The Motivational Power of Ideas in Institutions and Collective Action

The cognitive revolution has left its mark on institutional theory in sociology and political science. Cognitive structures – schemas, typifications, frames and ideas – are recognized as a crucial variable of social behavior, institutional development and collective action. However, while the assertion that “ideas matter” is widely shared, institutional theorists are still struggling with the question of how ideas matter, especially in motivational terms. The role of ideas not just as switchmen, in Weber’s terms, but as genuine drivers of collective action still lacks theoretical underpinning. This article aims at closing this gap. It elaborates a theoretical model to explain how ideas shift the structure of motivation from the individual to the collective level. Borrowing from motivational psychology and social philosophy, especially from the work of John Searle, two crucial mechanisms are explicated. The first one is imagination, a specific mental state that allows collectivizing the intentional structure of beliefs and affords a sense of self-efficacy in collective contexts. The second one is plural self-awareness, a mechanism that shifts intentions-in-action to the collective level.


Introduction
Institutional theory has come a long way in appreciating the importance of ideas, but there are still blind spots with regard to the mechanisms that explain the effects of ideas on institutional order and change. Ideas are defined here as any structured and conscious cognitive content, including schemas, beliefs and world views communicated through discourse and held by individuals. Often, ideas are treated as a residual and ancillary concept to the wider theory in which they are embedded (Blyth, 1997). Under the rationalist paradigm, they serve as focal points and coordinative tools, thus playing a largely instrumental role. Sociological accounts focus either on normative structures or on taken-for-granted scripts and typifications. Ideas are regarded as parts of the cultural and institutional environment that reduces uncertainty about the behavior of others and facilitates social interaction (e.g., Swidler, 1986;Zucker, 1977). The sociology of organizations conceives the institutional context on a larger scale as a nation-or worldwide cultural script that constitutes and orientates individual and collective actors (Jepperson, 2001;Meyer, 2010;Meyer & Rowan, 1977). For claiming legitimate actorhood, individuals and organizations perform the roles prescribed by institutional scripts and principles. As Meyer puts it, "successful institutionalization of social patterns, in this order, involves the construction on the social stage of purposive, competent, and motivated actors who appear to choose the correct and required actions" (2010: 5).
Yet, people in institutional contexts do not just perform the roles of motivated actors -in many cases, they are motivated. Individuals conform to institutional logics not just for legitimizing their position on the "social stage," but they regard collective purposes and principles as their own. Individual voters are convinced to make an effective contribution to the election of a new government and political activists devote considerable efforts to ensure the success of their party or movement. Empirically, the functions of institutions go beyond rules and cognitive maps (in the economic account), routines and normative orientations (in the sociological account), or myths and ceremonies (in organizational sociology). Institutions do not just determine the paths and scope of action, but rather bring individuals' motivation to act in line with institutional logics of thought and action. In fact, the very core of strong institutions lies in their capacity to create volition in their members -individual volition that refers to a collective goal (Zucker, 1977). The motivational aspect of institutions runs deeper than mimesis, internalized norms or cost-benefit-calculations -it involves a shift in the structure of motivation from the individual to the collective level. In certain instances, thinking and acting of individuals takes on a genuinely collective logic, and still, it emerges from individual minds and psychic dispositions. This leaves a gap between the individual and the collective level of motivation and action, which cannot be bridged by existing institutionalist approaches. To address this gap, the role of ideas needs to be reconsidered and extended.
Ideational approaches in institutional theory can be divided broadly into two major traditions, phenomenological (or constructivist) institutionalism and ideational institutionalism. The former was established by Berger and Luckmann (1967), who drew extensively on Alfred Schütz and the sociology of knowledge to theorize the cognitive foundations of society. They described these foundations as comprising "everything that passes for 'knowledge' in society" (1967: 26), but their emphasis was mainly on the taken for granted cognitive 'matrix' on the micro-level of everyday life (Lebenswelt), which defines "what is 'real' for the members of a society" (1967: 27). Ideas are regarded by Berger and Luckmann as higher-order knowledge structures emerging from the need for legitimating the institutional order in its totality and for integrating its different sub-systems. The so-called new institutionalism in organizational sociology builds on this tradition, but its focus is principally on the cognitive macro-level -shared scripts and institutionalized structures of meaning and control in which individuals, organizations and states are embedded (Finnemore, 1996;Meyer et al., 1997). The second major approach to ideas, ideational institutionalism, is found mostly in political science. It is originally associated with the work of Peter Hall, who applied Thomas Kuhn's concept of paradigms and paradigm shifts to institutional analysis (Berman, 2013;Schmidt, 2008). Ideational institutionalism is based on a narrower conception of ideas as problem definitions or policy solutions reflecting actors' interests and serving as strategic tools of legitimation or persuasion.
