Advancing Post-Structural Institutionalism: Discourses, Subjects, Power Asymmetries, and Institutional Change

ABSTRACT

Colin Hay’s and Vivien Schmidt’s responses to my previous critical engagement with their respective versions of neo-institutionalism raise the issue of how scholars may account for the ideational power of political processes and how ideas may generate both stability and change. Even though Hay, Schmidt, and I share a common philosophical ground in many respects, we nevertheless diverge in our views about how to account for ideational power and for actors’ ability to navigate a social reality that is saturated with structures and meaning. There continues to be a need for an analytical framework that incorporates discourse and a constitutive logic based upon the power in ideas. Post-structural institutionalism (PSI) analyzes discourse as knowledge claims by means of the concept of a constitutive causality, analytically identified in respect to institutions, such that the substantive content of ideas/discourse provides ideational power and generates immanent change.

Keywords:

• constructivist institutionalism
• discourses
• discursive institutionalism
• ideas
• neo-institutionalism
• post-structural institutionalism
• post-structuralism
• power/knowledge

I greatly appreciate the efforts of Colin Hay and Vivien Schmidt to respond to the criticism of their views presented in my article “Using Post-Structuralism to Explore the Full Impact of Ideas on Politics” (Larsson 2015b). Both are renowned scholars with busy schedules, and it is an honor that they were willing to engage in discussion concerning their respective analytical frameworks and the roles they ascribe to ideational elements and agency in respect to social stability and change.

Schmidt and Hay have done much to pave the way towards a more comprehensive understanding and acceptance of the roles played by ideas and discourses. Hay examined the foundations of social constructivism, as well as the causal and constitutive roles that ideas play in political processes, in his Political Analysis: A Critical Introduction. He observed that

ideas should be accorded a crucial role in political explanation, since actors behave the way they do because they hold certain views about the social and political environment they inhabit. Moreover, those ideas cannot simply be derived from the context itself. . . . Ideas are both real and have real effects. (Hay 2002, 213)

Likewise, Schmidt’s 2008 article “Discursive Institutionalism: The Explanatory Power of Ideas and Discourse” substantially advanced discussion of the role played by both ideas and discourse, making the case that ideational elements, including discursive elements, need to receive proper attention in political analysis. This publication was groundbreaking in that it spawned wider recognition of the importance of ideational elements in mainstream political science. Because of such pioneering work, I and many fellow social scientists who refer to ourselves as social constructivists, postmodernists, or post-structuralists enjoy a far better reputation and acceptance. More importantly, we now have the possibility of engaging in more advanced discussions regarding the various elements and causal powers possessed by ideas and discourses.

Both Hay and Schmidt have been very influential in introducing ideational dimensions into mainstream political science (Hay 2006, 2011; Schmidt 2008, 2010, 2011a, and 2018). Hay’s constructivist institutionalism (CI) and Schmidt’s discursive institutionalism (DI) have attained an agenda-setting status, such that scholars critically and explicitly engage with them order to present contrasting analytical frameworks and findings or else develop their own. The creators of DI and CI have continued to develop the basic tenets of these frameworks on the basis of additional empirical and theoretical insights since the publication of my original article on the issue (Schmidt 2011b; Schmidt and Thatcher 2014; Palier and Hay 2017; Carstensen and Schmidt 2016; Hay 2016). They have thus responded to certain points I raised, but not all (Schmidt 2017; Hay 2017). It is my intent in the present discussion to clarify our remaining differences and provide a detailed response to the advances and counter-arguments that Hay and Schmidt have presented.

My replies to Hay and Schmidt are structured as follows. First, I present the basic tenets of post-structuralism, a social/political approach that emphasizes nominalism, intersubjectivism, meaning creation, contingency, and the existence of an intimate relationship between power and knowledge. While elements of this discussion mirror my previous description of post-structuralist institutionalism, I now provide a more thorough philosophical account that specifies the distinctive elements of this perspective. Second, I revisit my criticism of CI and address Hay’s reply. This discussion centers, first, on whether ideas should be conceptualized as subjective and/or inter-subjective, and, second, on the degree of voluntarism we should ascribe to social and political agents situated in a given institutional context. Third, I turn to DI, focusing on specific conceptualizations of discourse and recent advances concerning how to theorize the relations between power and ideas. Fourth, I bring together elements of post-structuralism and the concept of institutions in order to illustrate that a careful combination of the two approaches provides an analytical framework that accounts for how ideational elements structure the social world and situate agency within contexts that are saturated with differing layers of meaning.

