanations of ideational stability and change without noting the theoretical problems pointed out above (e.g. Skogstad, Reference Skogstad1998; Hay, Reference Hay2001; Albrekt Larsen and Goul Andersen, Reference Albrekt Larsen and Goul Andersen2009; Hodson and Mabbott, Reference Hodson and Mabbott2009).

Institutionalization as internalization

The predominant focus on the institutionalization of ideas and the use of ideational punctuated equilibrium models has further strengthened the emphasis on actors’ dependence on – and sometimes even internalization of – ideas. One example is Parsons’s (Reference Parsons2003) analysis of the ideational and institutional construction of interests in the building of the European Union.Footnote 2 In his analysis of how community-oriented ideas were institutionalized over time, blocking other ideas from organizing the political project of the European Union, Parsons (Reference Parsons2003) provides strong arguments about how ideas gain political influence through institutionalization. However, we are told little about how ideas change or how actors work with (and perhaps around) ideas. It seems clear that Parsons understands actors as both strategic and interest-oriented (cf. Parsons, Reference Parsons2007: 98), but it remains unclear to what degree actors are able to think critically about the ideas they employ. Despite the effort to support the argument that ideas matter, then, a micro-theory about how actors actually work with ideas – besides depending on them for mutual cooperation and identifying their interests – is missing.

Other discursive institutionalists have done more to attach real transformative agency to their theories of ideational change. One example is Blyth (Reference Blyth2002), who in his five-step sequential model of ideas and institutional change argues that ideas have different causal effects in different periods. The degree of reflexivity actors, employ in their use of ideas varies between these periods. For example, in the first period – in moments of crisis – actors are very dependent on ideas because they serve to reduce uncertainty by providing an interpretation of the nature of crisis as a first step to construct new institutions. In the later period of delegitimizing crisis-ridden institutions, actors seem more reflexive about their use of ideas. Blyth’s model is helpful in moving beyond the structuralism of Hall’s policy paradigms, because Blyth conceptualizes actors as both dependent on the ideational structure and able to think about their own role inside this structure. However, though we are able to coax out the contours of more active agency, the relation between ideas and actors remains underspecified, attesting to the central analytical and theoretical role institutions have been assigned at the expense of active agency.

Schmidt (Reference Schmidt2002) comes closest to building a theoretical framework where the driving force of ideational change is placed with actors. By focusing on the interactive dimension of political discourse – and by distinguishing between ideas and discourse – Schmidt argues that actors are able to think outside the ideas they use to make sense of the world. Despite its many merits, however, Schmidt’s (Reference Schmidt2002) theoretical framework suffers from the important weakness that she too ends up with an equilibrium-oriented theory of ideational change that places the decisive drive for change outside agency. Though Schmidt admits the possibility of ideational change of a first or second order outside times of crisis,Footnote 3 she somewhat surprisingly follows Hall’s (Reference Hall1993) argument that comprehensive discursive and ideational change – that is, third-order change – always happens through crisis and only very rarely, which places the impetus for significant change at some distance from agents.

In the eagerness to show that ideas matter, an underestimation of the capacity of actors to create transformative change has almost unnoticeably followed. As we saw in the examples above, many ideational studies either understate the creative and critical faculty of actors or exclude further considerations. As a result, ideational theories lack the ‘micro-foundations’ of macro-historical patterns that specify how actors are able to change otherwise stable macro-cultural norms and ideas (Schmidt, Reference Schmidt2008a). This is all the more remarkable considering that most ideational scholars within political science would acknowledge that ideas do not matter when actors do not use them (see Risse, Reference Risse1994; Berman, Reference Berman1998; Blyth, Reference Blyth2002; Béland, Reference Béland2005).

