ANRV344-PL11-14 ARI 8 April 2008 15:11

Discursive Institutionalism:The Explanatory Power ofIdeas and DiscourseVivien A. SchmidtDepartment of International Relations, Boston University, Boston,Massachusetts 02215; email: vschmidt@bu.edu

Annu. Rev. Polit. Sci. 2008. 11:303–26

First published online as a Review in Advance onJanuary 22, 2008

The Annual Review of Political Science is online athttp://polisci.annualreviews.org

This article’s doi:10.1146/annurev.polisci.11.060606.135342

Copyright c© 2008 by Annual Reviews.All rights reserved

1094-2939/08/0615-0303$20.00

Key Words

institutions, institutional change, interests, norms, newinstitutionalism

AbstractThe newest “new institutionalism,” discursive institutionalism, lendsinsight into the role of ideas and discourse in politics while providinga more dynamic approach to institutional change than the older threenew institutionalisms. Ideas are the substantive content of discourse.They exist at three levels—policies, programs, and philosophies—and can be categorized into two types, cognitive and normative.Discourse is the interactive process of conveying ideas. It comesin two forms: the coordinative discourse among policy actors andthe communicative discourse between political actors and the pub-lic. These forms differ in two formal institutional contexts; simplepolities have a stronger communicative discourse and compoundpolities a stronger coordinative discourse. The institutions of dis-cursive institutionalism, moreover, are not external-rule-followingstructures but rather are simultaneously structures and constructsinternal to agents whose “background ideational abilities” within agiven “meaning context” explain how institutions are created andexist and whose “foreground discursive abilities,” following a “logicof communication,” explain how institutions change or persist. In-terests are subjective ideas, which, though real, are neither objectivenor material. Norms are dynamic, intersubjective constructs ratherthan static structures.

303

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

INTRODUCTIONThe turn to ideas and discourse in politicalscience has come to constitute a fourth “newinstitutionalism.” I call it discursive institu-tionalism (DI), distinct from rational choiceinstitutionalism (RI), historical institutional-ism (HI), and sociological institutionalism(SI). Political scientists whose work fits theDI rubric tend to have four things in com-mon. First, they take ideas and discourse seri-ously, even though their definitions of ideasand uses of discourse vary widely. Second,they set ideas and discourse in institutionalcontext, following along the lines of one oranother of the three older new institution-alisms, which serve as background informa-tion. Third, they put ideas into their “meaningcontext” while they see discourse as follow-ing a “logic of communication,” despite dif-ferences in what may be communicated howand where. Finally, and most importantly, theytake a more dynamic view of change, in whichideas and discourse overcome obstacles thatthe three more equilibrium-focused and staticolder institutionalisms posit as insurmount-able. What most clearly differentiates discur-sive institutionalists from one another is nottheir basic approach to ideas and discoursebut rather the kinds of questions they ask andthe problems they seek to resolve, which tendto come from the institutionalist tradition(s)with which they engage.

Although political scientists have beenexploring the explanatory power of ideas anddiscourse for a while now, the term used todefine this approach, discursive institution-alism, is of very recent vintage (see Schmidt2002a, 2006a,b). Although others have usedthe same term (see Campbell & Pedersen2001)—or similar ones, such as ideationalinstitutionalism (Hay 2001), constructivistinstitutionalism (Hay 2006), or strategic con-structivism ( Jabko 2006)—they have tendedto focus much more on the ideas that are thesubstantive content of discourse than on theinteractive processes involved in discourse.In addition, not all scholars who have turned

to ideas and discourse go so far as to posita fourth institutionalism (e.g., Campbell2004—but see Campbell & Pedersen 2001).This is mainly because their purpose is toblur the boundaries among all three olderinstitutionalisms, and to show how ideas anddiscourse can advance knowledge in the socialsciences across methodological approaches.This is a worthy goal, and one I share. But Ithink it necessary also to recognize the distinc-tiveness of approaches that focus on ideas anddiscourse—even though discursive institu-tionalists often speak less to one another thanto those who sit in the older “new institution-alism” in which they themselves have roots.

Within DI, moreover, although politicalscientists in recent years have generated lotsof ideas about ideas, they have engaged incomparatively little discourse about discourse.Why the turn to ideas? Why the reticence ondiscourse?

For many political scientists, the turn toideas has been a useful corrective to the lim-its of new institutionalist approaches and atacit acknowledgment of their difficulties inexplaining change. Importantly, large num-bers of new institutionalists, whether rationalchoice, historical, or sociological institution-alists, have sought to use ideas to counter thestatic and overly deterministic nature of in-stitutions in their explanations. The tippingpoint between those approaches to ideas thatremain within the confines of any one of thethree older new institutionalisms and thosethat belong to DI is fuzzy, but we can situ-ate it at the point at which the turn to ideasundermines the basic premises of the oldernew institutionalism, i.e., that institutions arein stable equilibria, with fixed rationalist pref-erences (RI), self-reinforcing historical paths(HI), or all-defining cultural norms (SI).

The reluctance of many of these same po-litical scientists to add discourse to their con-sideration of ideas stems primarily from pastusage of the term. “Discourse” conjures up ex-aggerated visions of postmodernists and post-structuralists who are assumed (often unfairly)

304 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

to interpret “texts” without contexts and tounderstand reality as all words, whatever thedeeds. But without using some term like “dis-course,” that is, a term that refers to talk-ing about one’s ideas, how does one discussthe process of putting one’s ideas across? Dis-course, as defined herein, is stripped of post-modernist baggage to serve as a more genericterm that encompasses not only the substan-tive content of ideas but also the interactiveprocesses by which ideas are conveyed. Dis-course is not just ideas or “text” (what is said)but also context (where, when, how, and whyit was said). The term refers not only to struc-ture (what is said, or where and how) but alsoto agency (who said what to whom).

But if the great innovation of DI is its abil-ity to explain change and continuity, then themain question is: How does it do so? And moregenerally, what is the explanatory power ofideas and discourse?

The first half of the article examines thewide range of approaches to ideas and dis-course in political science, without, at thisstage, differentiating among them in terms ofinstitutionalist tradition. First, I identify ideasboth in terms of their levels of generality (poli-cies, programs, and philosophies) and type ofcontent (cognitive and normative). Second, Idiscuss the two basic forms of discourse: thecoordinative discourse among policy actorsand the communicative discourse between po-litical actors and the public. Throughout thissection, I consider the attributes of successfulideas and discourse, along with the methodsthat serve to demonstrate their transformativepower and, thereby, their causal influence.

The second part of the article sets ideasand discourse into “new institutionalist” per-spective by contrasting DI with the threeolder new institutionalisms. Points of con-trast include definitions of institutions and in-stitutional change, interests and uncertainty,and norms and relativism. First, we defineinstitutions in DI as simultaneously struc-tures and constructs internal to agents whose“background ideational abilities” and “fore-ground discursive abilities” make for a more

dynamic, agent-centered approach to institu-tional change than in HI. Next, we show thatinterests in DI are “subjective” rather than ei-ther “objective” or “material,” as in RI, butnonetheless “real.” Third, we show that al-though DI has much in common with SI,norms in DI are more dynamic constructs. Weend with a discussion of how to conceive ofthe relationship of DI to the other three newinstitutionalisms.

My overall argument is that DI is a dis-tinctive approach that contributes to our un-derstanding of political action in ways thatthe older three institutionalisms cannot. Atthe very least, it adds another institutionalistapproach to our methodological toolkit. Buteven more than this, it provides insight intoan area of political action that political sci-entists have long neglected, largely becausethey could not account for it within the lim-its of their own methodological approaches.The result is that they have ignored some ofthe biggest questions in politics, the questionsthat political philosophers through the ageshave puzzled over, such as the role of ideasin constituting political action, the power ofpersuasion in political debate, the centralityof deliberation for democratic legitimation,the construction and reconstruction of polit-ical interests and values, and the dynamics ofchange in history and culture. Moreover, theyhave passed up the opportunity to weigh inon the substantive issues of political life, leav-ing to journalists and think-tanks the battleof ideas with regard to the policy questionsof the day. To policy makers and politiciansin particular, the very notion that one wouldneed to make a plea for taking ideas and dis-course seriously would appear ludicrous, be-cause the very essence of what they do is togenerate ideas about what should be done andthen communicate them to the general pub-lic for discussion and deliberation. This essayon DI, in short, takes it as a given that ideasand discourse matter, in order to focus on themore interesting set of questions for politicalscientists, namely how, when, where, and whyideas and discourse matter.

www.annualreviews.org • Discursive Institutionalism 305

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

THE EXPLANATORY POWEROF IDEAS AND DISCOURSE

The difference between scholars who use theterm discourse and those who limit them-selves to ideas is primarily one of empha-sis. Those scholars who focus exclusively onideas tend to leave the interactive processesof discourse implicit as they discuss the ideasgenerated, deliberated, and legitimated bypublic actors, the carriers of ideas. Thosescholars who speak of discourse address ex-plicitly the representation of ideas (how agentssay what they are thinking of doing) and thediscursive interactions through which actorsgenerate and communicate ideas (to whomthey say it) within given institutional contexts(where and when they say it). But whetherthey emphasize ideas or discourse, such schol-ars employ a range of methods to demon-strate the transformative power of ideas anddiscourse, that is, to show how they exerta causal influence in political reality and,thereby, engender institutional change (orcontinuity).

The Nature of Ideas

Defining ideas, the substantive content ofdiscourse, is no easy task because there areso many ideas about ideas (see Goodin &Tilly 2005, Pt. IV). Ideas have been seenas switches for interests, road maps, or fo-cal points (Goldstein & Keohane 1993); asstrategic constructions ( Jabko 2006) or strate-gic weapons in the battle for control (Blyth2002); as narratives that shape understandingsof events (e.g., Roe 1994) or as “frames of ref-erence” ( Jobert 1989, Muller 1995); and ascollective memories (Rothstein 2005) or na-tional traditions (Katzenstein 1996).