While phenomenological institutionalism is concerned with the cognitive dimension of social ontology, political science treats ideas as explanatory variable. The former's emphasis is on the structural side and the compulsory force of ideas, while the latter leaves more room for agency. Yet, both approaches largely conform to Weber's switchmen conception. In the phenomenological approach, the switchmen are invisible, pre-constituted and structurally engrained, while in political science they are visible and strategically used by actors, but both regard ideas as 'tracks' of individual and collective action. This leaves one question open: How do ideas matter in motivational terms?
To address this question, this article explores the cognitive mechanisms that allow bridging the gap between the level of individual motivation and action, on the one hand, and the collective logic of thought and action ingrained in institutional structures, on the other. 1 Clarifying the motivational power of ideas also means mitigating the dilemma between macro-cognitive structuralism and micro-level interactionism. While the former omits dynamics of institutional change, the latter fails to account for institutional order and permanence. 2 Through an integrated model of action, including not only the cognitive tracks but also the motivational engine, the volitional side of collective action is conceptually endogenized without theorizing agency out of institutional theory. Both, the cognitive and the volitional components of action on the individual level, are modified in such a way as to conform to the collective logics of institutions. The corresponding mechanisms are specified on the level of individual mental processes and intentional states. For this purpose, I borrow from motivational psychology and social philosophy to describe two crucial mechanisms acting upon individual motivation, imagination and plural self-awareness, the latter of which was most comprehensively developed by Hans-Bernhard Schmid (2014).
To bring these rather complex and contested concepts to fruition within a limited space, it is necessary to cut some edges and focus on a narrow range of conceptual 1 The term 'mechanism' is used here not in the sense of 'causal mechanism' or 'social mechanism,' as discussed, e.g., by Elster (1989), but instead in the sense of conceptual building blocks. 2 For an overview of this debate see Hirsch and Lounsbury (1997) and DiMaggio and Powell (1991). referents, mainly on Searle's account of intentional states and collective intentionality (Searle, 1983(Searle, , 1990(Searle, , 1995(Searle, , 1998(Searle, , 2008(Searle, , 2010(Searle, , 2015. The argument starts from the assertion that ideas are constitutive of interests, which is largely uncontroversial among ideational institutionalists. On that ground, the concept of interest is discussed to explicate the ambivalent role of ideas in institutional theory. Considering the fact that 'interest' implies both, the notion of choice (the 'paths of action') as well as the notion of desire (the 'engine of action'), it will be argued that ideas are constitutive not just of interests-as-choice, but also of interests-as-desire -a differentiation that will be expounded in the second section. On that ground, the main question will be addressed: How can ideas become motivationally effective on the collective level?
The specification of the corresponding mechanisms follows in the third and fourth sections with reference to two basic concepts from motivational psychology and philosophy of mind: self-efficacy and intentional control. Both are essential conditions for generating motivation to act and, therefore, both need to be modified for bringing individual structures of motivation into accordance with collective logics of action. The mechanisms modifying these conditions are imagination and plural self-awareness. In the third section, it will be shown that imagination affords a sense of self-efficacy by collectivizing mental states such as hopes, beliefs and desires. These states remain individualized as long as they are tied to perceptions. Collectivizing mental states means cutting the link between cognition and perception by the way of imagination. The fourth section turns to the second step, collectivizing action, which entails a shift of intentions-in-action from the individual to the collective level through plural self-awareness. Together, imagination and plural self-awareness fundamentally alter the structure of motivation and bring about a kind of genuinely collective agency.

Ideas and Interests in Institutional Theory: from Choice to Desire
For discussing the role of ideas in institutional theory, Max Weber's switchmen metaphor provides a valuable point of departure: "Not ideas, but material and ideal interests, directly govern men's conduct. Yet very frequently the 'world images' that have been created by 'ideas' have, like switchmen, determined the tracks along which action has been pushed by the dynamic of interests" (Weber, 1959: 280). Weber speaks of tracks and switchmen as opposed to interests, thus implying that he understands these concepts as separate and distinguishable. Yet, with the expression "material and ideal interests," he introduces a further differentiation between ideas, on the one hand, and ideal interests, on the other, with the latter corresponding to non-material kinds of motivation, akin to values (Lizardo & Stoltz, 2018). Both, material and ideal interests, are seen as the 'locomotive' motivating action, while ideas or world images serve as tracks and switchmen directing action.