The toolkit approach

An interesting approach in this regard is the ‘toolkit’ theory – represented primarily by Swidler (Reference Swidler1986, Reference Swidler2000, Reference Swidler2003) – arguing that ‘culture influences action not by providing the ultimate values towards which action is oriented, but by shaping a repertoire or “toolkit” of habits, skills and styles from which people construct “strategies of action” ’ (Swidler, Reference Swidler1986: 273). Toolkit theory proposes a relatively ‘lean’ model of the actor, which is seen as tightly interfaced with an external environment made up of culturally based and loosely structured symbols, frames, scripts, institutions, etc. (Lizardo and Strand, Reference Lizardo and Strand2010). Toolkit theory suggests – and empirical evidence supports the argument (DiMaggio, Reference DiMaggio1997; Lizardo and Strand, Reference Lizardo and Strand2010) – that actors cannot cognitively internalize highly structured symbolic systems. An important implication follows. Toolkit theory does not require that social agents reproduce an internal model of whatever is conceived by the researcher to be external (Lizardo and Strand, Reference Lizardo and Strand2010) – or in other words: culture is not inside the actors’ minds (Swidler, Reference Swidler2000).

How, then, do agents act in spite of their limited cognitive capacities? They rely on prefabricated links offered by existing cultural institutions. These prefabricated links work through schemas – ‘intrapersonal mental structures’ (Strauss and Quinn, Reference Strauss and Quinn1997: 6) – that actors employ to handle informational complexity. Schemas help actors choose which information to use, most often ignoring information that fits awkwardly with the schema; bias which pieces of reality are paid attention to; process new information more efficiently (but in no way necessarily ‘correctly’); and create a degree of coherence in beliefs and attitudes (see DiMaggio, Reference DiMaggio1997; Swidler, Reference Swidler2000; Jacobs, Reference Jacobs2009; Martin, Reference Martin2010). Schemas are not tightly structured, though. There is strong evidence that long-term memory involves not so much storing a tape of an event, but a network of ideas (Martin, Reference Martin2010). Storage occurs through sets of connections, and the difficulty actors experience in processing information – at least in connection to long-term memory – is a problem of limiting the relevant connections that pieces of information create in our heads. The flexibility of the schemas also shows itself through continual revision in the face of new experience (Swidler, Reference Swidler2000).

In this perspective, culture is a resource – a toolkit and not a coherent system – that exists between and not inside the minds of actors, and the use of culture thus demands some creativity and critical faculty of the actor. The logic that governs cultural practice is inherently practical: ‘the fit among cultural practices is due to their common link to practical problems generated within an institutional order’ (Swidler, Reference Swidler2003: 198). The tradition or cultural structure surrounding the actor does not determine the actor’s response to problems or new circumstances, because the structure cannot limit its own application – culture can only guide, not determine, how it is used by actors (Bevir, Reference Bevir1999; Bevir and Rhodes, Reference Bevir and Rhodes2003). Since culture and ideas do not form a coherent whole – but are rather constituted by contradictions, tensions, and overlap – scepticism, ranging from suspended judgement, to doubt, to outright rejection, is a usual, not exceptional, response to culture (Swidler, Reference Swidler2003: 14). In continually improving the usefulness of ideas by making them more relevant to contemporary issues, actors often do respond selectively to them. They accept some parts, modify some, and reject others (Bevir, Reference Bevir1999: 202). In this way, we are far from Hall’s (Reference Hall1993) actors, who are not able to scrutinize their paradigm.

Though actors do not ‘contain’ ideas, differences remain between how eagerly actors question ideas.Footnote 4 We may distinguish between concrete policy ideas that actors continuously present and debate in the political system; programmatic ideas that are related to a certain policy area and provide definitions of problems and solutions; and public philosophies that form the normative and cognitive backbone of a polity (cf. Mehta, Reference Mehta2010). In principle, these ideas are open to questioning, rearrangement, and bricolage – their meanings are never settled (Carstensen, Reference Carstensen2010a) – but actors will use them differently according to how fundamental they are for their understanding of politics, culture, and society. Thus, public philosophies are often unquestionable – which is why we often consider them public philosophies in the first place – whereas policy ideas, as part of the everyday functioning of the political system, are continuously questioned and debated. Programmatic ideas – working on a kind of meso-level – may remain unquestioned for considerable periods of time and then become part of a political controversy that places them in question. This dos not mean that public philosophies are internalized by actors, but rather that they are used differently from programmatic ideas and policy ideas. These distinctions do not obscure that actors are able to think creatively about how they use macro level ideas like public philosophies strategically, for example, in the framing of their programmatic ideas.