Political scientists’ uses of ideas tend tooccur at three main levels of generality (seeJ. Mehta, unpublished manuscript). The firstlevel encompasses the specific policies or “pol-icy solutions” proposed by policy makers.The second level encompasses the more gen-eral programs that underpin the policy ideas.These may be cast as paradigms that re-

flect the underlying assumptions or organiz-ing principles orienting policy (Majone 1989;Hall 1993; Schmidt 2002a, ch. 5); as framesof reference (referentiels) that enable policyactors to (re)construct visions of the worldthat allow them to (re)situate themselves inthe world ( Jobert l989, Muller l995); as“programmatic beliefs” (Berman 1998) thatoperate in the space between worldviewsand specific policy ideas; as “policy cores”that provide sets of diagnostics and prescrip-tions for action (Sabatier & Jenkins-Smith1993); or as “problem definitions” that set thescope of possible solutions to the problemsthat policy ideas address (Mehta, unpublishedmanuscript). These programmatic ideas are ata more basic level than the policy ideas be-cause they define the problems to be solvedby such policies; the issues to be considered;the goals to be achieved; the norms, methods,and instruments to be applied; and the idealsthat frame the more immediate policy ideasproposed to solve any given problem. At aneven more basic level are the “public philoso-phies” (Campbell 1998), “public sentiments”(Campbell 2004), or “deep core” (Sabatier &Jenkins-Smith 1993)—worldviews or Weltan-schauung that undergird the policies and pro-grams with organizing ideas, values, and prin-ciples of knowledge and society. Whereasboth policy ideas and programmatic ideas canbe seen as foreground, since these tend to bediscussed and debated on a regular basis, thephilosophical ideas generally sit in the back-ground as underlying assumptions that arerarely contested except in times of crisis (seeCampbell 2004, pp. 93–94).

Policies, programs, and philosophies tendto contain two types of ideas: cognitive andnormative. Cognitive ideas elucidate “what isand what to do,” whereas normative ideas in-dicate “what is good or bad about what is” inlight of “what one ought to do.” Cognitiveideas—also sometimes called causal ideas—provide the recipes, guidelines, and maps forpolitical action and serve to justify policies andprograms by speaking to their interest-basedlogic and necessity (see Jobert 1989; Hall

306 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

1993; Schmidt 2002a, ch. 5). Cognitive ideasspeak to how (first level) policies offer so-lutions to the problems at hand, how (sec-ond level) programs define the problems to besolved and identify the methods by which tosolve them, and how both policies and pro-grams mesh with the deeper core of (thirdlevel) principles and norms of relevant scien-tific disciplines or technical practices. Norma-tive ideas instead attach values to political ac-tion and serve to legitimate the policies in aprogram through reference to their appropri-ateness (see March & Olsen 1989). Normativeideas speak to how (first level) policies meetthe aspirations and ideals of the general pub-lic and how (second level) programs as wellas (first level) policies resonate with a deepercore of (third level) principles and norms ofpublic life, whether the newly emerging val-ues of a society or the long-standing ones inthe societal repertoire (Schmidt 2000; 2002a,ch. 5).

The big question for scholars of ideas iswhy some ideas become the policies, pro-grams, and philosophies that dominate politi-cal reality while others do not. The standardsand criteria they propose for evaluating ideastend to differ according to level.

For the first level of ideas, scholars iden-tify a range of purely political scientific fac-tors that help explain why specific policiesmay succeed and why they change. On pol-icy success, the main question for scholarsis: What specific criteria ensure the adoptionof a given policy? Hall (1989) speaks of theneed for policy ideas to have administrativeand political viability in addition to policy vi-ability; Kingdon (1984) argues that policiesmust come together with the other two crit-ical streams of problems and politics for apolicy idea to be adopted. Other ideationalfactors at play include the role of nationaltraditions in making a policy more or lessacceptable, as when state identities structurenational perceptions of defense and securityissues (Katzenstein l996), and the role of na-tional values and political culture in the adop-tion of transnational policy ideas, such as the

very different ways in which the word “precar-iousness” is understood and used (or not) inGermany and the United Kingdom by con-trast with France, Italy, and Spain (Barbier2004). Equally important is the matter of ex-pertise linked to the validation of ideas by re-search institutes and think-tanks (Rich 2004;J.L. Campbell & O. Pedersen, unpublishedmanuscript). The element of timing is also afactor in policy success, which helps explainwhy Scandinavian welfare states remain dis-tinct (Cox 2001). Another factor is genera-tional turnover, although this cannot accountfor the fact that certain ideas may persist fromone generation to the next, as in Austria andJapan with regard to World War II, while oth-ers may shift radically within a generation, asin Germany in the 1980s (Art 2006).

But although these criteria all help iden-tify the necessary factors for policy adoption,they cannot delineate the sufficient factors, inparticular those things that don’t get onto theagenda, since the selection bias of most suchstudies is toward successful ideas (see discus-sion by J. Mehta, unpublished manuscript).Moreover, such criteria often do little to spec-ify the ideational processes by which old ideasfade and new ideas come to the fore. And fi-nally, studies of policy ideas and discourse tendto have a built-in bias that seems to assumethat “good” ideas—meaning those that appearmore relevant to the problem at hand, moreadequate to the task, and more appropriateto the needs of society—succeed while “bad”ideas fail. But, in fact, sometimes good ideasfail and bad ideas succeed. How to respondto all of these issues? For answers we need togo on to the second and third levels of ideas,since scholars who focus on programmatic andphilosophical ideas tend to offer more generaltheories about ideational success and changeover time.

Scholars who concentrate on the secondlevel of ideas often look to the philosophyof science for the criteria that would ex-plain success and change in programs andthe policy ideas that emerge from them(e.g., Jobert 1989; Majone 1989; Hall 1993;

www.annualreviews.org • Discursive Institutionalism 307

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

Schmidt 2002a, pp. 222–25). These schol-ars generally liken programmatic ideas to the“paradigms” of Kuhn (1970) or the “researchprograms” of Lakatos (1970), and they linksuccess not only to the viability of a program’spolicy ideas but also to the program’s long-term problem-solving potential. Thus, theymay describe the revolutionary “third-order”change occurring in the United Kingdom un-der Prime Minister Thatcher, who replacedthe paradigm of Keynesianism with mone-tarism (Hall 1993), and the similar revolu-tionary change wrought in France by Mit-terrand with the “great U-turn” in macroe-conomic policy, by contrast with Blair’s “re-newal” of Thatcher’s neoliberal paradigm andSchroder’s attempt to “recast” the paradigmof the social market economy in Germany(Schmidt 2002a, ch. 6).

The use of the philosophy of science canonly go so far, however (see Schmidt 2002a,pp. 217–25). In science, programmatic suc-cess is judged by scientists alone; in society,programmatic success is judged not only bysocial scientists but also by citizens. The suc-cess of a program does not just depend onthe presence of cognitive ideas capable of sat-isfying policy makers that a given programwill provide robust solutions. It also dependson the presence of complementary normativeideas capable of satisfying policy makers andcitizens alike that those solutions also servethe underlying values of the polity. Moreover,whereas ideational change in science resultsfrom internal processes, when the Kuhnianparadigm expires because it has exhausted itsexplanatory potential, ideational change in so-cial science and society results also from exter-nal processes and events that create a receptiveenvironment for new ideas.

The difficulties in establishing criteria forfirst- and second-level ideas, that is, for poli-cies and programs, are even greater once weturn to the third level of ideas—the philoso-phies that underlie policies and programs. Be-cause these ideas are at a deeper level than theothers, and often left unarticulated as back-ground knowledge, it is difficult to prove that

a particular set of ideas constitutes a publicphilosophy. It is even more difficult to traceover time the development of one philosophyand its eventual replacement by another. Theidentification of such public philosophies hasoften been the domain of macrosociologists.The most notable is of course Max Weber,whose work on the ideas predominant in so-ciety has inspired numerous political scien-tists. More recent is the work of Bourdieu(1994), Foucault (2000), and Gramsci (1971),who present public philosophies as the ideasof the powerful who dominate society. Andyet it is often the case in a given society that,at a very basic level, “everyone knows” whatthe basic philosophy or worldview is, even ifthey may not be able to define it preciselyor describe how it developed or changed.This is why political scientists also often usemethods based on comparative case studiesand “process-tracing”—methods that demon-strate how such ideas are tied to action. Theseideas serve as guides to public actors for whatto do, as well as being the sources of justifica-tion and legitimation for what such actors do(see Berman 1998, 2006; Blyth 2002).

The literature includes several outstand-ing examples of such case studies andprocess-tracing. Hall’s (1989) edited volumeilluminates the philosophical as well as pro-grammatic reasons why advanced, industrial-ized countries on both sides of the Atlanticdid (or did not) adopt Keynesian economicideas. An edited volume by Dyson (2002) elu-cidates why European Union (EU) memberstates adopted or rejected the euro. Dobbin’s(1994) study concluded that the differing un-derlying philosophical ideas about the roleof the state in the economy ensured thatthe building of the railroads was state-led inFrance whereas in the United States it wasled by private actors. Berman (1998) draws ahistorical contrast between the German So-cial Democrats, who capitulated to Nazismlargely because they could not think beyondtheir long-held Marxist ideas, and the SwedishSocial Democrats, who succeeded not only infighting fascism but also in creating a social

308 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

democratic state because they were free ofany such ideational legacy and able to rein-vent socialism. McNamara’s (1998) accountof European monetary union posits a three-step learning process: first, policy failure; sec-ond, the search for new ideas that led toa neoliberal consensus on monetarism; andthird, the adoption of the German exem-plar. Blyth (2002) analyzes the role of foun-dational economic ideas at moments of eco-nomic crisis in Sweden and the United States,first in “embedding” liberalism in the 1930sand then “disembedding” it beginning in the1970s.

Another method involves directly address-ing the causal influence of discourse. Somescholars (e.g., Berman 1998, pp. 16–19; Blyth2002, Parsons 2003) show that ideas can be anindependent variable by demonstrating thatno other structural factors can account forthe clear changes (or continuities) in inter-ests, paths, or norms signaled by political ac-tors’ expressed ideas and intended actions. Inparticular, they rule out those structural fac-tors that follow from the three older new in-stitutionalisms: narrowly defined rationalistinterests, historical path dependencies, andcultural norms. For example, Parsons (2003)first shows why other new institutionalistaccounts of European integration cannot ex-plain outcomes before he traces the pro-cesses by which French leaders’ ideas anddiscourse about constructing the institutionsof the European Union became the insti-tutionalized ideas that constrained subse-quent French leaders’ ideas, discourse, andactions.

Despite the problems, then, there are a va-riety of ways in which political scientists es-tablish ideational success. But this leaves uswith one fundamental problem. We still haveno way of considering the process by whichsuch ideas go from thought to word to deed,that is, how ideas are conveyed, adopted, andadapted, let alone the actors who convey themto whom, how, where, and why. This raisesthe question of agency, which brings us to theconcept of discourse.