I. POST-STRUCTURALISM AND THE IMPORTANCE OF A SOCIAL ONTOLOGY

Post-structuralism is not a discrete theory, but comprises a social philosophy and an approach to the social world that builds on specific ontological and epistemological premises. The social ontology emphasizes the constitutive function of ideas. It stresses the importance of intersubjectivity and shared meanings, without which there would be no social or cultural reality (Pouliot 2007; Hacking 1999; Adler 1997; Wendt 1999). Social kinds exist as socially constructed entities that are upheld by social, political, and cultural conventions, and they require the active acceptance and support of actors/individuals by means of intersubjectively shared meanings (Wendt 1998; Pouliot 2007).

The central claim of post-structuralism is that social inquiry should recognize the causal power of both observable and unobservable entities. These may produce change regardless of whether they operate through individuals’ mental states. A secondary claim is that there is no independent “essence” or “nucleus” in events and objects that determines how they must be interpreted. They are unstructured; social inquiry is required in order to understand and explain this post-structural world (Derrida 1976). While post-structuralism leaves room for new interpretations, however, it emphasizes the social foundations of the structures that shape actors and argues that hegemonic discourses and present institutions might always be disrupted (Foucault 1990, 92-93). In addition, it is very likely that multiple and contradicting discourses, rationalities, and ideas are at play at one and the same time in any given setting (Larsson 2017a). Even if the setting appears to be highly stable, this is only due to its being located in the eye of the storm. Furthermore, the instability of our analytical and grammatical concepts is not valid only for ideas (Carstensen 2011) insofar as all situations are marked by a fundamental heterogeneity, instability, and tension among ideas, actors, and institutions that may not be recognized by, or even visible to, the actors involved. The emphasis in post-structuralism upon contingency, heterogeneity, and power/knowledge reveals that even those situations that appear stable actually contain hidden conflicts and suppressed voices, such that change and resistance to current institutional arrangements is always possible. Laclau and Mouffe (1985, 111) summarize this point by stating that “neither fixity nor absolute non-fixity is possible,” implying that a completely static relationship between individuals and institutions is not possible. Furthermore, since power is not to be equated with domination and surrender, but rather with struggle, no social situation or institutionalized discourse can ever be completely stabilized (Foucault 1990, 95).

Post-structuralists do not search for the essences of things, but rather for how “things” are constituted through language and meaning-making and thereby brought into being (Bacchi and Goodwin 2016; Bacchi 1999; Hacking 1999, 6). That is to say that we can never know the world beyond our concepts, knowledge, and existing discourses. We thus inhabit a world of differences in which language and texts, broadly understood, serve as barriers to signifiers that become interwoven in complex chains of intertextuality in which a given signifier, concept, or symbol is intimately connected with others. This leads to the creation of virtually infinite chains of signifiers, concepts, and symbols. Derrida’s statement that there is nothing beyond text should be understood metaphorically in the sense that created meanings, knowledges, and images are as far as we can go. The firming up of particular social arrangements and the creation and maintenance of institutions necessarily involves the key element of social and political power that is reinforced by prominent discourses, which, in turn, give way to specific institutions. From this perspective, discourse “is a constitutive dimension of social relations” (Griggs and Howarth 2016, 17).

Accordingly, discourse is not reducible to language or communication, for it consists of socially produced forms of knowledge that establish the limits of what it is possible to think, write, or say about any given social object or reality (Panizza and Miorelli 2012, 5). Michel Foucault is one of the more renowned scholars to champion the alternative view of discourse as knowledge claims rather than language and communication, observing, for example, that “discourse . . . is not a language, plus a subject to speak it. It is a practice that has its own forms of sequence and succession” (Foucault 1972, 169). Nelson Phillips, Thomas B. Lawrence, and Cynthia Hardy (2004) state in this regard that institutions are constituted through discourse and it is not action per se that provides the basis for institutionalization, but rather the texts that describe and communicate those actions (Phillips et al. 2004). Discourse thus provides structures and shapes institutional arrangements, including both formal and informal institutions. One may also argue, without taking a nihilist stance (Dillet 2017, 519), that post-structuralism does not regard knowledge as “true,” but rather as powerful truth claims or “in the true” (Bacchi and Goodwin 2016, 35).