The bricoleur

The toolkit approach ties in nicely with the second theoretical perspective of this paper, namely that of change through bricolage. While political scientists have so far not been too keen to exploit the explanatory potential of bricolage, it is no stranger to sociology (see Campbell, Reference Campbell2004, Reference Campbell2005). The bricoleur is ‘someone who works with his hands and uses devious means compared to those of a craftsman’ (Levi-Strauss, Reference Levi-Strauss1996: 16–17), loosely translated: a professional do-it-yourself man (for a general introduction, see Milner, Reference Milner2007). The anthropologist Claude Levi-Strauss (1966) first used the term to explain how pre-science, ‘primitive’ mythical thought, was able to produce an impressive body of reliable knowledge about matters such as pottery, weaving, agriculture, the domestication of animals, etc.

Unlike the engineer – in Levi-Strauss’s (1966) terms, the ideal-typical counterpart to the bricoleur – who picks his instruments based on the project at hand, the bricoleur picks from his finite set of heterogeneous tools that bear no relation to the current project. The usefulness of the instruments is not defined by a project, but from the practice of making do with ‘whatever is at hand’ (Levi-Strauss, 1962: 17–18). The work of the bricoleur is retrospective. In solving a task, he looks back to existing tools to find how these may be rearranged to fit the project at hand. The tools of the bricoleur are thus constrained by their prior use. Though they can be used differently than previously, and in this way their range of uses is infinite, they are at the same time drawn from the language of a previous practice, limiting the freedom of manoeuvre (Balkin, Reference Balkin1998). Thus, unlike the engineer who approaches his or her task with a blueprint and a pre-defined end-goal, the objective of the bricoleur is largely improvised. Note that this does not mean that the bricoleur works either illogically or irrationally. Rather, he works like a skilled practitioner, adept at making perceptive use of a limited universe of materials.

Following Campbell (Reference Campbell2005: 56), I define bricolage as ‘an innovative recombination of elements that constitutes a new way of configuring organizations, social movements, institutions, and other forms of social activity’, which entails ‘the rearrangement of elements that are already at hand, but it may also entail the blending in of new elements that have diffused from elsewhere’.

To sum up, this understanding of reasoning through bricolage – as theorized within anthropology and sociology – ties in nicely with the vision of actors using a toolkit of habits, beliefs, symbols, etc. The bricoleur cannot contain complex cultural systems but rather draws on his existing resources – in the words of toolkit theory: the schemas and institutions offered to him by culture. He does not use these resources dogmatically, following an abstract system, but rather focuses on getting his heterogeneous set of ideas to work, which means that they sustain the necessary degree of meaning to move forward.