The Dynamics of Discourse

Discourse is a more versatile and overarch-ing concept than ideas. By using the termdiscourse, we can simultaneously indicate theideas represented in the discourse (which maycome in a variety of forms as well as content)and the interactive processes by which ideasare conveyed (which may be carried by differ-ent agents in different spheres). The discur-sive processes alone help explain why certainideas succeed and others fail because of theways in which they are projected to whom andwhere. But the discourse itself, as representa-tion as well as process, also needs to be evalu-ated as to why it succeeds or fails in promotingideas. It is therefore a pity that political scien-tists have largely avoided the term because ofits original uses in postmodern literary crit-icism and philosophy, and stick to “ideas” intheir own discourse even when their own ideasare also about discourse.

In the representation of ideas, any givendiscourse may serve to articulate not only dif-ferent levels of ideas (policy, programmatic,and philosophical; see Hajer 2003) and differ-ent types of ideas (cognitive and normative)but also different forms of ideas—narratives,myths, frames, collective memories, stories,scripts, scenarios, images, and more. The“terms” of the discourse, in Connolly’s (1983)sense of “institutionalized structures of mean-ing that channel political thought and ac-tion in certain directions,” are multiple, pat-terning how arguments are made as well aswhich ideas are represented. Moreover, dis-course may intersperse technical and scien-tific arguments with more generally accessiblenarratives that fit together the specialists’ ar-guments with accounts of events, emblematiccases, and even doomsday scenarios to gen-erate compelling stories about the causes ofcurrent problems, what needs to be done toremedy them, and how they fit with the under-lying values of the society (see Schmidt 2002a,ch. 5). In addition, discourse can be highlyvaried in its use of ideas. For example, in thecase of European market integration, Jabko

www.annualreviews.org • Discursive Institutionalism 309

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

(2006, ch. 3) shows that actors used a reper-toire of “strategic ideas” of the market topresent it (a) as a constraint and inescapablereality in the financial arena, (b) as a norm tobe desired in the energy sector, (c) as a spacein structural policies focused on regional eco-nomic development, and (d ) as a symbol of anew source of discipline in the Economic andMonetary Union.

Discourse is not only what you say, how-ever; it includes to whom you say it, how,why, and where in the process of policy con-struction and political communication in the“public sphere” (see Habermas 1989, 1996).Political scientists tend to focus on a partic-ular part of the public sphere; some investi-gate the policy sphere, in which policy actorsengage one another in a “coordinative” dis-course about policy construction, while otherslook at the political sphere, in which politicalactors engage the public in a “communicative”discourse about the necessity and appropri-ateness of such policies (see Schmidt 2002a,ch. 5; 2005).

In the policy sphere, the coordinative dis-course consists of the individuals and groupsat the center of policy construction who areinvolved in the creation, elaboration, and jus-tification of policy and programmatic ideas.These are the policy actors—the civil ser-vants, elected officials, experts, organized in-terests, and activists, among others—who seekto coordinate agreement among themselveson policy ideas, which scholars have shownthey may do in a variety of ways in a widerange of venues. Thus, the coordinative dis-course may be the domain of individualsloosely connected in “epistemic communi-ties” in transnational settings on the basis ofshared cognitive and normative ideas abouta common policy enterprise (Haas 1992). Al-ternatively, it may consist of more closely con-nected individuals who share both ideas andaccess to policy making. Examples include“advocacy coalitions” in localized policy con-texts, as in water policy in California (Sabatier& Jenkins-Smith 1993); “discourse coalitions”in national settings across extended time pe-

riods, as in the ideas of “ordo-liberalism” thatunderpinned Germany’s postwar social mar-ket economy (Lehmbruch 2001); and “advo-cacy networks” of activists in internationalpolitics focused on issues of human rights,the environment, or violence against women(Keck & Sikkink 1998). But the coordinativediscourse may also contain “entrepreneurs”(Fligstein & Mara-Drita 1996, Finnemore &Sikkink 1998) or “mediators” ( Jobert 1989,Muller 1995) who serve as catalysts for changeas they draw on and articulate the ideas of dis-cursive communities and coalitions.

The communicative discourse occurs inthe political sphere. It consists of the individ-uals and groups involved in the presentation,deliberation, and legitimation of politicalideas to the general public. In a mass processof public persuasion, political leaders, gov-ernment spokespeople, party activists, “spindoctors,” and more communicate the policyideas and programs developed in the coordi-native discourse to the public for discussionand deliberation (see, e.g., Mutz et al. l996).But the communicative discourse encom-passes other political actors as well, includingmembers of opposition parties, the media,pundits, community leaders, social activists,public intellectuals, experts, think-tanks,organized interests, and social movements.These and other actors, often organized inthe “policy forums” of “informed publics”(Rein & Schon 1994) and the “public oforganized private persons” (Habermas 1989)as well as in the “strong publics” of oppositionparties, members of legislatures, and politicalcommentators (Eriksen & Fossum 2002),communicate their responses to governmentpolicies, engendering debate, deliberation,and ideally, modification of the policies underdiscussion. Finally, the general public of citi-zens and voters to whom this communicativediscourse is directed also contribute to it.As members of civil society, they engage ingrass-roots organizing, social mobilization,and demonstrations; as members of “mini-publics,” they express themselves in citizenjuries, issues forums, deliberative polls, and

310 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

the like (see Goodin & Dryzek 2006); and asmembers of the electorate, their voices areheard in opinion polls, surveys, focus groups,and, of course, elections—where actionsspeak louder than words.

The arrows of discursive interaction of-ten appear to go from the top down. Pol-icy elites generate ideas, which political elitesthen communicate to the public. Politi-cal elites often interweave the coordinativeand communicative discourses into a masterdiscourse that presents an at least seeminglycoherent political program. The master dis-course provides a vision of where the polity is,where it is going, and where it ought to go.The political elites then mediate the ensuingpublic debates. There is an extensive litera-ture on how elites shape mass public opinionby establishing the terms of the discourse andby framing the issues for the mass media and,thereby, for the mass public (e.g., Zaller 1992;see discussion in Art 2006, ch. 2). The arrowscan also go from the bottom up, however, as inthe discursive interactions of social activists,feminists, and environmentalists in nationaland international arenas (e.g., Keck & Sikkink1998). The arrows can even remain solely atthe level of civil society, in “public conver-sations” (Benhabib 1996), communicative ac-tion in the public sphere (Habermas 1989), or“deliberative democracy” in the supranationalsphere (Dryzek 1990, 2000).

Equally important, there may be no ar-rows between coordinative and communica-tive discourses. Coordinative policy ideas mayremain in closed debates out of public view,either because they might not be approved—as in the case of some of the more progres-sive immigration policy reforms (Guiraudon1997)—or because the issues are too techni-cal to capture the sustained interest of thepublic, as in the case of banking reforms(Busch 2004). The lack of connection be-tween spheres of discourse is a frequent oc-currence in the European Union (Schmidt2006a, ch. 1). As Howorth (2004) notes, theconcept of a European army was essentiallyaccepted by Prime Minister Blair in the co-

ordinative EU and national discourses, butthe “army” label was denied in his commu-nicative discourse once Fleet Street raised thealarm. Finally, because public debates cannotbe controlled by any one political actor or setof actors, even when a discourse starts fromthe top it very often escapes political leaders’control. In the case of Germany, for exam-ple, Art (2006) shows that when conservativeChancellor Kohl sought to “normalize” ideasabout the country’s Nazi past, the debate helaunched quickly became an opportunity forall manner of political actors to weigh in onthe issues, ultimately ensuring that the dis-course initiated by the left became the basisfor a “political correctness, German style” thatsilenced potential antisemitic and right-wingextremist speech.

Tracing discursive processes of coordina-tion and communication is a way of showingwhy ideas may succeed or fail. But discourse,like ideas, sometimes matters to that successand sometimes does not. There are, after all,always ideas and discourse, most of which tendto reinforce existing realities and only someof which promote change. When does dis-course exert a causal influence by promotingchange—first, in terms of its representation ofideas, and second, as the discursive process bywhich it conveys those ideas?

Discourse contributes to the success orfailure of ideas first of all by how it articu-lates their substantive content. What makesfor a successful discourse, in fact, encompassesa lot of the same things that make for success-ful ideas: relevance to the issues at hand, ade-quacy, applicability, appropriateness, and res-onance. But beyond this, the credibility of adiscourse is likely to benefit from consistencyand coherence across policy sectors, althougha modicum of vagueness or ambiguity is alsoto be expected (see Radaelli & Schmidt 2004).Vagueness especially helps in the context ofinternational diplomacy, when the same dis-course can be read in radically different ways,as in the case of Britain and France with regardto the concept of a European army (Howorth2004). The coherence of a discourse can add

www.annualreviews.org • Discursive Institutionalism 311

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

to its strength even when it uses ideas fordifferent strategic purposes in different con-texts. An example is the “double discourse”of the European Union Commission on agri-cultural policy, in which the discourse of “mul-tifunctionality” was used at the internationallevel to defend the EU from outside pres-sures for change and at the EU level to pushfor member-state reform (Fouilleux 2004).However, expectations of consistency and co-herence can lead to “rhetorical entrapment”(Schimmelfennig 2001): Political actors whodo not change their preferences neverthelessfeel obliged to follow the policy implicationsof discourses they have accepted in the past.This demonstrates that discourse is a lot morethan talk. It can not only commit the speakersthemselves to action, it can also constrain theideas, discourse, and actions of their succes-sors. Thus, not only did French political lead-ers find themselves having to honor their pre-decessors’ commitments, as noted above (seeParsons 2003), but they also found themselvestrapped by their predecessors’ communicativediscourse, in particular de Gaulle’s initial le-gitimating ideas about European integration(Schmidt 2007b).

