Taking ideas and discourse seriously 3 1995)or on 'advocacy coalitions'(Sabatier and Jenkins-Smith,1993).Yet others have resisted positing a fourth such institutionalism(e.g.,Campbell (2004),but see Campbell and Pedersen (2001)),mainly because they seek to blur the boundaries among the three older institutionalisms in order to show how ideas and discourse can serve to advance knowledge in the social sciences across methodological approaches.For our purposes,which is to demonstrate how the turn to ideas and discourse can take us beyond the limits of the three older neo- institutionalisms in order to explain the dynamics of change(and continuity),it is more useful to identify the commonalities in all approaches that focus on ideas and discourse as distinct from rationalist interests,path-dependent history,and cultural framing. Discursive institutionalism is an umbrella concept for the vast range of works in political science that take account of the substantive content of ideas and the interactive processes by which ideas are conveyed and exchanged through dis- course.On the substantive dimension of ideas and discourse,DI scholars consider ideas about 'what is and what ought to be'at different levels of generality (Schmidt,2008;Mehta,2010),going from policy ideas (e.g.,Kingdon,1984; Hall,1989)to programmatic ideas or paradigms(Hall,1993;Berman,1998)to deeper philosophical ideas(Campbell,2004).They also consider different types of ideas,including cognitive ideas justified in terms of interest-based logics and necessity (e.g.,Jobert,1989;Hall,1993;Schmidt,2002:Ch.5)and normative ideas legitimated through appeal to values and appropriateness(e.g.,March and Olsen,1989;Schmidt,2000).And they consider the representation of ideas through discourse,including frames,narratives,myths,collective memories, stories,scripts,and more (e.g.,Roe,1994;Hajer,2003). On the interactive dimension,DI scholars consider the discursive processes by which such ideas are constructed in a 'coordinative'policy sphere and deliberated in a 'communicative'political sphere (Schmidt,2000,2002).The coordinative discourse encompasses the wide range of policy actors engaged in the construction of policy ideas.They may be organized in 'epistemic communities'of elites with shared ideas (Haas,1992),'advocacy coalitions'of elites with shared ideas and policy access(Sabatier and Jenkins-Smith,1993),and'advocacy networks'of activists contesting ideas in international politics (Keck and Sikkink,1998).Or they may act as 'entrepreneurs'(Fligstein and Mara-Drita,1996)and 'mediators' (Jobert,1989;Muller,1995)who serve as catalysts for the ideas of such discursive communities. The communicative discourse encompasses the wide range of political actors who bring the ideas developed in the context of the coordinative discourse to the public for deliberation and legitimation.These actors may include political lea- ders involved in the top-down mass electoral process of public persuasion (see e.g.,Zaller,1992;Mutz et al.,1996),in public debates (Art,2006)or in the 'policy forums'of 'informed publics'(Rein and Schon,1991 engaged in 'com- municative action'(Habermas,1996).Or they may be members of civil society1995) or on ‘advocacy coalitions’ (Sabatier and Jenkins-Smith, 1993). Yet others have resisted positing a fourth such institutionalism (e.g., Campbell (2004), but see Campbell and Pedersen (2001)), mainly because they seek to blur the boundaries among the three older institutionalisms in order to show how ideas and discourse can serve to advance knowledge in the social sciences across methodological approaches. For our purposes, which is to demonstrate how the turn to ideas and discourse can take us beyond the limits of the three older neo￾institutionalisms in order to explain the dynamics of change (and continuity), it is more useful to identify the commonalities in all approaches that focus on ideas and discourse as distinct from rationalist interests, path-dependent history, and cultural framing. Discursive institutionalism is an umbrella concept for the vast range of works in political science that take account of the substantive content of ideas and the interactive processes by which ideas are conveyed and exchanged through dis￾course. On the substantive dimension of ideas and discourse, DI scholars consider ideas about ‘what is and what ought to be’ at different levels of generality (Schmidt, 2008; Mehta, 2010), going from policy ideas (e.g., Kingdon, 1984; Hall, 1989) to programmatic ideas or paradigms (Hall, 1993; Berman, 1998) to deeper philosophical ideas (Campbell, 2004). They also consider different types of ideas, including cognitive ideas justified in terms of interest-based logics and necessity (e.g., Jobert, 1989; Hall, 1993; Schmidt, 2002: Ch. 5) and normative ideas legitimated through appeal to values and appropriateness (e.g., March and Olsen, 1989; Schmidt, 2000). And they consider the representation of ideas through discourse, including frames, narratives, myths, collective memories, stories, scripts, and more (e.g., Roe, 1994; Hajer, 2003). On the interactive dimension, DI scholars consider the discursive processes by which such ideas are constructed in a ‘coordinative’ policy sphere and deliberated in a ‘communicative’ political sphere (Schmidt, 2000, 2002). The coordinative discourse encompasses the wide range of policy actors engaged in the construction of policy ideas. They may be organized in ‘epistemic communities’ of elites with shared ideas (Haas, 1992), ‘advocacy coalitions’ of elites with shared ideas and policy access (Sabatier and Jenkins-Smith, 1993), and ‘advocacy networks’ of activists contesting ideas in international politics (Keck and Sikkink, 1998). Or they may act as ‘entrepreneurs’ (Fligstein and Mara-Drita, 1996) and ‘mediators’ (Jobert, 1989; Muller, 1995) who serve as catalysts for the ideas of such discursive communities. The communicative discourse encompasses the wide range of political actors who bring the ideas developed in the context of the coordinative discourse to the public for deliberation and legitimation. These actors may include political lea￾ders involved in the top-down mass electoral process of public persuasion (see e.g., Zaller, 1992; Mutz et al., 1996), in public debates (Art, 2006) or in the ‘policy forums’ of ‘informed publics’ (Rein and Scho¨n, 1991 engaged in ‘com￾municative action’ (Habermas, 1996). Or they may be members of civil society Taking ideas and discourse seriously 3