A final note before we move on to comparing the bricoleur and decision-makers working inside a paradigm. The above theoretical discussion might raise the question: does the claim that ideas are not internalized by actors make ideas epiphenomenal? Does it rob them of the bulk of their importance? The answer is: it does not. The general argument of the ideational literature – that actors need ideas to frame and schematize the vast informational complexity that social interaction creates to act meaningfully – applies whether we see ideas as internalized or not. The choice between understanding ideas as internalized or not is not a matter of how great an impact ideas have (in both perspectives, they have a great impact), but rather constitutes a difference in how actors are believed to work with ideas. In other words, the matter of whether ideas matter is not at stake here. To argue that ideas are ‘outside the minds of actors’ does not mean that actors do not really believe in them, but rather that ideas are not ready-made general templates. Actors cannot sit back and let the ideas do the thinking for them. Instead, political actors must employ ideas creatively and pragmatically to make them work, both in matters of intellectually grasping their world and in the strategic endeavour to satisfy their political preferences. Put differently, ideas are used by actors, not the other way around. Thus, the active cognitive work on behalf of the strategic actor that the toolkit perspective places so centrally does not lend credence to the rational choice perspective of ideas as a variable that captures variance, which traditional variables cannot explain (Blyth, Reference Blyth1997). On the contrary, the vision of the bricoleur with his toolkit of ideational resources strengthens the case for the importance of ideas by zooming in on the mechanisms that make ideas so important in political analysis. By providing a micro-foundation for analysing how ideas matter to actors – a signal aim of this piece – we may thus better understand how ideas matter for political actors.

Mode of reasoning

The fundamental distinction between paradigm man and the bricoleur is found in their mode of reasoning, that is, in how they use ideas to make sense of a complex and uncertain world. In Hall’s (Reference Hall1993) understanding of actors working inside paradigms, the ideas at the core of the paradigm structure the policies and instruments used by officials and politicians. In this way, actors more or less knowingly deduce their response to new circumstances from the paradigm’s core. A process of top–down reasoning – applying general ideational templates to specific situations – is thus implicitly suggested to structure elite decision-making. Actors are not able to grasp or use ideas from other paradigms and therefore strictly adhere to the dogmas of their paradigm. Political problems must thus be solved inside the confines of the paradigm. Theories differ as to how creatively actors might work with the paradigm, but these differences generally remain implicit.

The mode of reasoning used by the bricoleur resembles the paradigm perspective in that solutions are sought using existing ideas and institutions. To identify problems and find solutions, bricoleurs are both cognitively and normatively dependent on ideas already present, and so new ideas may be used to the extent that they can be fitted to the existing set of ideas. The bricoleur, however, differs in important ways from paradigm man. First, the bricoleur is – in contrast to the dogmatism of actors working inside paradigms – above all pragmatic. That means that the bricoleur takes stock of his existing set of ideas, policies, and instruments and reinterprets them in the light of concrete circumstances. In this way, the bricoleur applies creativity and pragmatism to solve problems (Bevir, Reference Bevir1999: 238).

Second, and following his pragmatic point of departure, the bricoleur works across different epistemological domains ‘ “piecing together” what he knows from different sources in different ways’ (Freeman, Reference Freeman2007: 485). Policymakers often need to combine what from a rationalist viewpoint are incommensurable types of knowledge, which makes it necessary to cast different understandings as more coherent than they really are. This obviously implies greater ideational flexibility than is normally assigned to actors working inside paradigms. This also helps explain how actors are able to think strategically about their ideas. Through mixing different types of knowledge – practising what may be termed ‘inter-paradigm borrowing’ (Hay, Reference Hay2010) – actors may better take multiple political, cultural, or functional perspectives into consideration than within the confines of a single paradigm (see below).

The structure of ideas

Echoing this last point, the bricoleur perspective works with a significantly different understanding of the structure of ideas than that represented by paradigm theory. In paradigm theory – and more generally in theories that deal with the institutionalization of ideas – systems of ideas are most often conceptualized as closed, coherent, stable, and well defined from a core of meaning. In short, the ideas that actors use in paradigm theory are generally thought to be homogeneous, and only a crisis may destabilize this homogeneity. Indeed, this is what characterizes ‘successful’ ideas: when actors subscribe to the core tenets of a paradigm, the basis of the paradigm remains unquestioned, stable, and coherent for political actors (Carstensen, Reference Carstensen2010a; see also Parsons, Reference Parsons2007: 128).