This dualistic view can be found in different varieties of institutional theory. In organizational sociology, it becomes visible in the 'decoupling' of formal 'tracks' of action -roles, scripts and legitimizing principles -on the one hand, and actors' interest-driven, mostly informal behavior, on the other (Meyer & Rowan, 1977). In ideational institutionalism, the (dualistic) switchmen conception is expressed, for example, in Berman's assumption that actors with similar interests, but with different ideas "will make different decisions, even when placed in similar environments" (Berman, 1998: 33). At the same time, however, a more comprehensive conception of ideas has gained currency, particularly in the constructivist branch of ideational institutionalism, which regards ideas as providing "the framework through which interests are defined" and as constitutive of the very agents that hold them (Gofas & Hay, 2010: 30; see also Blyth, 2002;Hay, 2011;Wendt, 1999).
At the core of these conceptual ambiguities is an inherent ambivalence of the concept of interest itself, which includes the notion of choice -the decisional side of action -as well as the notion of desire -the motivational or volitional side. 3 To use an example: The 'interest' in going to a restaurant to have lunch can be understood in two different ways. In a decisional sense, the interest refers to a choice between different (cognitively constituted) 'tracks' of action -eating at a restaurant, at the university canteen or skipping the meal. This decision is influenced by ideas, which not only determine the (perceived) costs and benefits of each choice but also constitute the available options. Without knowing that there is a university canteen, I cannot consider eating there, no matter how hungry I am (although I might employ heuristics to find out). Ideas in this sense are decisional devices that relate to choice. Yet, there is also a motivational side to the term 'interest,' which refers to the mental force or volitional impulse that drives action. If ideas have motivational force, as constructivist institutionalism is implying and as I am trying to show in this article, their effect must bear on this second dimension, interest-as-desire, too. In the present example, the actual intention to go to a restaurant stems from the physical feeling of hunger.
The differentiation between interest-as-choice and interest-as-desire is essential for clarifying the constitutive role of ideas. Conceptions of ideas as strategic or coordinative tools refer to interests-as-choice. They are supposed to reduce information costs on the individual level and provide shared goals and cognitive maps on the collective level. Much of the ideational literature in political science is concerned with the construction of collective "tracks of action" such as austerity (Blyth, 2013), developmentalism (Sikkink, 1991), neoliberalism (Hay, 2001;Mirowski & Plehwe, 2009), or the European monetary union (Brunnermeier, 2016;McNamara, 1998). These studies show how ideas constitute interests on the collective level and, thereby, render actors like parties or governments capable of acting collectively. Yet, the constitutive role of ideas remains confined to interests-as-choice, while interestsas-desire are attributed (implicitly, most of the time) to material incentives and to individually grounded reasons for action.
This needs not to be a problem as long as collective action is seen as an aggregation of individual wants and needs or the congruence of collective interests-as-choice with individual interests-as-desire is presupposed. Often, however, these presuppositions do not hold. As soon as genuine collective goods are involved, the classical collective action dilemma comes to the fore. For a group to develop the idea of building a pyramid or of fighting climate change is one thing, supposing the involved individuals contribute actively to the achievement of these goals is another. Collective interests-as-choice cannot be expected in themselves to be in accordance with individual structures of motivation. Certainly, the building of pyramids was not a voluntary endeavor. Nonetheless, many instances of collective action are hardly explainable in terms of compulsion and domination alone. As Weber notes, "[t]he pyramids appear preposterous unless we realize that the subjects firmly believed in the king as god incarnate" (1978: 1169). Thus, even in this case, some degree of motivation has to be presupposed in order to account for collective action.
What is needed in such cases is a motivational component on the collective level, a collectivized interest-as-desire. But how can motivation be transferred from the individual to the collective level? To elucidate the pertinent mechanisms, I will take up two concepts from psychology and the philosophy of mind: self-efficacy and intentional control. The notion of self-efficacy is central to psychological models of motivation. It is defined as "beliefs in one's capabilities to organize and execute the courses of action required to manage prospective situations" (Bandura, 1995: 2). Volition depends on beliefs about the actor's self and her capacity to act, that is, about her capability of bringing about a desired state required for satisfying her wants and needs. If the grapes are hanging too high, the intention to grab them will not arise, as in the fable of the fox (Elster, 1983;Kay et al., 2002).
In this fable, the fox's desire for grapes and his ability to reach for them are both independent and necessary conditions of action, in accordance with Hume's formula "Desire plus belief equals action". The concept of self-efficacy, however, goes beyond that formula. It implies that beliefs have an effect on desires, at least on those desires that derive from psychological needs. While basic physical needs (like the fox's hunger) can be regarded indeed as independent determinants of action, this is not equally the case with acquired or psychological needs like those described by McClelland (1988): achievement, affiliation and power. These latter needs are particularly relevant to social life and their effect on motivation does not only depend on personality traits, but also on ideas about potential goals and on beliefs about an actor's capability of achieving these goals. Desires like starting a business, writing a book or producing a piece of art will only arise if beliefs confer a sense of self-efficacy in achieving these goals. In such cases, desire is not an independent condition of action, but a function of beliefs.