But are there no agents who inhabit this type of social ontology? An important theoretical concept in post-structuralism is subject-position, which explains how subjects come into being through social processes of subjectification. Subjects, including actors, targets, and identities, are thereby understood as the effects of politics and the products of power-knowledge relations (Golder 2010). This particular manner of imagining political subjects and actors calls into question the more widespread approach to the rational autonomous individual that is typical of Enlightenment humanism (Bacchi and Goodwin 2016, 49). Instead, subjects and political actors are thoroughly contingent, emergent, and constituted through discourse and particular institutional contexts. This ontological approach to emergent subjects is often understood and utilized in policy analysis in terms of target groups that stand in need of correction, support programs, and empowerment management, such as obese adults and children, the poor, criminal youth, the unemployed, immigrants, and so forth (Schneider and Ingram 1993; Hacking 1999). While social actors obviously occupy multiple conflicting subject positions, it is important to note that they are also structurally constrained by others’ perceptions and denials of recognition (Dagg and Haugaard 2016).

How, then, can we account for resistance and change? I follow Foucault (1982, 780) in this regard, who argues that we may understand resistance as a “chemical catalyst so as to bring to light power relations, locate their position, and find out their point of application and the methods used.” This implies that if resistance is a result of power relations, then a given counter-action itself is a power strategy/relation that dictates the types of resistance that not only may eventually prevail, but may even be logically entailed in light of the concrete configurations of discourses/knowledge/power and existing institutions (Lilja and Vinthagen 2014). In turn, while “institutions” comprise an analytical category that is useful for understanding patterned behavior and the “rules of the game” (North 1990), institutions are also messy and contingent. Contestation and change can emerge from within the instability of those power relations that are produced through discourses and made manifest in institutions, formal and informal (Beunen and Patterson 2016). I recognize that we need to think carefully about the roles that social and political actors play in respect to institutional change, including the fact that various actors may perceive the same institutions differently and possess differing abilities/resources to change them. Much of my controversy with Hay and Schmidt concerns precisely this issue, namely, the degree of leverage we should ascribe to individuals/actors in respect to institutional change.

Hay and Schmidt describe their motive for introducing their respective analytical frameworks as dissatisfaction with existing neo-institutional approaches, insofar as they proved unable to account for institutional change, at least not without referring to external shocks or disturbances as the primary cause. Ideas could provide a means of explaining incremental and internal change (Larsson 2015b, 184). In the neo-institutional tradition itself, ideas help to explain informal institutions, norms, traditions, cultures, and ideologies: intersubjective structural elements that are shared among the various actors but that incline them toward stability rather than change. Both Hay and Schmidt seek to explain institutional change, undermining the notion that institutions are necessarily “sticky,” by introducing subjective ideas, thus challenging the neo-institutional restriction of ideas to intersubjective and thus binding elements of formal and informal institutions.

II. THE IMPORTANCE OF INTERSUBJECTIVITY AND THE DANGERS OF VOLUNTARISM

In order to provide the leverage necessary for actors to cause change, Hay points to “interests” that supposedly are not created by contexts and structures, but rather result from interpretive processes in which actors conceive of their interests and subsequently pursue institutional change. This understanding of change starts with actors’ ideas (Hay 2011, 69). In a previous text that presents the basic framework of CI, Hay observes that “ideas in the form of perceptions matter” insofar as they shape the individual’s orientation towards her context, making it necessary to separate “institutions and ideas of institutions” (Hay 2011, 69). He also maintains that perceptions are “socially constructed,” which takes on an individualistic and subjective cast when individuals’ ideas do the construction. Hay, however, claims that such a reading is unfortunate because, if my criticism is correct, CI would basically ignore intersubjective elements and promote a subjectivist view that ushers in voluntarism, which makes institutional change appear far simpler than it actually is (Hay 2017, 237). Moreover, he argues, “things that are socially constructed (social facts) are not, and can never be, purely subjective,” adding that there are “important subjective elements to how we encounter, experience and act with respect to them” (ibid.). Thus, if I am correct in describing CI as subjectivist, then it could not be genuinely constructivist. This summarizes my concerns perfectly: CI may not be as consistent with social constructivism as it initially appears to be.

In part, Hay supports his position by reference to a work that was not yet published at the time of my original article. In this later text (Hay 2016), he appeals to Peter Berger, Thomas Luckmann, and John Searle, all of whom are well-known constructivists, arguing that social facts are different in kind from natural facts, in that they acquire both their facticity and the ontologically distinct character of that facticity from processes that are intersubjective rather than subjective. He also maintains in the same passage that only if something is asocial (that is, purely subjective) and/or non-contingent (in the sense that it could not be otherwise) is it not political (Hay 2017, 239). Hay then states that he regards this position as serving to demonstrate that, as “the focus on social/discursive construction would imply, any genuine constructivism cannot be guilty of the subjectivist bias Larsson detects” (ibid.). All of this, however, merely reiterates the social foundations of constructivism by building upon the notion that social facts, by both definition and necessity, are intersubjective. But my original criticism of CI did not regard the foundations of social constructivism in general, but rather their relation to the perception of interests: whether actors subjectively can initiate change by means of their independent ideas concerning interests, and the extent to which this entails a voluntarist view of human agency that ignores the socio-structural aspects of human life.