The interactive processes of discourse mayalso exert a causal influence beyond what dis-course does in representing ideas. Most gen-erally, discourse serves not just to express oneset of actors’ strategic interests or norma-tive values but also to persuade others of thenecessity and/or appropriateness of a givencourse of action. Some, following Habermas(1989, 1996), see a need to distinguish “argu-ing,” which involves persuasion, from “bar-gaining,” which is a strategic action (e.g., Risse2000). Although this is an evocative distinc-tion, most discursive interactions actually in-volve both arguing and bargaining; one canargue to defend one’s interests while beingstrategic in persuading others as to the appro-priateness of one’s viewpoint (see Holzinger2004, Radaelli & Schmidt 2004). Equally im-portant, although discourses are most oftensuccessful if true, coherent, and consistent,they need not be any of these things. Suc-

cessful discourses may be manipulative, theymay lie, they may be “happy talk” or “spin”to obscure what political leaders are reallydoing, and they may even be vehicles for elitedomination and power, as Bourdieu (1990),Foucault (2000), and Gramsci (1971) argue.But this is where public debates in demo-cratic societies come in. They can expose thebad ideas of the particular discourse of anypolitical actor or set of actors. For example,Art (2006) demonstrates the causal influenceof public debates when he links the failureof the extreme right in Germany to the far-ranging public debates, together with protestsand social mobilization, that isolated and dele-gitimated extreme right parties, and links itssuccess in Austria to the lack of any such exten-sive debate or social action. Political leaders’discourse alone, however, can have a majorimpact, as I show in a matched pair of casesin which all factors are controlled for otherthan the discourse. Lasting public acceptancefor neoliberal reform in the United Kingdomwas due in large measure to the communica-tive discourse through which Prime Minis-ter Thatcher sought to persuade the publicof what she believed as she reformed; its lackof acceptance in New Zealand had much todo with the lack of communicative discourseof political leaders beginning with FinanceMinister Douglas, who assumed that peoplewould come to believe what he believed afterhe reformed (Schmidt 2000, 2002b).

The formal institutional context also hasan impact on where and when discourse maysucceed. Different forms of discourse may beemphasized in different institutional settings.In “simple” polities, where governing activ-ity tends to be channeled through a singleauthority—primarily countries with majori-tarian representative institutions, statist pol-icy making, and unitary states such as Britainand France—the communicative discourse tothe general public tends to be much moreelaborate than the coordinative discourseamong policy actors. This is because gov-ernments that tend to impose policies with-out much consultation with the most affected

312 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

interests face sanctions ranging from interestgroup protest to loss of public confidence andloss of elections if they fail to provide a suffi-ciently legitimating communicative discourseabout those policies to the public. By contrast,in “compound” polities, where governing ac-tivity tends to be dispersed among multipleauthorities—countries with proportional rep-resentation systems, corporatist policy mak-ing, and/or federal or regionalized states suchas Germany and Italy—the coordinative dis-course among policy actors tends to be muchmore elaborate than the communicative dis-course to the public. This is because it isdifficult to communicate in anything morethan vague terms to the public the results ofthe negotiations among the many policy ac-tors involved without jeopardizing any of thecompromises they make in private (Schmidt2002a, pp. 239–50; 2005; 2006a, pp. 223–31).An exception among compound polities is theUnited States, which has a strong communica-tive discourse as a result of its majoritarianpolitics and presidential system, along witha strong coordinative discourse as a result ofits pluralist processes and federal structures—although these often work at cross purposes.The highly compound European Union, bycomparison, has the weakest of communica-tive discourses as a result of the lack of anelected central government—and its depen-dence on national leaders to speak for it—and the strongest of coordinative discourses,thanks to its highly complex, quasi-pluralistprocesses and quasi-federal structures (seeSchmidt 2006a).

More specific institutional settings are alsoimportant. Discourses succeed when speak-ers address their remarks to the right audi-ences (specialized or general publics) at theright times in the right ways. Their messagesmust be both convincing in cognitive terms(justifiable) and persuasive in normative terms(appropriate and/or legitimate). A successfuldiscourse “gets it right” in terms of a given“meaning context” according to a given “logicof communication.” This suggests not onlythat the ideas in the discourse must “make

sense” within a particular ideational settingbut also that the discourse itself will be pat-terned in certain ways, following rules andexpressing ideas that are socially constructedand historically transmitted (but more on thisbelow).

IDEAS AND DISCOURSEIN INSTITUTIONALISTPERSPECTIVE

The so-called new institutionalism emergedin the mid-1980s in response to an overem-phasis on agency without structure (i.e., ratio-nal choice methodology) or, worse, on agencywithout sentient agents or structures (i.e., be-haviorism). The new institutionalists broughtinstitutions “back in” in an effort to rightthe balance, but they may have tipped it toofar in the other direction. The problem forall three of the older new institutionalismsis that in their effort to develop explanationsthat took account of institutions, the institu-tions they defined have had a tendency to beoverly “sticky,” and the agents (where they ex-ist) have been largely fixed in terms of pref-erences or fixated in terms of norms. Theturn to ideas and discourse by scholars inall three of the new institutionalisms repre-sents their effort to unstick institutions andto unfix preferences and norms. In so doing,however, those who really took ideas and dis-course seriously, whom I call discursive insti-tutionalists (whether or not they would labelthemselves as such), have challenged the basicpremises of the older new institutionalisms.The challenge is both ontological (about whatinstitutions are and how they are created,maintained, and changed) and epistemologi-cal (about what we can know about institutionsand what makes them continue or change withregard to interests and norms).

Institutions and Institutional Change

For the most part, the three older new institu-tionalisms treat institutions (once created) asgiven, whether as continuing structures (the

www.annualreviews.org • Discursive Institutionalism 313

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

historical regularities of HI) or as the contextwithin which agents act (the incentive struc-tures of RI or the cultural norms of SI). Suchinstitutions are thereby external to the actorscollectively. Institutional rules about acting inthe world serve mainly as constraints, whetherby way of RI’s incentives that structure action,HI’s paths that shape action, or SI’s normsthat frame action. Action in institutions in thethree older new institutionalisms conforms toa rule-following logic, whether an interest-based logic of calculation, a norm-based logicof appropriateness, or a history-based logicof path dependence. But if everyone followsrules, once established, how do we explain in-stitutional change? And how do we explainagency? RI, HI, and SI effectively leave us with“unthinking” actors who are in an importantsense not agents at all.1 This subordination ofagency (action) to structure (rules) is the keyproblem for HI, SI, and RI, and it is why allmanner of new institutionalists have turned toideas and discourse in recent years.

DI simultaneously treats institutions asgiven (as the context within which agentsthink, speak, and act) and as contingent (asthe results of agents’ thoughts, words, and ac-tions). These institutions are therefore inter-nal to the actors, serving both as structuresthat constrain actors and as constructs createdand changed by those actors. As a result, ac-tion in institutions is not seen as the product ofagents’ rationally calculated, path-dependent,or norm-appropriate rule-following. Instead,it is the process in which agents create andmaintain institutions by using what I call theirbackground ideational abilities. These under-pin agents’ ability to make sense of and in agiven meaning context, that is, in terms of theideational rules or “rationality” of that setting.But institutional action can also be predicatedon what I call foreground discursive abilities,through which agents may change (or main-tain) their institutions. These discursive abil-ities represent the logic of communication,

1I thank Robert Goodin for this suggestion.

which enables agents to think, speak, and actoutside their institutions even as they are in-side them, to deliberate about institutionalrules even as they use them, and to persuadeone another to change those institutions or tomaintain them. And it is because of this com-municative logic that DI is better able to ex-plain institutional change and continuity thanthe older three new institutionalisms.

This said, DI can be seen as com-plementary to the other three institution-alisms. Institutions—whether understood asRI’s incentive-based structures, HI’s histori-cally established patterns, or SI’s socially con-stituted norms—frame the discourse. Theydefine the institutional contexts within whichrepertoires of more or less acceptable (andexpectable) ideas and discursive interactionsdevelop. As such, the older three new in-stitutionalisms could be seen to providebackground information for what one nor-mally expects, given the structural constraints,as opposed to what one often gets—theunexpected—which may better be explainedby DI. Importantly, DI can explain the un-expected not just because it can account forunique events by reference to individuals’ideas and discourse, but also because the unex-pected may actually be expectable when anal-ysis is based on a particular set of ideationalrules and discursive regularities in a givenmeaning context following a particular logicof communication—rather than being basedon rationalist interests following a logic ofcalculation, historical regularities following alogic of path dependence, or cultural normsfollowing a logic of appropriateness.

Most political scientists who take ideasand discourse seriously intuitively assumethat institutions are simultaneously structureand construct (agency) in which agents haveboth background ideational and foregrounddiscursive abilities, and they generally use thestructural accounts of one or more of the threeolder institutionalisms as background infor-mation. But they rarely articulate it. This is astrue for discursive institutionalists engagingwith the RI tradition as it is for those in the HI

314 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

tradition. In fact, only discursive institution-alists working in the SI tradition have donemuch to elaborate on the ontological issues,influenced by such continental philosophersand macrosociologists as Bourdieu (1994),Foucault (2000), Habermas (1989, 1996), andGiddens (1984). Wendt (1987, pp. 359–60),for example, whose “structurationalist”theory derives largely from the work ofGiddens, sees social structures as having “aninherently discursive dimension in the sensethat they are inseparable from the reasons andself-understandings that agents bring to theiractions,” while agents and structures are “mu-tually constitutive.” But what does this mean?And how can one establish this duality, with-out falling into the trap of emphasizing struc-ture over agency or agency over structure?

Although philosophers in the continen-tal tradition have done most to address thesequestions, I turn first to the work of Searle, aphilosopher in the analytic tradition, for simi-lar kinds of insights on the construction of so-cial reality. Searle (1995) defines “institutionalfacts” as those things which exist only throughcollective agreement about what stands foran institution. Although such facts are con-sciously created by sentient agents throughwords and action, people may quickly losesight of this, not only because they grow upassuming the existence of such facts but alsobecause the whole hierarchy of institutionalfacts evolves as people use institutions. ForSearle (1995, pp. 140–45), this hierarchy ofinstitutional facts makes up the structure ofconstitutive rules to which agents are sen-sitive as part of their background abilities,which encompass human capacities, disposi-tions, and know-how (knowledge of how theworld works and how to cope with it). Suchbackground abilities are internal to agents,enabling them to speak and act without theconscious or unconscious following of exter-nal rules assumed by RI (rationalist calcu-lation), HI (path dependence), or SI (normappropriateness).

The concept of background abilities is notunique to Searle, as he himself acknowledges.

He sees it as the focus of Wittgenstein’s (e.g.,1968) later work and notes that it is present inBourdieu’s notion of the habitus (Searle 1995,pp. 127–32). Bourdieu’s habitus is in fact quitesimilar to Searle’s concept of background abil-ities, in that he sees human activity as neitherconstituted nor constitutive but both simul-taneously. Human beings act “following theintuitions of a ‘logic of practice’’’ (Bourdieu1990, p. 11). The theory of cognitive disso-nance in psychology also comes close to whatwe are talking about here, at least insofar as itrefutes assumptions about the rule-followingnature of behavior, because it shows that peo-ple generally act without thinking of any rulesthey may be following, but then check whatthey are doing against the various rules thatmight apply. Consciousness about the rulescomes into play mainly where cognitive disso-nance occurs, that is, when the rules are con-tradictory (Harmon-Jones & Mills 1999).