In group action, the condition of self-efficacy poses a major challenge. Any type of interaction entails a measure of uncertainty -to marry, to dance the tango, to paint a house together or to carry a couch upstairs. The influence of individual action over the intended outcome is inevitably limited as soon as the actions of others come into play. How then can the condition of self-efficacy be met in such situations? Ideas offer a way out insofar as they provide shared knowledge and mutual reassurance. As an individual acting in a collective context, I can gauge the intentions of others to a certain degree from observational hints, from experiences and from a network of knowledge, often referred to as 'culture'. Shared ideas reduce uncertainty about the intentions of others and, therefore, enhance the level of perceived self-efficacy in achieving the common goal. Still, the corresponding desires remain individualized in this conception.
The second relevant concept raised above is intentional control (or control over action), which relates to a more fundamental level of motivation and action. To tap into this level, a closer examination of the concept of intentionality is required. Intentionality is conceived by Searle as "that feature of the mind by which it is directed at, or about, or of objects and states of affairs in the world" (Searle, 2015: 13). That is, intentions are purposeful operations of the mind that aim at bringing about a certain state. Intentional states, therefore, have conditions of satisfaction -truth conditions for beliefs, fulfillment conditions for desires (Searle, 1995: 129). In the case of beliefs, intentions aim at creating mental representations fitting an object in the world. In the case of desires, intentions aim at fitting the outside world to the mental content. The desire to get a house painted aims at fitting an object to the mental image of a freshly painted house. Accordingly, desires (as well as hopes or fears) have a worldto-the-mind direction of fit, while beliefs have a mind-to-the-world direction of fit. As in both cases the conditions of satisfaction are defined by the relation between the mind and the world, intentions can either succeed or fail. If the restaurant is closed, my desire to eat there as well as my belief about the availability of food have failed. Still, the intention to walk to the restaurant has succeeded, although on the ground of an erroneous assumption.
Different from beliefs and desires, intentions-in-action cannot fail. The intention to walk to the restaurant or to raise an arm is at the same time cause and effect of the intended action. It is causally self-referential, as Searle puts it, because it can only be had if it is satisfied. For that reason, intentions-in-action require intentional control. According to Baier, intentional control circumscribes the very range of intentions a person is capable of conceiving. We cannot intend what we cannot control. "I cannot intend the sun to stop, nor can I intend to turn the moon around to see its other face. Both of these, if I am ignorant or credulous or confident enough, can figure among my goals, among the things I am hoping and planning and working to bring off. The proper objects of intending, unlike the proper objects of aiming at, seem limited to my actions (not the sun's) and to things I can do" (Baier, 1970: 649). Baier calls this the "principle of the delimited sovereignty of intention". Bratman takes this notion of intentional sovereignty up and calls it 'settle condition': "I may only intend what I think my so intending settles" (Bratman, 1999: 149). I may or may not intend to raise my arm, because I can control it, but I am incapable of intending to raise the arm of some other person. I may think about it or ask for it, but I cannot intend it.
Intentional control is a more demanding condition than mutual reassurance and efficacy calculations. It is predicated on a fundamental sense of agency and ownership, which, according to Jeannerod (2003), stems from the sensual experience of the body and bodily movement. The intention to raise my arm instantly has the effect of my arm going up -control of action is nearly perfect. This kind of intentional state is different from planning and cognitively conceiving an action -it is immediate. Planning to raise my arm is a question of cognition and confidence in my future physical ability; actually raising the arm, by contrast, requires an intimate awareness of the body and a sense of intentional control. Motivation to act, therefore, hinges on both conditions, a sense of self-efficacy provided by beliefs and knowledge as well as intentional control, conferred by bodily awareness.
In this way, the motivational foundation of joint action like painting a house together can be broken down into a knowledge component and a control component. The knowledge component includes a shared idea of the desired end-state and a mutual commitment between my colleagues and me (to paint one side each, say). Without sufficient reason for believing in the contribution of my colleagues, I lack a sense of self-efficacy and, therefore, the motivation to do my part. 4 The second component is intentional control that allows forming an intention-in-action, that is, to move the body in accordance with the task to paint the color on the wall. This intention-in-action does not refer to painting the entire house, but just to executing the individual part of the common plan. This is what I intend to do and what is under my full intentional control. Intentions-in-action, therefore, still have an individualized structure. Insofar, action is not shared in a genuinely collective sense, it is rather coordinated (see Tollefsen, 2014). Painting a house together may be regarded an instance of shared action, but the fact of sharing only refers to the knowledge component.