Not so in processes of bricolage. It follows from the non-dogmatic use of ideas by the bricoleur that the sets of ideas used in bricolage do not necessarily fit together in any coherent way. Often, ideas contain elements of meaning that may conflict with other elements in the idea. This only poses a problem for the political bricoleur to the extent that it makes the idea difficult to communicate. Since the meaning of ideas is defined by the relation between different elements of meaning rather than a core (Carstensen, Reference Carstensen2010a), it is – at least in principle – possible for actors to put together ideas in an infinite number of combinations. What matters is if other political actors, not to mention the general public, are willing to accept the combination – and not whether the combination is acceptable within the logic of a certain abstract paradigm.

Above all, the process of putting together ideational elements is strategic in nature. What constrains the process of weaving together ideas in a way that creates political resonance is not the structure of a certain paradigm but rather the previous application of the idea. That is, since new ideas must be hooked on to older ideas, ideational path dependence in practice limits the possible combination of ideas (Campbell, Reference Campbell2004; Carstensen, Reference Carstensen2010a). The above discussion may create a picture of actors choosing the different elements an idea should contain, but a more proper picture is probably one of kneading or moulding an idea to try to get it to hang together and gain the acceptance of other actors – often while at the same time other actors knead and mould their related ideas. In this way, ideational battles function much as Kingdon’s (Reference Kingdon1984) anarchic ‘multiple streams’ model would suggest.

Since the bricoleur focuses on combining ideational elements to create resonance in the public and support in the political system – rather than upholding stringency in a paradigm – actors oftentimes practise inter-paradigm borrowing by using elements from the ideas of their political adversaries. One example can be found in Jabko’s (Reference Jabko2006: Ch. 3) analysis of how in European market integration, political actors used a repertoire of ideas about the market to present it as an inescapable reality in the financial arena; as a norm to be desired in the energy sector; as a space in structural policies focused on regional economic development; and as a symbol of discipline in the Economic and Monetary Union. There are many strategies that actors can follow, such as utilizing symbols or causal stories that support one’s definition of the problem; tying one’s definition to widely accepted cultural symbols; linking across problem definitions to expand the scope of the conflict and thus involve more groups (Benford and Snow, Reference Benford and Snow2000; Béland, Reference Béland2009). These strategies entail bricolage, weaving together ideational elements from different paradigms and logics, in effect creating heterogeneous sets of political ideas.

Political viability of ideas

What then determines whether an idea is useful to the political aspirations of paradigm man or the bricoleur, respectively? Here, too, we find both similarities and differences. It is evident that in both theoretical perspectives, actors try to achieve as much political impact as possible. However, the relation between politics and paradigms remains somewhat unclear. In Hall’s (Reference Hall1993) version of paradigms, actors are only able to consider first- and second-order change – that is, change in instruments and their settings – but unable to change the ideas at the core of the paradigm. How, then, do decision-makers adjust the paradigm significantly in the face of, for example, electoral defeat? And how do decision-makers strike agreements with political adversaries? More generally, are solutions to strategic political issues also deduced from the core of the paradigm? Answers to such questions unfortunately remain unclear. Thus, we strike upon the non-pragmatic nature of decision-making inside paradigms: since solutions are by definition deduced from the core of the paradigm, it is difficult to understand how politicians are able to react forcefully to new strategic circumstances. In Hall’s (Reference Hall1993) version of policy paradigms, they just do, without any explication of how this relates to the workings of a paradigm, or how the conception of strategic actors ties in with a strong focus on the dogmatic following of a paradigm.

Things are different for the bricoleur. Above all, the bricoleur works pragmatically. Because of his undogmatic approach to politics, the bricoleur focuses on putting ideas together that may create the support necessary for them to get through the political process. In this way, he resembles Lindblom’s (Reference Lindblom1959) actors who spend most of their time ‘muddling through’. Rather than ranking preferences, identifying all relevant policy alternatives, and then choosing the policy that maximizes these preferences – as a rational political actor would – a policymaker will instead more or less consciously choose a goal, review the couple of policy alternatives that occur to him – probably well known from earlier policymaking processes – and then make a choice. Through successive limited comparisons, decision-makers build from the current situation, step-by-step, and by small degrees. By basing new policy on the existing policy, the reliance on theory is greatly reduced, and the ‘test of a “good” policy is typically that various analysts find themselves directly agreeing on a policy’ (Lindblom, Reference Lindblom1959: 81). In this perspective, the actor cannot depend on a paradigm, because the actor focuses on the political viability of the policy and existing policies. This is a pragmatic agent and resembles the vision of agency presented in this piece.