The ideational processes by which agentscreate and maintain institutions, whether weuse Searle, Bourdieu, or cognitive psychol-ogy to ground them, can be summarized bythe concept of background ideational abili-ties. This generic concept is useful in signi-fying what goes on in individuals’ minds asthey come up with new ideas, but it does notexplain much about the processes by whichinstitutions change, which is a collective en-deavor. It also underemphasizes a key compo-nent in human interaction that helps explainsuch change: discourse.

We undersell DI if we equate the ontol-ogy of institutions with background ideationalabilities alone, neglecting foreground discur-sive abilities (people’s ability to think andspeak outside the institutions in which theycontinue to act). For this, we could turn forsupport to Habermas (1989, 1996), with hisconcept of communicative action. But it is alsoin line with much of the literature on “discur-sive democracy” and deliberative democracy(e.g., Dryzek 1990, 2000), which is about theimportance of discourse and deliberation inbreaking the elite monopoly on national andsupranational decision making while ensuring

www.annualreviews.org • Discursive Institutionalism 315

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

democratic access to such decision making.Even though Searle (1995) does not talk aboutforeground abilities, one could argue that hetoo is open to this discursive side. He recog-nizes the importance of language, in particularof “speech acts,” and he insists that institu-tional change can be not only unconscious, asagents start to use the institutions differently,but also conscious, when they decide to usethem differently. The main point of all thesephilosophies is that discourse as an interactiveprocess is what enables agents to change in-stitutions, because the deliberative nature ofdiscourse allows them to conceive of and talkabout institutions as objects at a distance, andto dissociate themselves from them even asthey continue to use them. This is becausediscourse works at two levels: at the everydaylevel of generating and communicating aboutinstitutions, and at a meta-level, as a second-order critical communication among agentsabout what goes on in institutions, enablingthem to deliberate and persuade as a preludeto action.

These foreground discursive abilities alsoprovide a direct response both to propo-nents of the older new institutionalisms,who emphasize structural rule-following overagency, and to continental philosophers andmacro-sociologists who imply ideational rule-following. Among the latter, Bourdieu (1994)argues that the doxa (worldview) of elites whodominate the state creates the habitus that con-ditions people to see the world in the way cho-sen by the elites; and Foucault (2000) suggeststhe impossibility of escape from the ideationaldomination of the powerful. Foreground dis-cursive abilities enable people to reason, de-bate, and change the structures they use—apoint also brought out by Gramsci (1971),who emphasizes the role of intellectuals inbreaking the hegemonic discourse.

Thus, by combining background idea-tional abilities with foreground discursiveabilities, DI puts the agency back into insti-tutional change by explaining the dynamicsof change in structures through constructivediscourse about ideas. In so doing, DI also

provides an answer to the problems of HIs inparticular, both in accounting for agency andin explaining the dynamics of institutionalchange (see Schmidt 2007a).

One of the main problems with HI, infact, is that despite the reference to historyin its title, it tends to be rather ahistorical.Change is explained mainly by reference tocritical junctures or “punctuated equilib-rium” (Krasner 1988), or history is givenvery limited play through path dependence,with its “lock-in effects” and “positive re-inforcement” mechanisms (Pierson 2000).Discursive institutionalists in the HI traditionalso often explain change as coming at criticaljunctures—periods of “third-order change”(Hall 1993); “critical moments” in which “col-lective memories” are made and/or changed(Rothstein 2005, ch. 8); critical junctureswhen public debates serve to reframe howcountries “come to terms with the past” (Art2006); “great transformations” when ideasserve to recast countries’ political economicpolicies (Blyth 2002); or moments when a“window of opportunity” (Kingdon 1984)opens and the search for a new policy programbegins. But whereas for historical institution-alists such critical moments are unexplainabletimes when structures shift, for discursive in-stitutionalists these moments are the objectsof explanation through ideas and discourse,which lend insight into how the historicallytransmitted, path-dependent structures arereconstructed. Importantly, however, discur-sive institutionalists also consider change ina more evolutionary manner. The literatureincludes some notable examples. Berman(2006) traces the slow transformation ofsocialists into Social Democrats as theirpolitical ideas shifted in the effort to findworkable and equitable democratic solutionsto the economic challenges of globalizingcapitalism. Crawford (2006) charts thedevelopment of the idea of trusteeshipfrom the discourse of colonialism to that ofcontemporary international institutions.

Even recent innovations in HI (e.g.,Streeck & Thelen 2005) that elaborate on

316 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

the incremental processes of change result-ing from actors’ use of mechanisms of lay-ering, conversion, and interpretation mainlydescribe such change rather than explain it byreference to what actors themselves think andsay that leads to change. In fact, when histor-ical institutionalists concern themselves withagency, they tend to turn either to RI for acalculus-oriented agency or to SI for a culture-oriented one (see Hall & Taylor 1996), andthey have thereby ended up with another kindof static, equilibrium-focused explanation (seeSchmidt 2006b, 2007a). This is why, in recentyears, increasing numbers of historical insti-tutionalists have turned to ideas and discoursefor agency. King (1999) focuses on the role ofideas and knowledge in the making of illiberalimmigration policy in Britain and the UnitedStates; Lieberman (2005) combines institu-tions and ideas in his history of racial incor-poration in America; and Weir (2006) arguesthat organized labor’s efforts to redefine itselfas a political actor in the United States and tobuild new coalitions can best be explained byconsidering how organizational leaders “puz-zled” and “powered” over questions of iden-tity, alliances, and values as well as interests.

How complementary are these ap-proaches? Do the historical rules and regu-larities brought out in HI investigations serveas unquestionable background informationfor DI explanations of agency? If we assumethat HI analysis elucidates structures andDI illuminates agency, to what extent arewe papering over the differences betweenthese two approaches? My investigation ofdemocracy in Europe (Schmidt 2005, 2006a)shows that a HI explanation of the differentialinstitutional impact of European integrationon simple polities such as Britain and Franceversus compound polities such as Germanyand Italy helps describe the challenges tothese countries’ organizing principles ofdemocracy, but it does not account for theirresponses, because institutional design is notdestiny. Only by adding a DI explanation ofthe role of legitimating ideas and persuasivediscourse in promoting (or not) public

acceptance of the EU can we fully understandnational responses to the EU.

Interests and Uncertainty

In RI, the turn to ideas has also been rela-tively recent. It has shed light on problemsthat could not be solved in terms of inter-ests alone, such as how preferences are createdand how they may change. The turn to ideashas only gone so far, however, because ratio-nal choice institutionalists continue to assumethat preferences remain fixed, that objectiveinterests are analytically separable from sub-jective ideas, and that ideational explanationis useful only when and if explanation interms of objective or material interests is in-sufficient (e.g., Goldstein & Keohane 1993;see critiques by Blyth 2002, Gofas & Hay2009).

For discursive institutionalists generally—and in this they are in agreement with so-ciological institutionalists—the fundamentalflaws of RI’s approach to ideas are its assump-tions that rationality is mainly instrumental,that objective or material interests exist, thatthey are separable from ideas, and that theycan also represent the incentive structures forrational action. Against such instrumental ra-tionality, Boudon (2003) summarizes many DIas well as SI arguments when he contends thatactors are not motivated by self-interest alonebut rather have a wider range of reasons foracting—including moral, prudential, and “ax-iological” (norm-based)—many of which arenot commonplace, do not have consequencesfor others, and do not directly affect their ownself-interest. Rationality for Boudon (2003,p. 18) is cognitive rather than instrumental,so action needs to be explained in terms of itsmeaning to the actor, as grounded in a systemof reasons (similar to meaning context) thatthe actor sees as strong. And in this cognitivesystem, as most DI scholars agree, one can-not distinguish objective interests from ideas;all interests are ideas, and ideas constitute in-terests, so all interests are subjective (see Hay2006).

www.annualreviews.org • Discursive Institutionalism 317

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

Interest-based behavior certainly exists,but it involves ideas about interests that mayencompass much more than strictly utilitar-ian concerns. As Jabko (2006) shows in hisdiscussion of the political strategy for unitingEurope, ideas can certainly be used strategi-cally in political life to advance certain inter-ests against others, but such strategic ideasare rarely reducible to instrumental inter-ests arrived at through utilitarian calculi; in-stead, they draw from a larger repertoire ofideas, in his case, about the European sin-gle market. Moreover, as Seabrooke (2006)demonstrates, even in the international finan-cial markets—an arena that seems to demandthe most interest-based calculations of costsand benefits—economic social norms comeinto play, serving as the social sources of states’international financial power.

But if interests are subjective and norm-driven, then the RI assumption that they canserve as neutral incentive structures is alsoflawed, especially when rational choice in-stitutionalists assume that narrowly instru-mental behavior can lead to the establish-ment of credible institutions. Rothstein (2005,pp. 137–66) argues that institutions shouldnot be seen as neutral structures of incentivesor (worse) immutable products of culture thatlead to inescapable “social traps,” but are bet-ter understood as the carriers of ideas or “col-lective memories.” As such, they can be theobjects of trust or mistrust, and are thereforechangeable over time as actors’ ideas and dis-course about them can change in response tochanges in their performance.

However, if everything is related to ideasand discourse, with no neutral incentive struc-tures and no objective and material interests,one might think that DI leads to some sort ofextreme idealism in a radically uncertain, im-material world. Far from it. Most discursiveinstitutionalists do not deny the existence ofmaterial reality; they just oppose the confla-tion of material reality and interests into “ma-terial interests.” Material reality is, rather, thesetting within which or in response to which

agents may conceive of their interests. Discur-sive institutionalists problematize RI’s notionof objective material interests by theorizinginterests as subjective responses to materialconditions, and they take the actual responsesto material reality as their subject of inquiry(see Schmidt 2006b, Hay 2006, Gofas & Hay2009).

We are left with two final questions: Whatis material reality? And how do we deal withrisk and uncertainty in a material world? Ra-tional choice institutionalists tend to assumea correspondence view of the world, i.e., thatmaterial reality is out there for agents to seeand that scholars are in the business of dis-covering it. DI positions range widely, fromthose who hold something akin to a corre-spondence view—for example, through a kindof “rump materialism” determining a hierar-chy of needs in economic life (Wendt 1999,pp. 109–10)—to those who assume that mostof reality is constructed by the actors them-selves beyond a very basic level (e.g., Hay2006; M. Blyth, unpublished manuscript).