Social philosophy pays much attention to cognitive and normative foundations of collective action such as conventions or commitments (Gilbert, 1990(Gilbert, , 2014Lewis, 1969). Shared knowledge, however, has its limitations. Searle insists that there is more to shared action than individual intentions plus some connecting mechanism based on mutual observation, commitments or conventions (Searle, 1998: 118-121). Coordinating individual actions is not enough to dance the tango, for example. Knowing or believing that my dancing partner will conform to the agreed upon dance form is something utterly different from moving the two bodies in sync with the music. Even if the dancing partners know each other and have committed themselves explicitly to dancing together, it is still too demanding a task to anticipate and observe every move of the other and to act accordingly, while, in turn, the dancing partner anticipates and observes these moves and so on. Everyone who has once had trouble getting past a person coming from the opposite direction on a sidewalk knows how difficult coordination can be even in simple situations of interaction. Dancing the tango is much more demanding and it obviously exceeds the cognitive capacities of mutual observation and adaptation. Even without practical expertise, I dare to assert that it cannot be reduced to the sum of individual contributions or the coordinated actions of individual dancers. It is collective on a more fundamental level, which involves the very structure of intentionality.
In institutional theory, it is (again implicitly, most of the time) supposed that intentional states have an individualized structure. It is assumed that to believe, to hope, to desire and to act occur in the first person singular. The core of Weber's methodological individualism is still prevalent insofar as the analyzability of social facts into individual intentional states is mostly taken for granted (but see, e.g., Epstein, 2015). In this view, ideas can be shared -as mechanisms to coordinate individual goals (in the rationalist account) or as social constructions that constitute collective actors and their interests (in the constructivist account) -but the fact of sharing does not extend to the structure of intentionality. However, acting jointly while intending individu-ally leaves a gap between the individual and the collective logics of motivation and action.
This gap becomes even wider in the case of large groups, in which not just the condition of intentional control but also the condition of self-efficacy is seemingly impossible to satisfy because individual action has no significant effect on the group level. Unlike painting parts of a house, the contribution of an individual voter to the outcome of an election is negligible -the effective influence of each single vote tends towards zero. Without self-efficiency, an essential motivational requirement is lacking. Ideas provide no solution because knowing about the intentions of others does not render my individual contribution more effective. While collective action may in its totality be effective in achieving the intended goal, no single participant has noteworthy control over this outcome. Obviously, collective action like voting cannot be reduced to individual motivation based on self-efficacy and intentional control.
What is needed is a shift in the subject structure of intentionality from the individual to the collective level. More precisely, a two-fold shift is required, because two kinds of intentional states are involved in producing motivation to act. The first shift conveys a sense of self-efficacy by collectivizing the subject-structure of beliefs and, thereby, creates interests-as-desire on part of the involved individuals; the second shift brings about intentions-in-action as a function of a collectivized kind of intentional control. The corresponding mechanisms, imagination and plural self-awareness, will be elaborated in the following sections.

Collectivizing the Structure of Intentionality: from Perception to Imagination
In the subsequent paragraphs it will be shown that the intentional structure of beliefs and desires can indeed be collectivized. To approach the corresponding mechanism, the notion of ideas-as-representations (of real-world objects) has to be substituted (or complemented) by ideas-as-imaginations.
The concept of imagination has enjoyed increasing scholarly interest in recent years - Adams et al. (2015) speak of it as a 'paradigm in the making'. The main thrust of the corresponding literature, however, is concerned with 'social imaginaries' in the sense of collective representations or institutionalized meaning systems. As such, social imaginaries are by definition collective, their emergence from the individual level is not problematized. Beckert (2016), for example, refers to imaginaries as discursively constituted 'interpretive frames' for reducing uncertainty and projecting economic decisions into the future, which comes close to the constructivist conception of ideas. In a more fundamental sense, Charles Taylor speaks of social imaginaries as a "common understanding that makes possible common practices and a widely shared sense of legitimacy," which determines how 'ordinary people' conceive of themselves as social beings (Taylor, 2004: 23). Similarly, Ricoeur stresses the "constituting nature of the social imaginary," and Castoriadis describes social imaginaries as institutional facts exogenous to the individual mind and as "underivable from the psyche in itself as such" (Castoriadis, 1987: 247;Ricoeur, 1978: 17).
While these conceptions are valuable for understanding imagination as a social fact, they lack the micro foundation required for linking social imaginaries to individual structures of intentionality. To establish this link, imagination is conceived here as a mental mechanism changing the intentional structure of beliefs and, thereby, giving way to a collective subject on the individual level. To carve out this mechanism, Searle's account of intentionality again provides a helpful conceptual lever.