Hay remarks that “social facts are interpreted and interacted with (at least in part) subjectively,” and that political space is necessarily intersubjective, which supposedly renders all political processes intersubjective “even if they involve only the clash of narrowly subjective preferences” (ibid., 239ff). Hay adds that if we are not to collapse subjectivity into intersubjectivity, as he appears to maintain that I recommend, we need some conception of subjectivity. To this end, he introduces the notion of intra-subjectivism in order to explain outcomes in a context that is acknowledged to be political and hence intersubjective. By intra-subjective, I take it that Hay means ideas that, however they might have been acquired or held, are specific to the individual subject in question, in the sense that no other subject is assumed to hold them in quite the same way for quite the same reasons. Put differently, we cannot derive an actor’s (intra-)subjectivity from her social context or conditioning. Actors (even similarly situated actors) are not, in short, interchangeable (ibid., 240ff). Hay argues that intra-subjectivity taken in this sense is “perfectly compatible with constructivism” (ibid., 241), and he asks whether it could be otherwise since political subjects are “at the locus” of decisions involved in “managing the tradeoffs” between “motivational dispositions” (ibid., 241).

Hay reminds us that even though he uses the familiar notion of perceived interests, he does not regard purely instrumental considerations associated with the pursuit of self-interest to be of great importance in motivating political conduct. In this respect, his critique of rational-choice institutionalism is that actors are not interchangeable, and that they take into consideration a wide range of motivations for their actions, including moral, ethical, and sympathetic concerns for the economic, political, and social well-being of others. Actors as reflective creatures may thus either conform to intersubjective pressures, or place a greater significance upon one or another motivation for action. The response and action to which this gives rise is “neither purely intersubjective nor purely intra-subjective, but a product of the interdependence of the two” (ibid., 241). This is an important point that has substantial methodological implications for neo-institutionalism. If we seek to trace institutional change, as well as the power to initiate such change, to individual actors and their perceptions of interests, we would need to get data about particular individuals.

Hay argues that if we entertain the view that social and political facts/things—potentially including institutions as well—are intersubjective by definition, then we must also investigate whether they somehow exist on the individual and subjective level. I regard the latter as not necessarily the case, since institutions are not upheld by individual minds; it is their intersubjective character that brings them to life and sustains them, regardless of any critical thinking and resistance on the part of individual actors. Obviously, the claim to have identified a specific institution may be verified by suggestions about how it is sustained by either the mindset or behavior of individual actors. In fact, many informal institutions are not codified, and they are maintained by the attitudes and actions of social actors who may have no particular critical perception of them, such as occurs with gendered norms and racist attitudes. While formal institutions may be altered by a single legislative or executive decision, the transformation of informal institutions requires gradual change that includes transforming the ways in which most people think and act.

The common feature of all versions of neo-institutionalism is that they seek to theorize, explain, and understand the relations between structure and agency. The advance Hay makes in this field is in providing space for (social) constructivism, ideas, and perceptions of interests. It is interesting to note in this respect that a recent discussion concerning game theory also devotes significant attention to cultural and social institutions (Burns et al. 2018). My criticism of Hay’s position was primarily concerned with the consequences that follow from the importance it ascribes to individual ideas and perceptions, or rather, individuals + ideas, which in fact turns ideas into instruments that can be utilized by individual actors. Jon Elster, who has supported methodological individualism, makes the ontological claim that “the elementary unit of social life is the individual human action. To explain social institutions and social change is to show how they arise as the result of the action and interaction of individuals (Elster 1989, 13). This observation is useful for restating my criticism, which is that CI, at least in the manner presented in the texts I originally investigated, has a marked tendency towards methodological individualism and voluntarism. This reduces both ideational elements and constructivism to individual interpretations and perceptions, thereby ignoring some of the more prominent propositions of the social ontology that forms the foundation for social constructivism and post-structuralism alike.