What, then, determines which ideas the bricoleur uses? In general, the bricoleur seeks to answer multiple logics simultaneously. To simplify somewhat, we may speak of four general logics that ideational bricolage needs to answer: the party political logic, the general electoral logic, the national institutional logic, and the actor’s own position within the larger context. First, and perhaps most importantly, an abundance of party political factors affect the success of an idea, and we may only name a few of them. A new idea needs to fit the overall goals of the ruling parties as well as the interests of potential coalition partners, such as the opposition parties or interest organizations (Hall, Reference Hall1989). To attract the attention of other actors, an agent may attach his or her idea to a salient problem and in this process provide cognitive and normative arguments capable of satisfying policymakers and citizens alike (Schmidt, Reference Schmidt2002: 221). By arguing for the necessity of the idea – with due attention to the interest policymakers have in power and re-election – ideas can be useful in ironing out differences between coalition partners, because they provide a common ground to negotiate from (Culpepper, Reference Culpepper2008). It is thus an important task for actors to mould a set of ideas in such a way that other actors – coalition partners as well as adversaries – come to accept it.

Second, the idea has to appeal to the public, or at least as large a part of the voters as possible. The obvious reason is that politicians to gain re-election have to attend to – or at least make it look like they attend to – the interests of their constituents. In communication to the public, the bricoleur focuses on providing a ‘master’ discourse with regard to the country’s present and future that legitimizes government policy (Schmidt, Reference Schmidt2002: 235). It is possible to create policies that propose cutbacks and retrenchment without losing an election, but it is often necessary to create a discourse about crisis (Hay, Reference Hay2001) or ‘the need to reform’ (Cox, Reference Cox2001) to convince the ‘armies of beneficiaries’ (Pierson, Reference Pierson1994) that reform is necessary. In the effort to construct an effective discourse, politicians also need to uphold some consistency in their use of ideas, and the demand for consistency might ‘entrap’ actors to follow the policy implications of discourses they have accepted in the past but no longer wish to endorse (Schmidt, Reference Schmidt2007; Schmidt, Reference Schmidt2008b; Carstensen, Reference Carstensen2010b). Considering that public debate is immensely complex and very difficult to control by politicians (Legro, Reference Legro2000; Art, Reference Art2006), answering to electoral logic by communicating effectively with the public is a daunting task for any politician – a complex task they often fail to fufil to their own satisfaction.

Third, concerning the role of the national institutional context, a number of historical institutionalists (e.g. Hall, Reference Hall1989; Weir, Reference Weir1989) have argued that ideas can be expected to gain wider acceptance if the bureaucracy is open to them, which would depend on factors like how well the new ideas fit with existing institutions and routines in the administration; the resources available in the bureaucracy to adopt new ideas; and the openness of the state bureaucracy to the ideas of academia (see Skocpol and Weir, Reference Skocpol and Weir1985). From a sociological perspective, translation theory has argued that ideas need to be accommodated to the national context, where actors have to be very selective about the parts of an idea they choose to put together to form a consistent new idea. Ideas are not frozen cultural objects that are transferred from one country to another (Kjær and Pedersen, Reference Kjær and Pedersen2001; cf. Campbell, Reference Campbell2010) – rather, actors must adjust the meaning structure of the idea with due consideration of the knowledge regime (Campbell and Pedersen, Reference Campbell and Pedersen2010).