Intentional states like beliefs, hopes or desires are -according to Searle (2010: 29) -representations of their conditions of satisfaction: the "conditions in the world which must be satisfied if the intentional state is to be satisfied". These conditions constitute a relationship between the mind and the world, with perceptions connecting both spheres. Intentional states are purely mental, but for being satisfied they require a perceptual link to the factual world. As reflections of real-world referents, perceptions are true by definition because their content is caused by the objects they reflect (Searle, 1983(Searle, , 2010(Searle, , 2015. Moreover, in Searle's conception, intentions depend on a network of knowledge and commitments that define the range of intentional contents. Searle calls this 'the background' and attributes to it "all of those abilities, capacities, dispositions, ways of doing things, and general know-how that enable us to carry out our intentions and apply our intentional states generally" (Searle, 2010: 31). 5 It is because of these constraints that one cannot command someone to believe or to hope something (Searle, 2010: 40).
Ideas-as-imagination are of a different nature. The claim 'Imagine!' makes perfect sense. This is not because imaginations have no conditions of satisfaction, but because they are not constrained in the same way as perceptually anchored intentional states are. Imaginations are all those mental contents that have no referent in the physical world. Examples of such contents include not just Santa Claus or unicorns, but also the demos, the nation, God, the saints, paradise, climate change, gross domestic product or the class struggle. The common denominator of imaginative mental contents is that they are not representations of states of the world. In a state of imagination, the mind is not fed by perceptions, but rather it feeds itself. The intentional content is decoupled from, or at least takes priority over, real-world referents and so do all kinds of intentional states (beliefs, fears, hopes or desires) that are directed towards the imagined objects. Only a small chunk (if any) of these objects is accessible to sensual experience -and still, in many cases their existence is taken for granted. This raises the question of how such an intentional state can be satisfied. After all, there is no perceptual input provided by the senses for satisfying such beliefs, fears, hopes or desires.
Searle describes imagination as having no direction of fit and no conditions of satisfaction. He equates imagination to fiction, stating that "the commitment to the conditions of satisfaction are deliberately suspended" (Searle, 1983: 18). 6 Yet, this would presuppose that imaginations are consciously recognized as fictions, which is often not the case. The existence of a nation or the value of money are experienced 5 From social psychology we know that mental processes are geared towards maintaining or restoring consistency in a person's knowledge, beliefs and actions (Festinger, 1962); Pettit (2003) describes this kind of rational unity as the definitional core of intentional subjectivity. 6 Similarly, Ricoeur (1978Ricoeur ( , 1979) describes fiction as a kind of conscious imagination. as real by many people. The same holds for God, the saints and paradise, which in their mental representations are often endowed with the same status as the physical world. Otherwise, imaginations would be inconsequential in motivational terms. They would be undirectional and, consequentially, ineffectual in creating interestsas-desire. Yet, imaginative intentional states such as believing in God or hoping for the well of the nation do have real consequences and, therefore, are to be regarded as full-fledged intentional states, including conditions of satisfaction.
To be sure, imagined contents are usually not completely decoupled from perceptions. After all, it is hard for any person to imagine something totally alien to her experiential universe. For example, Beckert's imagined futures are largely extensions and extrapolations from past and present experiences. Ultimately, the building blocks of imaginations stem from the physical environment. Still, there is a difference between ideas-as-representations and ideas-as-imaginations: In the latter case, perceptual inputs are selected, ignored or manipulated to fit the imagined content. While ideas-as-representations presuppose a correspondence between the mind and perceived objects in the world, ideas-as-imaginations render perceptions subservient to mental contents. Perceptions play no primary epistemic role but merely serve as a support for imagination. The use of symbols gives testimony to this logic: Objects in the world are manipulated and they are purposefully sought out to give substance to imagined contents. When someone prays to a saint, the referent of this intentional state is imaginative. The imagination of the saint is largely independent of its symbolic or figurative representation. The same is true for political imaginaries. While a substantial part of material institutional reality is made up of symbolic representations such as statues, representative buildings, emblems or flags (Jones et al., 2017), it would, nonetheless, be mistaken to regard these material artefacts as the essence of institutions. Rather, this essence is found on the level of imagination. Thus, although the link to the world is not entirely cut, imaginative intentional objects are mainly mental creations and, therefore, unconstrained by external reality.
Yet, imagination not only involves the objects an intentional state is directed at, but also the intentional subject. This becomes evident, for example, from Anderson's account of national imaginaries. Anderson points to the strong attachment "peoples feel for the inventions of their imaginations" and the fact that they are even "ready to die for these inventions". He shows that the imagination of the nation transcends individual beliefs and desires, inspiring an intimate sense of community, "the beauty of gemeinschaft" (Anderson, 1983: 141-143). Such a kind of imagination inevitably involves a shift in the subject structure of intentionality, which is not sufficiently appreciated yet in theories of imagination. Imagination is defined by the non-existence of the intentional object, while the intentional subject is supposed to be unaffected by such a state. Hoping for the blessing of God or believing in the strength of the nation are intentional states with an imagined object, but their subject structure remains individualized. It is me, as an individual, who puts my hopes in God or believes in the nation.