Hay’s response to my critique thus resides upon a commitment to social constructivism and the intersubjective character of facts and things that also provides space for resistance on the part of actors. But this is nonetheless accompanied by the danger of assigning too much space to individual abilities to change institutional structures, which is what we do when we move towards methodological individualism and fail to include the complex institutional settings of a socially constructed world. One specific route to institutional change is also omitted, namely, change as it can be brought about by alternative discourses through the constitutive effects of discourse upon the social world. I will provide an example in the concluding section of institutional change that was generated by changes in the discourses central to security and crisis management, which illustrates how social actors can be of assistance in a process of substantial institutional change without directly driving that process forward in a voluntarist fashion.

III. DISCOURSE: A TERM FOR CONFUSION WITHIN NEO-INSTITUTIONALISM

At first glance, the fact that Schmidt has developed an analytical framework that she terms “discursive” institutionalism (DI) suggests that my focus on discourse is redundant. However, there are important differences between DI and an alternative approach that may be termed post-structural institutionalism (PSI), based on substantially different conceptions of discourse. These ontological and epistemological differences are relevant to the preceding discussion of intersubjectivity.

Schmidt (2017, 249) regards my claim that DI displays an orientation towards subjectivism and voluntarism to be unwarranted, insofar as she herself highlights “the interconnections of the subjective and intersubjective nature of ideas in many different works, empirical as well as theoretical” (Schmidt 2017, 249). “Individuals,” she argues,

act within the context of ongoing, existing institutions even as they may seek to change or maintain them. Background ideational abilities and foreground discursive abilities operate in tandem . . . [making it possible] to think outside the institutions in which [individuals] continue to act, to talk about such institutions in a critical way, to communicate and deliberate about them, to persuade one another to change their minds about their institutions, and then to take collective action to change them. (Ibid., 252ff)

I argued that this understanding of the “discursive” overemphasizes the autonomy of reflexive agents’ ideas and actions as well as their usefulness in explaining change. It also omits subject-position and power asymmetries (Young 1996; Hendriks 2009; Mansbridge 1994). In her reply, Schmidt (2017, 253) states that she uses the term discourse “because it spans the divide between the substantive content of ideas and the interactive processes of discourse through its embodiment of both.” Although she refers to “the benefits of post-structuralism in exploring the meaning content of discourse in innovative ways” (ibid.), she remarks in her 2010 article that she stripped the word discourse of “postmodernist baggage to serve as a more generic term” that captures “the interactive process by which ideas are conveyed” (Schmidt 2010, 190). An endnote in her 2017 Critical Review article relates how “discourse” was viewed by mainstream political scientists as a dangerous word in the late 1990s and early 2000s. Schmidt says that this “should help explain [her] comment about using discourse as a generic term ‘stripped of post-modernist baggage’”: the term otherwise would have made political scientists turn a deaf ear to the importance of ideas altogether (Schmidt 2017, 261-62). She continues: “What I intended to do . . . was to develop an application of the term [discourse] in a manner that was different from the way in which it had been used before” (ibid., 262). One might thus safely say that DI diverges, quite deliberately, from some of the basic tenets of post-structuralism in that it aims to provide a different notion and application of the term. I do not understand, then, Schmidt’s unwillingness to agree that DI presents an understanding of “discourse” that differs significantly from that which is directly associated with post-structuralism.

Briefly stated, the way in which Schmidt conceptualizes discourse emphasizes the exchange of ideas and the logic of strategic communicative action. She remarks in this regard that scholars who speak of discourse address the representation of ideas (how agents say what they are thinking of doing) and the discursive interactions through which actors generate and communicate ideas (to whom they say it) within given institutional contexts (where and when they say it) (Schmidt 2008, 306). I argued in response that this conception of discourse leaves out of consideration important social processes through which actors become specific types of subjects, which involves what it means to inhabit an identity and a subject-position shaped by norms and restrictions as well as how that inevitably shapes one’s ability to engage in strategic communication (Larsson 2015b; Bacchi and Rönnblom 2014). Feminist Institutionalism (FI) has made great progress in this regard by explicitly theorizing how formal and informal institutions are gendered, requiring us to take into account how actors’ gendered perceptions of the institutional context influence their behavior and strategic choices (Chappell and Waylen 2013; Erikson 2018). For instance, studies of the Swedish Parliament have revealed that, in spite of formal rules concerning equality, men and women have different professional opportunities and encounter a broad set of norms in their everyday work associated with gender that dictate their conduct and strategies of speaking and acting (Erikson 2017). Another current example is the global #MeToo campaign, which has revealed the very widespread, and even institutionalized, practice of sexual harassment that had previously been blamed on women rather than men, silencing the former. Who speaks, and from which subject-position/identity, makes a great difference in institutionalized discourses (Epstein 2011; Phillips et al. 2004). If discourse is regarded as comprising knowledge claims, rather than merely as discursive communicative action, then discourse has different implications for power relations and resistance than Schmidt is willing to acknowledge.