Fourth, it is worth pointing out that the institutional position of an actor greatly influences how the actor should approach the promotion of a policy idea. There are many relevant types of actors in ideational policy processes: decision-makers, state bureaucrats, epistemic communities (Haas, Reference Haas1992), advocacy-coalition networks (Sabatier and Jenkins-Smith, Reference Sabatier and Jenkins-Smith1993), more loosely coupled networks (Campbell, Reference Campbell2010: 99), interest groups like business and trade associations (Campbell, Reference Campbell2004), social movements (Benford and Snow, Reference Benford and Snow2000), officials from international organizations, etc. The strategy an actor chooses thus needs to fit the actor’s position. An actor outside parliament is, for example, more free to articulate new and controversial ideas, because the actor does not need to concern himself or herself with re-election – the actor might rather speak on behalf of whoever is paying the salary. In other words, some actors are freer to present new ideas than others – and perhaps in turn much less influential in deciding if the idea is actually put to use.

Large parts of the literature on the role of ideas in politics have thus rather consistently shown that to get an idea to the top of the political agenda, it is necessary to work creatively with the idea and adjust it according to a long list of demands – and conversely that actors working within the confines of a Kuhnian paradigm will find it difficult to succeed. All these constraints attest to the importance of combining different ideational elements in a strategic way. Rather than dogmatically following the prescriptions of a policy paradigm, actors have to work creatively and pragmatically with the ideational, political, and institutional resources at hand. Put simply, in the case of actors working inside a paradigm, ideas are chosen for their stringency, whereas the bricoleur chooses ideas that may answer multiple logics simultaneously.

Change through bricolage

Change through bricolage always entails significant continuity. In this way, bricolage is different from the paradigm perspective, where – as discussed above – significant ideational and institutional change occurs as paradigm shift, which resembles the critical junctures and punctuated equilibria so popular within historical institutionalism. Not so with bricolage. As mentioned above, when trying to change or reconfigure institutions, actors’ choices are to a large degree fixed by the existing institutional principles and practices, and so ‘the new institutions that actors build resemble the old ones, containing many elements from the past’ (Campbell, Reference Campbell2004: 70; see also Campbell, Reference Campbell2010). Since new ideas and institutions must be grafted onto existing ideas and institutions, bricolage generally leads to evolutionary change. This should not be confused with path dependence. As Campbell (Reference Campbell2005: 59) points out, ‘evolution is a concept that depicts change; path dependence is a concept that depicts stability’. Thus, to talk of change, we need to witness the introduction of new elements to the existing repertoire of ideas and institutions, and ‘the more new elements diffuse from elsewhere, become part of the available repertoire, and are translated into the bricolage, the more revolutionary and less evolutionary change becomes’ (Campbell, Reference Campbell2005: 60).

Despite these differences between path dependence and evolutionary change through bricolage, the mechanisms behind bricolage do show affinity with traditional historical institutionalist explanations of stability and change. As argued by Balkin (Reference Balkin1998: 34), bricolage is characterized as cumulative, economical (i.e. exhibiting increasing returns to scale) and having unintended consequences. The central difference between bricolage and the way agents act in historical institutionalist theory lies in the conception of stability and change and the role that a micro-theory of actors plays in the two approaches. While the latter focuses on macro-historical processes, where a self-reinforcing sequence of events creates ‘deep equilibrium’ over time through positive feedback mechanisms (Schmidt, Reference Schmidt2009), theories of bricolage work with a micro-level theory of actors working inside a cultural and political-institutional setting that never reaches equilibrium but instead continually develops through cumulative reinterpretation. Though historical institutionalist arguments are frequently couched in agency-oriented terms, they often lack further specification of the individual-level mechanisms that drive change. The vision of the bricoleur may thus help spell out the micro-theory that underpins ideational change.