In such a conception, a disjunction arises between an individualized, perceptually anchored intentional subject and an imaginative intentional object. Imagination remains, so to say, incomplete, which explains the fact of why it can be recognized as fiction. One might ask, however, why the mind is capable of creating imagina-tive objects and a whole imaginative universe, while the intentional subject remains individualized. If it is possible to imagine a nation, why should it not be possible to imagine oneself being a people? Indeed, what Anderson's account of imagined communities implies is a total shift in the structure of intentionality, including the intentional subject. Such a shift allows collectivizing intentional states such as hopes, beliefs or desires, which no longer occur in the first person singular. Instead, the intentional subject takes on the form of a collective actor such as a demos, a people, a proletariat, God's chosen people or mankind -something which Taylor (2004) supposedly has in mind when he speaks of 'metatopical agency'. Subjectivity becomes collective: A praying individual imagines herself as embodying the community of the faithful and a person participating in an election takes on the subject structure of the sovereign demos, at least partially. 7 As noted by Taylor, "[p]art of the implicit knowledge that makes sense of each act of voting is our awareness of the whole action, involving all citizens, each choosing individually, but from among the same alternatives, and the compounding of these microchoices into one binding, collective decision" (2004: 24). Acting in a group becomes acting as a group; the intentional subject shifts from the individual to the collective level and renders actions like voting genuinely collective in nature.
Once the intentional subject is collectivized along with the intentional object, there is no disjunction in the structure of intentionality anymore -imagination becomes total and, therefore, consequential in motivational terms. In the realm of imagination, intentional subjects are able to move beyond the constricting requirements imposed by perceptions and background commitments. When the bond between the mind and the world is loosened, a nearly unlimited space of possibilities is opened up. This entails the creation of collective identities and collective agents with far ranging attributions. In such a state, it does not matter anymore if the grapes are hanging too high, because agents can transform themselves into imagined giants. For an imagined giant, the grapes come close; for an imagined demos, the election of a new government becomes an effective possibility. Imagination, therefore, allows resolving the puzzle of self-efficacy in large groups because the members of a group no longer perceive themselves as individuals, but as collective agents capable of bringing about the shared goal of, say, saving the nation, electing a new government or "putting a man on the moon," as in the famous anecdote about President Kennedy in Cape Canaveral.
Crucially, ideas-as-imaginations are not to be understood as exogenous forces influencing individual behavior. Imaginations should not be equated to myths or shared scripts that control human conduct or legitimate actorhood (as in organizational institutionalism). They also differ from what sociologists describe as internalized norms, schemata and meanings stemming from processes of socialization, personal interaction and conditioning. 8 Rather, imaginations are endogenous to individual structures of motivation. They are actively and deliberately appropriated from available discourse because they are experienced as empowering. They serve as enabling tools for satisficing individual wants and needs beyond the constraints of background knowledge and perceived reality. As imagined giants, people are able to reach the grapes and as an imagined community they are able to change the course of history. The mechanism of imagination, therefore, fundamentally alters the parameters of motivation and action -provided that imaginations are shared and instantiated.
Yet, as a merely mental state, imagination remains inconsequential for social reality. Social and institutional change only occurs through collective action, for which additional conditions are required. Ideas-as-imaginations are necessary for affording a sense of self-efficacy in collective contexts, but they are insufficient for bringing about manifest action. Collective action is subject to a stronger sort of constraint because to act means to engaging directly with the world.

Collectivizing Action: from Bodily Awareness to Plural Self-Awareness
So far, we have regarded intentions-in-action to be bound to bodily awareness and control. Individual action is conceived as resulting from a cognitively constituted desire (to eat at a restaurant) and a sense of bodily control that allows an intention-inaction to arise (setting the body into motion to walk to the restaurant). This conception leads to trouble as soon as an imagined community becomes an acting group. While imagination allows collectivizing the cognitive component of individual action orientations, this seems to be impossible with the control component of action. Imaginations may well be detached from perceptions and bring about a collective actor on the level of (individual) minds and (shared) discourse, but when physical action is involved, illusions about a collective intentional subject must invariably collapse. If an intention-in-action is tied to bodily awareness, it cannot be other than individual intentionality of the form 'I intend'. Perceptions of bodily movements and sensual experiences all buttress the individual character of action. In bodily experience, individuality imposes itself. Given these natural limitations of intentional control, how can it be that imaginations have real consequences for political life and institutional development? How can it be that an imagined We gives way to a manifest collective actor?