It would nevertheless be unfair to suggest that she completely ignores the relations between ideas/discourses and power. In a recent article with Martin Carstenson, she provides a chart that provides theoretical insight into how ideas and power may be interrelated in order to clarify the basic tenets of DI. Carstenson and Schmidt (2016) distinguish three perspectives: power through, power over, and power in ideas. Referring to this article in her reply to my intervention, she specifies that “ideational power [is] the capacity of actors (whether individual or collective) to influence other actors’ normative and cognitive beliefs through the use of ideational elements. We note that acts of ideational power—whether successful or not—occur in only a subset of the relations relevant for understanding how ideas matter, namely when actors seek to influence the beliefs of others by promoting their own ideas at the expense of others (Schmidt 2017, 258). She then identifies three ways of theorizing about the power of ideas and discourse that involve persuasive power through ideas by means of discourse, coercive power over ideas and discourse, and structural or institutional power in ideas and discourse (Schmidt 2017, 259). Although both power through and power over ideas give substantial preference to a conception of actors + ideas, power over ideas can also be used to explain how powerless actors with limited access to resources can exert influence over others insofar as discursive means can be used to shame powerful actors to act in ways they otherwise would not have. This is the case with various progressive social movements, including the #MeToo campaign, when relatively weak actors persuade powerful actors to change not only their behavior, but even the “rules of the game.” We may say in such cases that ideational power levels the field of power relations by means of immaterial resources, a specific logic of communication, and successful framing and argumentation.

Powerful examples of change in this respect include the bans on land mines (Price 1998) and whaling (Epstein 2008), as Schmidt notes, but these in fact serve to showcase the importance of framing and linking issues to alternative discourses. As such, this requires that the actors in question are located in circumstances where foreground discursive abilities prevail, and that they know their interests and concerns, are aware of the formal and informal contexts they inhabit, and have (probably) tried a number of different approaches to convince others of their position, both allies and adversaries. It is also important that those who have the power to change existing institutions need not take full responsibility for their existence. Another instance of such a situation concerns the persistence of racial discourses and practices by means of informal institutions, regardless of the “colorblindness” of formal institutions. For example, the historical intertwining of colonial and racial dispossession in the United States continues to play a very large role in the everyday life of African-Americans and other ethnic groups in the country. Racism remains a pressing issue in spite of the abolition of slavery, the annulment of Jim Crow laws, the successes of the civil rights movement of the 1960s and 1970s, and the election of Barack Obama as the first black president in 2008. Although the Black Lives Matter campaign that was launched in 2012 aims to counter the extreme police brutality against Africans-Americans in the United States as well as socio-economic divisions, both mass incarceration and deadly violence against unarmed African-Americans continue. The March 2018 killing of Stephon Clark by police officers, who fired 20 rounds at him as he was holding a cell phone in his grandmother’s backyard, claiming they feared for their lives, is telling evidence. Ibram Kendi discusses the persistence of racist ideas and the racial consequences of policies in his Stamped from the Beginning: The Definitive History of Racist Ideas in America. He describes how while though there no longer are any official and formal apartheid structures in the United States, there is also little investigation into and understanding of norms, attitudes, and the historical legacy of previous formal and informal institutions (Kendi 2016). Alyosha Goldstein argues in this regard that anti-discrimination law, as with jurisprudence in the United States more broadly, serves to enforce evidentiary logics that weigh against adjudicating systemic culpability. A fundamental premise of such jurisprudence is that discrimination is a discrete set of identifiable and attributable acts that, even if pervasive within a government agency or institution, remain external to the logic of that agency or institution’s purpose (Goldstein 2018). As a result, each individual case and each item of antidiscrimination litigation ultimately reinforces the legitimacy of racialized institutions, along with the social, economic, and political norms predicated on the corresponding logics. Since official policy, and most likely foreground discursive abilities as well, are anti-discriminatory in character, each case or instance of racial discrimination is regarded as a violation of formal policy, and there is little willingness to admit the existence of underlying racist attitudes and related problems. This explains why there is virtually no real change in spite of overt resistance, awareness, and public debate concerning issues that are at least as important as whaling or the ban on land mines. This also illustrates the importance of subject-positions and of who speaks in relation to prevailing discourses. We should thus be cautious about limiting our understanding and definition of ideational power to the capabilities of the actors involved without accounting for the more complex interplay of powerful formal and informal discourses.