Processes of change through bricolage also show affinity to Streeck and Thelen’s (Reference Streeck and Thelen2005) theory of evolutionary and incremental change. According to Streeck and Thelen (Reference Streeck and Thelen2005), the gap between rule and enactment opens up opportunities for strategic actors to create transformative endogenous change (p. 13). This gap exists because ‘enactment of a social rule is never perfect’ (p. 14). Streeck and Thelen’s emphasis on the gap between rule and enactment is based on the insight that rules are never self-evident but always subject to interpretation: ‘applying a general rule to a specific situation is a creative act that must take into account, not just the rule itself, but also the unique circumstances to which it is to be applied’ (Streeck and Thelen, Reference Streeck and Thelen2005: 14). Considering that Streeck and Thelen conceptualize actors as driven by their interpretation of complex and uncertain situations, it is a bit surprising that they do not turn to ideas as an explanation for action (Schmidt, Reference Schmidt2010b). In Streeck and Thelen’s perspective, actors seem primarily interest-oriented, but it remains unclear how these interests take shape or where they originate. With their focus on interpretive processes, the most plausible candidate seems to be ideas. Though Streeck and Thelen (Reference Streeck and Thelen2005) do not take the full step to include ideas in their explanatory model or explicitly conceptualize actors as bricoleurs, their introductory article, as well as other contributions to the edited volume, empirically support the claim that via creative reinterpretations of institutions, actors may circumvent and gradually change existing institutions without critical junctures or policy punctuations.

Empirical studies have thus offered support for the claim that actors can change institutions and ideas incrementally through bricolage by using the repertoire of multiple institutional legacies (see, e.g. Schmidt, Reference Schmidt2002; Campbell, Reference Campbell2004; Hacker, Reference Hacker2004; Streeck and Thelen, Reference Streeck and Thelen2005; Palier and Martin, Reference Palier and Martin2007; Carstensen, Reference Carstensenn.a.; Palier and Thelen, Reference Palier and Thelen2010). The process of bricolage, however, does not apply only to evolutionary change: bricolage also occurs in crises. One recent example can be found in the process of creating the bailout policies issued to save otherwise faltering financial institutions. In an analysis of how Danish politicians and officials handled the financial crisis (Carstensen, Reference Carstensenn.a.), it is shown how political actors – unable to turn to their standard solutions – reinterpreted their repertoire of policy instruments, as well as finding hooks in their existing web of ideas on which to hang new ideas imported from abroad. Since no new ideas had yet come along that could restructure the institutional setting, and because actors felt it necessary to move fast, politicians and officials resorted to what they had at hand, notably previous experiences with specific bank guarantees; expanding the borrowing possibilities of the National Bank; beefing up the Danish FSA; and a state guarantee seen in use abroad. In this way, new and old ideas were joined in a policy that senior officials judged to be ungraceful in economic theoretical terms, but nonetheless relatively effective in defrosting the Danish bank’s relations with the money market.

That bricolage also occurs in times of crisis is also recognized in sociological theory. Thus, McAdam and Scott (Reference McAdam and Scott2005: 28) argue that bricolage is particularly likely to occur during times of rapid change, because ‘by employing recognizable structural vocabulary and action schemas so that change is perceived to be incremental rather than discontinuous, they enable more radical changes than would otherwise be possible’ (see also Armstrong, Reference Armstrong2005). The argument that a dramatic breakdown of institutionalized ideas almost never occurs – because actors draw on the existing set of ideas to get hold of new ideas – thus seems relevant in situations of extreme uncertainty and complexity, that is, crisis. The continual re-interpretation of existing ideas and institutions that characterizes bricolage thus translates into a perspective on change that emphasizes neither strong stability nor revolutionary change. Rather, the focus is on the incremental change that in times of both crisis and general stability follows when bits and pieces from several institutional and ideational legacies are blended to answer to the continual stream of challenges the political system faces.

To sum up, Table 1 provides an overview of the central differences between the bricoleur and actors working inside paradigms.

Table 1 Paradigm man vs. the political bricoleur