But to ask if material reality exists is thewrong question. We do better to ask what ismaterial and real, and what is real even if it isnot material. Institutions may be real in thesense that they constitute interests and causethings to happen, even though they are so-cially constructed and thus not material in a“put your hand or rest your eyes on it” sense.Searle (1995) helps elucidate this point whenhe distinguishes between “brute facts” such asmountains, which are material because theyexist regardless of whether sentient (inten-tional) agents acknowledge their existence orhave words for them, and “social facts,” ofwhich institutional facts are a subset. Insti-tutions are not material because they do notexist without sentient agents; but, like money,they are real and have causal effects.

To clarify questions of certainty or uncer-tainty related to social facts, we could turn toWittgenstein. His On Certainty (1972) makesa little-noticed but important distinction be-tween two kinds of language games: those

318 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

based on our experience and those based onour pictures of the world. Language gamesbased in our everyday experiences in the worldordinarily admit of no doubts and mistakes—such as knowledge of one’s own name, address,actions, and history; of the number of handsand toes one has; and of the meaning of thewords one uses. By contrast, language gamesbased on our pictures of the world, whichoften follow from our (social) scientific in-terpretations of the world—such as belief inthe existence of the earth 100 years ago, inthe events of history, in the temperature atwhich water boils—always allow for doubts,mistakes, and even gestalt switches or radi-cal conversions, although much less often forlanguage games at the “foundation” of ourpicture of the world, which “stand fast” be-cause they are part of the very “scaffolding”of our thoughts (Wittgenstein 1972, §s 211,234). Taleb & Pilpel (2003; see discussion byM. Blyth, unpublished manuscript) make asimilar point when they demonstrate that theworld in which we live is a lot more uncer-tain than the world generally assumed by riskeconomists and rational choice institutional-ists. This is because of the “problem of thenon-observability of probability generators,”that is, the impossibility of knowing, let alonestatistically predicting, the effects of all theforces that may have an impact on economicand political realities.

What does this mean for political scien-tists? It means that our own generalizationsmay have varying degrees of certainty, de-pending on their objects of knowledge andexplanation. But how do we operationalizethis? And where is the line between RI andDI? Blyth (unpublished manuscript) providesthe beginnings of an answer when he pointsout that RI notions of uncertainty are reallyabout risk, because agents assume some di-rectly observable world that they can perceivemore or less well and in which they can cal-culate the subjective probability of the likelyoutcomes of their preferences, such as in theUS Congress. Under real (Knightian) un-

certainty, agents are not simply unsure abouthow to achieve their interests but unsure ofwhat their interests are in a world that is notdirectly observable, such as in the global econ-omy. Here, we do best to make sense of actors’policies, say, about flexible labor markets andfree trade, in terms of their programmaticideas and discourse in response to their per-ceptions of the challenges of globalization.For Blyth, much of social science exists inthis more uncertain world in which ideasare fundamental to explanation, and which isat odds with the older new institutionalists’taken-for-granted assumptions about institu-tional equilibrium, linear causality, exogenousforces for change, and normality. As a result,although Blyth would accept an RI interest-based explanation as perfectly adequate tothe task in certain instances, he finds that,for the most part, it represents expressions ofsocial scientists’ and social actors’ desires fora certain world rather than the world itself.

But what, then, does this tell us about therelationship between RI and DI? It suggestsat the very least that RI can serve as back-ground information to DI in two ways: First,when RI appears sufficient to explain humanaction, it can be seen as providing DI with ashortcut by way of its account of “interest-based ideas,” which are nothing more thanthe range of responses to material realitiesthat can be expected (although not predicted),given what we know about human rationalityand irrationality. Second, when RI fails to ex-plain, it can serve as a jumping-off point forDI, indicating what discursive institutional-ists could usefully investigate and might doa better job explaining. This does not mean,however, that we should turn to ideas onlywhen RI does not explain—the view of thosewho see ideas as switches or focal points. Fol-lowing this logic would imply that DI explainsonly the unexpected, by accounting for uniqueevents. DI may also explain the expected inunexpected (for RI) ways as well as the seem-ingly irrational, by analyzing ideas and actionsin a given meaning context.

www.annualreviews.org • Discursive Institutionalism 319

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

Norms and RelativismIn SI, we cannot talk about a turn to ideas oreven discourse as such, since SI is all aboutideas and discourse, in particular with re-gard to questions of norms, cognitive frames,and meaning systems, and the ways in whichthey are created and changed. What distin-guishes sociological institutionalists from dis-cursive institutionalists working in the SI tra-dition is more a difference of degree than kind.The difference is the extent to which ideasare treated as dynamic constructs (DI) or asstatic structures (SI). The difficulty in distin-guishing between the two, especially in inter-national relations, is that scholars on eitherside of the static/dynamic divide call them-selves constructivists because they see identityand interests as endogenous and socially con-structed, in contrast with the neo-utilitarians,or rational choice institutionalists, for whomidentities and interests remain exogenous andgiven (see Ruggie 1998, p. 864). Ruggie clari-fies the distinction when he notes that SI con-structivists such as Katzenstein (1996) and hiscolleagues “cut into the problem of ideationalcausation at the level of ‘collective representa-tions’ of ideational social facts and then tracethe impact of these representations on be-havior. They do not, as Weber tried, beginwith the actual social construction of mean-ings and significance from the ground up”(Ruggie 1998, pp. 884–85).

DI constructivists in the SI tradition areall those who engage dynamically with theconstruction of ideas and discourse. In in-ternational relations, these include, for ex-ample, Wendt (1987, 1999), discussed above;Finnemore & Sikkink (1998), who examinethe diffusion of international norms to devel-oping countries; and Risse (2001), who con-siders how different European countries suc-cessively constructed and reconstructed theirstate identities and ideas about European in-tegration. In comparative and internationalpolitical economy, DI constructivists includeAbdelal (2006), who contends that the rulesfor global finance changed not because of a

Washington consensus but because of a “Parisconsensus,” in which European policy mak-ers conceived and promoted the liberal rulesnow structuring the international financialmarkets; Hay (2001) and Hay & Rosamond(2002), who detail the ways in which politi-cal leaders crafted discourses about the chal-lenges of globalization to legitimate neolib-eral reform at home; and Seabrooke (2006),who shows that a state’s influence in the in-ternational financial order is based not onthe resources of its elite financial actors but,rather, on the legitimacy that emerges fromits everyday dealings with ordinary people inlower income groupings. Significantly here,Seabrooke (2006, ch. 2) shows that states de-velop international financial capacity not be-cause their political leaders have a persuasivetop-down “master” communicative discoursefor the masses, nor even because of a success-ful top-to-top coordinative discourse amongstate and financial elites influenced by dis-cursive coalitions and entrepreneurs. Rather,states develop such capacity on the basis ofthe legitimacy they gain through a bottom-up communicative discourse consisting of thedeliberative interactions and contestations be-tween state actors and economic actors withincomes below the median level.

Important questions are related to whichway the arrows go—top-down, top-to-top, orbottom-up—and who is seen as the carrier ofideas. The main problem with the top-downideational and/or discursive process is that le-gitimation is seen as hierarchical, with elitesin charge and entrepreneurial actors jumpingthrough windows of opportunity in momentsof uncertainty to produce a shift in ideas. AsCampbell (1998, p. 383) notes, this leaves therest of us as “institutional dopes blindly fol-lowing the institutionalized scripts and cuesaround them.” For Seabrooke (2006, pp. 4–42), the problem is also that with such top-down approaches scholars end up presentinglegitimacy as a condition tied to beliefs morethan as a process of ongoing contestation indeliberative discursive processes.

320 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

Discursive institutionalists in the SI tra-dition share with constructivists of all stripesthe rejection of the RI emphasis on the in-dividual in favor of a more collective ap-proach to the creation of ideas, with intersub-jective meanings underpinned by culture andnorms. Therefore, sociological institutional-ists, including discursive institutionalists inthe SI tradition, make no universalistic claimsabout rationality. Moreover, if discursive insti-tutionalists who engage with the RI traditioncan be seen to focus on cognitive rationality,then discursive institutionalists working in theSI tradition can be described as emphasizingnormative rationality.

But because they often focus on expla-nation within cultures rather than acrossthem, and on normative ideas rather thaninterest-based (cognitive) ideas, discursiveinstitutionalists in the SI tradition (as wellas sociological institutionalists more gener-ally) are sometimes accused of an implicitrelativism. The question is raised as towhether they can make any cross-nationalgeneralizations at all, or even if there is any-thing mutually recognizable across cultures.Sikkink (1991) was criticized as leaving herselfopen to charges of relativism because she saweverything as socially constructed within agiven culture (see Jacobsen 1995). In fact,generalizations are possible even when onetakes a strongly normative and culture-basedview of rationality, by invoking similaritiesas well as differences in cultural norms andidentities. One could argue even here thatcertain ideas and norms are more universalthan others—those based on Wittgenstein’s“experience games” as opposed to “picturegames.” Moreover, there are certain bases tohuman rationality that allow for universalism,illustrated in Wittgenstein’s (1968, II, xi,p. 223) famous observation that “if a lioncould talk, we would not understand him.”And if all interests and norms are ideas, and allideas are constructed, it is just as possible, al-though not as easy, to construct internationalideas about interests and norms. What is thetwentieth-century notion of human rights,

after all, if not that (see Risse et al. 1999)? Thepoint, in short, is that norms are intersubjec-tive and discursively constructed and, as such,can for the most part be understood acrosscultures even when they are not shared.

Norms, moreover, are everywhere. This isargued most forcefully by philosophers andmacrosociologists such as Foucault, Bourdieu,and Gramsci with regard to the inevitabilityof elite domination of norm construction. Butwe need not take as radical a view of powerto make the point that ideas and values in-fuse both the exercise and the study of power(Lukes 2005). We can apply this argument toRI as well, since even critics within that tra-dition note that rational choice institutional-ists do little to question the institutional ruleswithin which rational actors seek to maximizetheir utility, instead mostly assuming themto be good (Moe 2003, p. 3) and/or efficient(North 1990). The problem with ignoring thevalues embedded in our research, and believ-ing that our work is value-neutral and there-fore objective, is not only that it may skewresearch findings. It is also that political scien-tists lose an important opportunity to engagewith politics, which is clearly all about values.This is most obvious in the world of think-tanks: Conservative think-tanks, which pro-duce unabashedly political and value-laden re-search, have gotten a much bigger bang fortheir buck in Washington than more progres-sive think-tanks, which seek to be (or at leastto appear to be) more value-neutral and ob-jective (see Rich 2004).