Schmidt (2017, 260) acknowledges that there is a power in ideas that may be regarded “as even more ‘powerful’ in some sense than coercive or structural power.” She maintains, however, that even though this power may be considered to be structural or institutional in character, it results from the actions of agents who have established “hegemony over the production of subject-positions, and [this] is generally the focus of post-structuralists” (Schmidt 2017, 260). She allows that inasmuch as this position recognizes that such power is a consequence of institutions imposing constraints upon ideas that agents have to take into consideration, it falls within the domain of historical institutionalism. She and Carstensen thus observe that while other forms of ideational power concentrate on the interactions between ideational agents, power in ideas mostly concerns the deeper-level ideational and institutional structures that actors draw upon and relate their ideas to in order for them to gain recognition from elites and the mass public (Schmidt 2017, 260; Carstensen and Schmidt 2016, 329-31).

It is unclear to me why power in ideas should be restricted to having an historical and temporal quality. In contrast, post-structuralism regards this type of power as making possible an exploration of the substantive content of discourses as knowledge claims that are ever-present, form the basis for subject-positions and the structuring of the social world, and are continuously being produced. The definition of ideational power provided by Schmidt thus seems to me to illustrate the central point of my argument: that DI gives analytical preference to an actor’s ability to influence another actor’s normative and cognitive beliefs through the use of ideational elements. As a result, discursive institutionalism, in spite of its name, is biased towards actors + ideas, with ideas being used instrumentally by actors in political controversies. This leaves “power in ideas” with little or no opportunity to cast light upon contemporary institutional change.

For such reasons, I continue to maintain that there is a need for a neo-institutional framework that starts from the original post-structuralist understanding of discourse as comprising knowledge claims, such that it is constitutive of the social world and gives rise to specific institutional arrangements (Larsson 2017b). The concluding section ties these elements together and provides a current example of how institutional changes can stem from discursive alterations.

IV. POST-STRUCTURAL INSTITUTIONALISM IN THE HERE AND NOW

Why is it a good idea to bring together post-structuralism and neo-institutionalism? Why not stick with unadulterated post-structuralism?

The view that discourse theory primarily provides an analysis of the power in ideas seems to me to be correct to an extent, revealing some of the shortcomings of post-structuralist analysis. Thus, on the one hand, post-structuralists often display great interest in the content and representation of realities but only later add the theoretical element of constitutive causality, which can transform a given discourse into a supposedly indisputable social fact without investigating either its intersubjective or subjective status. While this type of discourse analysis serves to help reveal knowledge claims (often in texts), it encounters difficulties in recognizing that a particular discourse may exercise constitutive causality in particular instances and acquire a hegemonic position. An increasing number of researchers therefore adopt the view that discourse analysis needs to be complemented with other methods of empirical research in order to be able to verify the mechanisms involved in such claims of causality (Marttila 2015; Bacchi and Goodwin 2016).

Some post-structuralists argue that the move towards an institutional framework is not only redundant, but violates the ethos of contingency in post-structural thought (Bacchi and Rönnblom 2014). Other scholars maintain that insofar as both discursive institutionalism and post-structuralist discourse theory comprise discursive approaches to the study of politics, they can mutually enrich their each other’s frameworks (Panizza and Miorelli 2012). In this respect, David Moon proposes that it is possible to move towards a post-structural institutionalism by integrating a variety of “discursive-constructivist approaches” (Moon 2013).

I have endeavored throughout the present discussion to illustrate how post-structural institutionalism, in contrast to Moon’s framing, involves particular ontological and epistemological disagreements with both CI and DI that cannot be reduced to simply growing “tissue on the bones” (Moon 2013). This is why I am inclined to use the term post-structural institutionalism rather than constructivist or discursive institutionalism (to add to the confusion, that latter term was already in use prior to Schmidt’s adoption of it; see Campbell and Pedersen 2001, 2). Terminology aside, the key advantage of combining post-structuralism with an institutional framework is that we can thereby substantiate the claims that a given discourse possesses constitutive causality, and that its relative influence is manifested through formal and informal institutions. We can also then argue that evidence of the power or weakness of a given discourse is provided by its relative institutionalization, which illustrates how individual perceptions, actions, and resistance become structured in relation to the substantive content of discourses (Phillips et al. 2004).

Theories of institutional change need to account for how discourse as knowledge claims affects the actors who navigate and potentially alter institutional arrangements. Different actors will obviously possess diverse capabilities for engaging in institutional change (Alasuutari 2015; Epstein 2011; Bisschops and Beunen 2018). While political, economic, and intellectual elites enjoy significant resources, have access to knowledge, engage in ongoing analysis, and may critically contemplate the rules of the game (Parsons 2016, 447), those who are uninformed, vulnerable, and opposed to the dominant ideas and discourses may possess meager resources for changing existing structures and institutions in spite of any foreground discursive abilities they may have. Focusing on resourceful elites who are consciously engaged in either institutional change or maintaining the status quo tends to overemphasize voluntarism and methodological individualism, such that all social and political actors are regarded as able to think in terms of structures/agents and act as institutional entrepreneurs (Battilana et al. 2009). More importantly, the complexity of the interplay between discourses and institutions often hinders attempts to strategically engage with, manipulate, and control the effects of new ideational elements, regardless of any attempts to alter either discourses or institutions (Beunen and Patterson 2016). We thus need a more nuanced understanding of who is in fact able to bring about institutional change and how.

Such an understanding can be fostered by incorporating post-structuralist notions of discourse and subject-position without ignoring the perceptions of interest emphasized by CI or the discursive action emphasized by DI, which can be useful to a degree in explaining certain types of institutional change. It should be noted that the relative importance of these various factors may differ from case to case, which has been illustrated by the various examples that Hay, Schmidt, and I have provided in presenting our respective theoretical positions. Perhaps it is therefore time to move the focus of discussion to empirical investigation and take advantage of the various insights garnered from examining particular cases of institutional change, in which perceptions of interests, the logic of communication, or discourse as knowledge claims may be dominant depending on the circumstances. One example from my own research illustrates the potential importance of the latter.

After the end of the Cold War, Sweden adopted a broader understanding of security that shifted attention from national security, defense of borders, and military capability to societal and human security. This change might very well have been impossible if analysis of the geopolitical situation had not indicated that the risk of war in the near future was low (Larsson 2015a, 15-17). In turn, this new conception of security made it possible for critical security scholars, through the application of securitization theory, to direct attention to security objects that differed from those with which the state had previously been concerned (Buzan et al. 1998).

Securitization theory maintains that issues, objects, and processes can be securitized by means of speech acts that become accepted by a given audience. The move away from strategic military studies to societal security also invited the participation of a broad set of private actors who either managed social issues or sought to contribute to the development of new governance strategies, marking a transition from the use of administrative hierarchies to the use of networks within a political field that had traditionally been characterized by the exercise of sovereign and absolute power (Larsson 2017a). In addition, this more general change altered the politics of citizenship. Prior to the acceptance of a broader concept of security, the rationality of the state in large part remained consistent with the ideas of Hobbes and other social contract theorists insofar as it centered on its responsibility to prevent harm to citizens. The process of responsibilization now underway in Sweden, which is led by the public authorities, instead promotes the view that capable individuals are not to presume that the state will immediately provide assistance in times of crisis or emergency, but should rather be prepared to contribute resources in order to unburden the state and public agencies so that the crisis management system can function more effectively (Rådestad and Larsson 2018). These fundamental alterations in the legal framework, governance, and governmentalities are incorporated in both formal and informal institutions that regulate policies and practices in this field. More importantly, these changes have resulted neither from actors proceeding in accordance with their perceived interests, nor from public debates among conflicting voices. They have instead emerged through the gradual public acceptance of a new security concept that was initially developed by critical security scholars who sought to direct greater attention to human rather than state security and foster a clearer understanding of the social construction of security and security objects. The case in point makes clear how substantial formal and informal changes can take place due to discursive alterations that in themselves possess an enabling logic such that social actors do not have to fall back upon perceived interests or engage in discursive conflict. One caveat in this respect is that there are very large differences among various political fields, and that security and crisis management are typically not subject to public debate. This places more or less all social actors in the same situation insofar as everyone wishes to be secure and live in a society that is safe and resilient.

It seems to me that we need to further diversify the notion of institutional change by taking into consideration the interplay between formal and informal institutions and their respective degrees of flexibility in respect to change. We also need to address how the various subjects involved either change or sustain the rules of the game; the ways in which power asymmetries intermingle with ideational elements; and, most importantly, how change is constant and immanent both within separate and across multiple discourses. For the subject content of such ideational elements as discourses may generate processes of institutional change in which social actors provide assistance rather than directly drive changes forward in a subjectivist and voluntarist fashion.