CONCLUSION

The objects of discursive institutionalist ex-planation consist of both ideas and discourse.Ideas differ in levels of generality: They maybe specific to policy, encompass a wider pro-gram, or constitute an underlying philos-ophy. They also differ in type: Cognitiveideas are constitutive of interests and norma-tive ideas appeal to values. Discourse servesnot just to represent ideas but also to ex-change them through interactive processes of

www.annualreviews.org • Discursive Institutionalism 321

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

(a) coordination among policy actors in pol-icy and program construction and (b) com-munication between political actors and thepublic in the presentation, deliberation, andlegitimation of those ideas, against a back-ground of overarching philosophies. Institu-tional context also matters—both the formalinstitutional context (simple polities tend tohave a more elaborate communicative dis-course, compound polities a more elaboratecoordinative discourse) and the more specificmeaning context.

DI differs from the three older new insti-tutionalisms in terms of its logic as well as itsobjects of explanation. First, institutions inDI, rather than serving as external structuresfor rule-following, are simultaneously struc-tures and constructs internal to the agentsthemselves. Agents’ background ideationalabilities enable them to act in any given mean-ing context to create and maintain institutionswhile their foreground discursive abilitiesenable them to communicate critically aboutthose institutions and so to change or maintainthem. Institutional change in DI, therefore,as opposed to in HI, is dynamic and explain-able across time through agents’ ideas anddiscourse, rather than largely static because ofpath-dependent structures and unexplainablecritical moments. Second, interests in DI,as opposed to in RI, are neither objective(because interests are ideas and, as such, sub-jective) nor material. However, the discursiveinstitutionalist is not giving way to total

uncertainty or denying that there is a materialreality out there, because subjective interestsas well as institutions can be real even if notmaterial. Third, norms in DI, as opposed to inSI, are dynamic constructs rather than staticstructures. Here, the intersubjectivity of nor-mative ideational constructions and discursiveinteractions guards against relativism.

Can we have our cake and eat it too?That is, can we accept DI without rejectingthe other three institutionalist approaches?I would like to think so. Political realityis vast and complicated. No one method-ological approach is able to explain it suffi-ciently. Each gets at a different piece of re-ality, at different levels of abstraction, withdifferent kinds of generalizations, and dif-ferent objects and logics of explanation. Itis for this reason that DI can treat the re-sults of the other institutionalist approachesas background information. Such results maybe taken for granted as common-sense ideasand discourse about political reality, or may beproblematized and investigated. I have previ-ously suggested that political scientists “givepeace a chance” (Schmidt 2006b), abandon-ing their methodological wars in order to ex-plore the boundaries between their method-ological approaches. I reiterate this appealhere. But I also maintain that only with aclearer view of approaches that take ideasand discourse seriously can political scien-tists begin to explain the fullness of politicalreality.

DISCLOSURE STATEMENT

The author is not aware of any biases that might be perceived as affecting the objectivity ofthis review.

ACKNOWLEDGMENTS

For their comments at various stages of manuscript preparation, I thank Peter Katzenstein,David Art, Mark Blyth, Jean-Claude Barbier, Sheri Berman, Charlotte Epstein, Robert Goodin,Jean Leca, Pierre Muller, Fritz Scharpf, and Mark Thatcher.

LITERATURE CITED

Abdelal R. 2007. Capital Rules: The Construction of Global Finance. Cambridge, MA: HarvardUniv. Press

322 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

Art D. 2006. The Politics of the Nazi Past in Germany and Austria. New York: Cambridge Univ.Press

Barbier JC. 2004. A comparative analysis of ‘employment precariousness’ in Europe. Eur. Res. Cent.,Cross-Natl. Res. Pap. 7(3)

Benhabib S. 1996. Toward a deliberative model of democratic legitimacy. In Democracy andDifference, ed. S Benhabib, pp. 67–95. Princeton, NJ: Princeton Univ. Press

Berman S. 1998. The Social Democratic Movement: Ideas and Politics in the Making of InterwarEurope. Cambridge, MA: Harvard Univ. Press

Berman S. 2006. The Primacy of Politics: Social Democracy and the Making of Europe’s TwentiethCentury. New York: Cambridge Univ. Press

Blyth M. 2002. Great Transformations: Economic Ideas and Institutional Change in the TwentiethCentury. New York: Cambridge Univ. Press

Blyth M. Four (plus two) reasons to take ideas very seriously indeed. In Ideas and Politics inSocial Scientific Analysis, ed. D Beland, R Cox. Unpublished manuscript

Boudon R. 2003. Beyond rational choice theory. Annu. Rev. Sociol. 29:1–21Bourdieu P. 1990. In Other Words: Essays towards a Reflexive Sociology. Stanford, CA: Stanford

Univ. PressBourdieu P. 1994. Raisons Pratiques. Paris: Le SeuilBusch A. 2004. National filters: Europeanisation, institutions, and discourse in the case of

banking regulation. West Eur. Polit. 27(2):310–33Campbell JL. 1998. Institutional analysis and the role of ideas in political economy. Theory

Society 27(3):377–409Campbell JL. 2004. Institutional Change and Globalization. Princeton, NJ: Princeton Univ. PressCampbell JL, Pedersen O. 2001. Introduction. In The Rise of Neoliberalism and Institutional

Analysis, ed. JL Campbell, O Pederson, pp. 1–24. Princeton, NJ: Princeton Univ. PressCampbell JL, Pedersen O. Knowledge regimes and comparative political economy. In Ideas

and Politics in Social Scientific Analysis, ed. D Beland, R Cox. Unpublished manuscriptConnolly WE. 1983. The Terms of Political Discourse. Princeton, NJ: Princeton Univ. PressCox RH. 2001. The path dependency of an idea: why Scandinavian welfare states remain

distinct. Soc. Policy Admin. 38(2):204–19Crawford N. 2006. How previous ideas affect later ideas. In The Oxford Handbook of Contextual

Political Analysis, ed. R Goodin, C Tilly, pp. 266–83. Oxford, UK: Oxford Univ. PressDobbin F. 1994. Forging Industrial Policy: The United States, Britain, and France in the Railway

Age. New York: Cambridge Univ. PressDryzek J. 1990. Discursive Democracy: Politics, Policy, and Political Science. Cambridge, UK:

Cambridge Univ. PressDryzek J. 2000. Deliberative Democracy and Beyond. Oxford, UK: Oxford Univ. PressDyson K, ed. 2002. European States and the Euro: Europeanization, Variation, and Convergence.

Oxford, UK: Oxford Univ. PressEriksen EO, Fossum JE. 2002. Democracy through strong publics in the European Union?

J. Comm. Mkt. Stud. 40(3):401–24Finnemore M, Sikkink K. 1998. International norm dynamics and political change. Int. Org.

52:887–917Fligstein N, Mara-Drita I. 1996. How to make a market: reflections on the attempt to create

a single market in the European Union. Am. J. Sociol. 102:1–32Foucault M. 2000. Power. Essential Works of Foucault, 1954–1984, Vol. 3, ed. JD Faubion. New

York: New PressFouilleux E. 2004. Reforms and multilateral trade negotiations: another view on discourse

efficiency. West Eur. Polit. 27(2):235–55

www.annualreviews.org • Discursive Institutionalism 323

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

Giddens A. 1984. The Constitution of Society: Outline of a Theory of Structuration. Cambridge,UK: Polity

Gofas A, Hay C. 2009. The ideas debate in political analysis: towards a cartography and criticalassessment. In The Role of Ideas in Political Analysis: A Portrait of Contemporary Debates, ed.A Gofas, C Hay. London: Routledge. In press

Goldstein J, Keohane R. 1993. Ideas and Foreign Policy: Beliefs, Institutions and Political Change.Ithaca, NY: Cornell Univ. Press

Goodin RE, Dryzek JS. 2006. Deliberative impacts: the macro-political uptake of mini-publics.Polit. Soc. 34(2):219–44

Goodin RE, Tilly C, eds. 2006. Oxford Handbook of Contextual Political Analysis. Oxford, UK:Oxford Univ. Press

Gramsci A. 1971. Selections from the Prison Notebooks. New York: Int. Publ.Guiraudon V. 1997. Policy change behind gilded doors: explaining the evolution of aliens’ rights in

contemporary Western Europe 1974–1994. PhD thesis, Dep. Gov., Harvard Univ.Haas PM. 1992. Introduction: epistemic communities and international policy coordination.

Int. Org. 46:1–35Habermas J. 1989. The Structural Transformation of the Public Sphere. Transl. T Burger,

F Lawrence. Cambridge, MA: MIT PressHabermas J. 1996. Between Facts and Norms: Contributions to a Discourse Theory. London: PolityHajer M. 2003. A frame in the fields: policymaking and the reinvention of politics. In Deliberative

Policy Analysis: Understanding Governance in the Network Society, ed. M Hajer, H Wagenaar,pp. 88–112. Cambridge, UK: Cambridge Univ. Press

Hall P. 1989. Conclusion. In The Political Power of Economic Ideas: Keynesianism across Nations,ed. P Hall, pp. 361–92. Princeton, NJ: Princeton Univ. Press

Hall P. 1993. Policy paradigms, social learning and the state: the case of economic policy-making in Britain. Comp. Polit. 25:275–96

Hall P, Taylor R. 1996. Political science and the three new institutionalisms. Polit. Stud. 952–73Harmon-Jones E, Mills J. 1999. Cognitive Dissonance: Progress on a Pivotal Theory in Social

Psychology. Washington, DC: Am. Psychol. Assoc.Hay C. 2001. The “crisis” of Keynesianism and the rise of neoliberalism in Britain: an

ideational institutionalist approach. In The Rise of Neoliberalism and Institutional Analysis, ed.JL Campbell, O Pedersen, pp. 193–218. Princeton, NJ: Princeton Univ. Press

Hay C. 2006. Constructivist institutionalism. In The Oxford Handbook of Political Institutions,ed. RAW Rhodes, S Binder, B Rockman, pp. 56–74. Oxford, UK: Oxford Univ. Press

Hay C, Rosamond B. 2002. Globalisation, European integration and the discursive constructionof economic imperatives. J. Eur. Pub. Pol. 9(2):147–67

Holzinger K. 2004. Bargaining through arguing: an empirical analysis based on speech acttheory. Polit. Comm. 21(2):195–222

Howorth J. 2004. Discourse, ideas and epistemic communities in European security and defencepolicy. West Eur. Polit. 27(1):29–52

Jabko N. 2006. Playing the Market. Ithaca, NY: Cornell Univ. PressJacobsen JK. 1995. Much ado about ideas: the cognitive factor in economic policy. World Polit.

47:283–310Jobert B. 1989. The normative frameworks of public policy. Polit. Stud. 37:376–86Katzenstein PJ, ed. 1996. The Culture of National Security: Norms and Identity in World Politics.

New York: Columbia Univ. PressKeck ME, Sikkink K. 1998. Activists Beyond Borders: Advocacy Networks in International Politics.

Ithaca, NY: Cornell Univ. Press

324 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

King D. 1999. In the Name of Liberalism: Illiberal Social Policy in the United States and Britain.Oxford, UK: Oxford Univ. Press

Kingdon J. 1984. Agendas, Alternatives and Public Policies. New York: LongmanKrasner S. 1988. Sovereignty: an institutional perspective. Comp. Polit. Stud. 21:66–94Kuhn T. 1970. The Structure of Scientific Revolutions. Chicago: Univ. Chicago Press. 2nd ed.Lakatos I. 1971. Falsification and the methodology of scientific research programmes. In Criti-

cism and the Growth of Knowledge, ed. I Lakatos, A Musgrave, pp. 91–196. Cambridge, UK:Cambridge Univ. Press

Lehmbruch G. 2001. Institutional embedding of market economies: the German model and itsimpact on Japan. In The Origins of Nonliberal Capitalism: Germany and Japan in Comparison,ed. W Streeck, K Yamamura, pp. 39–93. Ithaca, NY: Cornell Univ. Press

Lieberman RC. 2005. Shaping Race Policy: The United States in Comparative Perspective.Princeton, NJ: Princeton Univ. Press

Lukes S. 2005. Power, Second Edition: A Radical View. Basingstoke, UK: Palgrave MacmillanMajone G. 1989. Evidence, Argument, and Persuasion in the Policy Process. New Haven, CT: Yale

Univ. PressMarch J, Olsen J. 1989. Rediscovering Institutions. New York: Free PressMcNamara K. 1998. The Currency of Ideas: Monetary Politics in the European Union. Ithaca, NY:

Cornell Univ. PressMehta J. From “whether” to “how”: the varied role of ideas in politics. In Ideas and Politics in

Social Scientific Analysis, ed. D Beland, R Cox. Unpublished manuscriptMoe T. 2003. Power and political institutions. Presented at conf. on Crafting and Operating

Institutions, Yale Univ., New Haven, CT, Apr. 11–13Muller P. 1995. Les politiques publiques comme construction d’un rapport au monde. In La

Construction du Sens dans les Politiques Publiques: Debats autour de la Notion de Referentiel, ed.A Faure, G Pollet, P Warin, pp. 153–79. Paris: L’Harmattan

Mutz DC, Sniderman PM, Brody RA. 1996. Political Persuasion and Attitude Change. Ann Arbor:Univ. Mich. Press

North DC. 1990. Institutions, Institutional Change, and Economic Performance. Cambridge, UK:Cambridge Univ. Press

Parsons C. 2003. A Certain Idea of Europe. Ithaca, NY: Cornell Univ. PressPierson P. 2000. Increasing returns, path dependence, and the study of politics. Am. Polit. Sci.

Rev. 94:251–68Radaelli C, Schmidt V. 2004. Conclusions. West Eur. Polit. 27(2):364–79Rein M, Schon DA. 1994. Frame Reflection Toward the Resolution of Intractable Policy Controversies.

New York: Basic BooksRich A. 2004. Think Tanks, Public Policy, and the Politics of Expertise. New York: Cambridge Univ.

PressRisse T, Ropp SC, Sikkink K, eds. 1999. The Power of Human Rights: International Norms and

Domestic Change. Cambridge, UK: Cambridge Univ. PressRisse T. 2000. Let’s argue. Communicative action in world politics. Int. Org. 54(1):1–39Risse T. 2001. Who are we? A Europeanization of national identities. In Europeanization and

Domestic Change, ed. M Green Cowles, J Caporaso, T Risse, pp. 198–216. Ithaca, NY:Cornell Univ. Press

Roe E. 1994. Narrative Policy Analysis: Theory and Practice. Durham, NC: Duke Univ. PressRothstein B. 2005. Social Traps and the Problem of Trust. Cambridge, UK: Cambridge Univ.

PressRuggie J. 1998. What makes the world hang together? Neo-utilitarianism and the social con-

structivist challenge. Int. Org. 52(4):855–85

www.annualreviews.org • Discursive Institutionalism 325

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

ANRV344-PL11-14 ARI 8 April 2008 15:11

Sabatier P, Jenkins-Smith HC, eds. 1993. Policy Change and Learning: An Advocacy CoalitionApproach. Boulder, CO: Westview

Schimmelfennig F. 2001. The community trap: liberal norms, rhetorical action, and the Easternenlargement of the European Union. Int. Org. 55(1):47–80

Schmidt VA. 2000. Values and discourse in the politics of adjustment. In Welfare and Work in theOpen Economy. Vol I: From Vulnerability to Competitiveness, ed. FW Scharpf, VA Schmidt,pp. 229–309. Oxford, UK: Oxford Univ. Press

Schmidt VA. 2002a. The Futures of European Capitalism. Oxford, UK: Oxford Univ. PressSchmidt VA. 2002b. Does discourse matter in the politics of welfare state adjustment? Comp.

Polit. Stud. 35(2):168–93Schmidt VA. 2005. Democracy in Europe: the impact of European integration. Persp. Polit.

3(4):761–79Schmidt VA. 2006a. Democracy in Europe: The EU and National Polities. Oxford, UK: Oxford

Univ. PressSchmidt VA. 2006b. Give peace a chance: reconciling the four (not three) new institutionalisms in

political science. Presented at Annu. Meet. Am. Polit. Sci. Assoc., PhiladelphiaSchmidt VA. 2007a. Bringing the state back into the varieties of capitalism and discourse back into

the explanation of change. Program for the Study of Germany and Europe Work. Pap. 07.3,Cent. Eur. Stud., Harvard Univ.

Schmidt VA. 2007b. Trapped by their ideas: French elites’ discourses of European integrationand globalization. J. Eur. Pub. Pol. 14(4):992–1009

Seabrooke L. 2006. The Social Sources of Financial Power. Ithaca, NY: Cornell Univ. PressSearle J. 1995. The Construction of Social Reality. New York: Free PressSikkink K. 1991. Ideas and Institutions: Developmentalism in Brazil and Argentina. Ithaca, NY:

Cornell Univ. PressStreeck W, Thelen K, eds. 2005. Beyond Continuity. Oxford, UK: Oxford Univ. PressTaleb NN, Pilpel A. 2005. On the very unfortunate problem of the nonobservability of the probability

distribution. http://www.fooledbyrandomness.com/knowledge.pdfWeir M. 2006. When does politics create policy? The organizational politics of change. In

Rethinking Political Institutions: The Art of the State, ed. I Shapiro, S Skowronek, D Galvin,pp. 171–86. New York: NY Univ. Press

Wendt A. 1987. The agent-structure problem in international relations theory. Int. Org.41(3):335–70

Wendt A. 1999. Social Theory of International Politics. Cambridge, UK: Cambridge Univ. PressWittgenstein L. 1968. Philosophical Investigations. Oxford, UK: Basil BlackwellWittgenstein L. 1972. On Certainty. New York: HarperZaller J. 1992. The Nature and Origins of Mass Opinion. New York: Cambridge Univ. Press

326 Schmidt

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

AR344-FM ARI 8 April 2008 18:44

Annual Review ofPolitical Science

Volume 11, 2008Contents

State FailureRobert H. Bates � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �1

The Ups and Downs of Bureaucratic OrganizationJohan P. Olsen � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 13

The Relationships Between Mass Media, Public Opinion, and ForeignPolicy: Toward a Theoretical SynthesisMatthew A. Baum and Philip B.K. Potter � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 39

What the Ancient Greeks Can Tell Us About DemocracyJosiah Ober � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 67

The Judicialization of Mega-Politics and the Rise of Political CourtsRan Hirschl � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 93

Debating the Role of Institutions in Political and EconomicDevelopment: Theory, History, and FindingsStanley L. Engerman and Kenneth L. Sokoloff � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �119

The Role of Politics in Economic DevelopmentPeter Gourevitch � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �137

Does Electoral System Reform Work? Electoral System Lessons fromReforms of the 1990sEthan Scheiner � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �161

The New Empirical BiopoliticsJohn R. Alford and John R. Hibbing � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �183

The Rule of Law and Economic DevelopmentStephan Haggard, Andrew MacIntyre, and Lydia Tiede � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �205

Hiding in Plain Sight: American Politics and the Carceral StateMarie Gottschalk � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �235

Private Global Business RegulationDavid Vogel � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �261

Pitfalls and Prospects in the Peacekeeping LiteratureVirginia Page Fortna and Lise Morje Howard � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �283

v

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.

AR344-FM ARI 8 April 2008 18:44

Discursive Institutionalism: The Explanatory Power of Ideasand DiscourseVivien A. Schmidt � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �303

The Mobilization of Opposition to Economic LiberalizationKenneth M. Roberts � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �327

CoalitionsMacartan Humphreys � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �351

The Concept of Representation in Contemporary Democratic TheoryNadia Urbinati and Mark E. Warren � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �387

What Have We Learned About Generalized Trust, If Anything?Peter Nannestad � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �413

Convenience VotingPaul Gronke, Eva Galanes-Rosenbaum, Peter A. Miller, and Daniel Toffey � � � � � � � � �437

Race, Immigration, and the Identity-to-Politics LinkTaeku Lee � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �457

Work and Power: The Connection Between Female Labor ForceParticipation and Female Political RepresentationTorben Iversen and Frances Rosenbluth � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �479

Deliberative Democratic Theory and Empirical Political ScienceDennis F. Thompson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �497

Is Deliberative Democracy a Falsifiable Theory?Diana C. Mutz � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �521

The Social Processes of Civil War: The Wartime Transformation ofSocial NetworksElisabeth Jean Wood � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �539

Political Polarization in the American PublicMorris P. Fiorina and Samuel J. Abrams � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �563

Indexes

Cumulative Index of Contributing Authors, Volumes 7–11 � � � � � � � � � � � � � � � � � � � � � � � � � �589

Cumulative Index of Chapter Titles, Volumes 7–11 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �591

Errata

An online log of corrections to Annual Review of Political Science articles may be foundat http://polisci.annualreviews.org/

vi Contents

Ann

u. R

ev. P

olit.

Sci

. 200

8.11

:303

-326

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

CA

PES

on 0

6/22

/08.

For

per

sona

l use

onl

y.