To approach this question, we may consider the phrase 'What if everyone did that'. When people recur to that phrase, they seem to experience their own action as integral part of a larger context of collective action, which would break down if they would stop acting (Elster, 1985). We also know from empirical observation that individuals in collective contexts behave differently from isolated individuals and that collective or shared activities are experienced not as activity in a group, but as activity as a group. Dancing the tango is something fundamentally different from dancing alone; playing soccer on a team is different from running or swimming. Even if collective activities are cognitively framed as individual contribution to a group activity ("what did you do yesterday?" -"I played soccer with my friends"), they are experienced as genuinely collective activities on a more fundamental level. What is actually experienced is better described as "we played soccer". In such a state, a group can be said to become an acting body.
Such a shift from the I-mode to the We-mode of acting is referred to as collective intentionality, which is subject to an intense debate. A major controversy revolves around the question whether the state of collectivity resides inside individual minds or, rather, is constituted by some external mechanism like conventions or agreements. 9 Here, the former position is taken by assuming that collective intentionality is internal to individual minds. Nonetheless, Searle understands collective intentions as irreducible to individual intentions: " [It] is not the same as the summation of individual intentional behavior " (1990: 402). This seeming paradox is spelled out by Searle in this way: "[Collective intentionality] is of the form 'we intend' even though it is in my individual head " (1998: 119). It is phenomenologically collective, but it is created within individuals. In Searle's words: "It is not just that I am doing this and you are doing this, but we are doing this together; and this fact is represented in each of our heads in the form of collective intentionality" (2008: 446).
Collective intentionality may be understood as synchronization and merging of action orientations as can also be observed in the animal world, in herds or swarms. Searle describes this mechanism through the use of an example, the joint preparation of a sauce hollandaise, by analogy with polynominal individual actions such as the firing of a gun: "[J]ust as I fire the gun by means of my pulling the trigger, so We make the sauce by means of Me stirring and You pouring". Just as the action of firing a gun constitutes a single intentional operation, so the preparation of the sauce is to be understood as a unitary process: "We intend to make the sauce by means of Me stirring and You pouring" (Searle, 2010: 412), whereby the pouring of my cooperation partner constitutes not an external contribution to our common endeavor, but an integral part of it. In such a state, individual contributions are not perceived and experienced as distinct from group action. To be sure, collective actions are carried out, ultimately, by individuals and they usually come with a share of individual intentionality. Steps like putting on a jacket, leaving the house, walking to the polling place and so on are intended individually, but they are derivative of an overall collective intention. I intend to cast my ballot -but only because and insofar as we intend to elect a new government.
Collective intentionality implies that intentional control is not contingent upon physical limitations anymore. Just as I can intend to raise my arm, we can intend to defeat a competing team or to build a pyramid. This intuition is supported by evidence from experimental psychology, which shows that bodily awareness can be manipulated and detached from the physical body (e.g., Blanke & Metzinger, 2009;Lenggenhager et al., 2007). The fact that metaphorical references to the body are omnipresent in political contexts -body politics, the legislative body, the electoral body, the nation as body -points to the analogous nature of bodily awareness and the sense of common agency and ownership observable in social life. 10 We may think here of a phenomenon akin to what Schmid calls 'plural self-awareness' and what he describes as a "background awareness of plural selfhood," prior to any conscious kind of intentionality, "a sort of mental integration in which the participants are aware of some of their attitudes as theirs, collectively" (Schmid, 2014: 18f.). Indeed, the concept of plural self-awareness provides a theoretical avenue for apprehending the collectivization of action and, therefore, resolving the problem of intentional control in collective contexts, including macro-institutional contexts.
Yet, plural self-awareness is not directly observable and, moreover, it is often diluted by the fact that group members tend to account for group action in individual and aggregative terms. They eventually recall an event as "I played soccer with my friends," although in fact the soccer game was a genuinely collective experience. Thus, the state of plural self-awareness, on which the collectivization of action rests, has to be inferred indirectly from empirical instances of collective intentionality. It reveals itself with particular clarity during exceptional episodes of collective arousal or 'moments of madness,' in Zolberg's terms, which are often (but not necessarily) associated with crucial historical turning points. In such moments, intentionality transcends the limits of individual cognition, which is experienced by group members as a state of enthusiasm and joy. This state, supposedly, stems from a kind of liberation and empowerment, when the constraints imposed by the individualizing sense of bodily awareness are removed. From Zolberg's account of crucial episodes in French history -May 1968, the Paris commune of 1871, the Revolution of 1848 or the French Revolution of 1789 -the typical features of plural self-awareness become evident: "[W]hen the carefully erected walls which compartmentalize society collapse," a state of "political harmony" arises. "Minds and bodies are liberated; human beings feel that they are in direct touch with one another as well as with their inner selves" (Zolberg, 1972: 186, 196).
It is striking how strongly the testimonies of such episodes resemble each other. Zolberg refers to Henry Lefébre, who notes with view on the Paris Commune of 1871: "[S]ocial action wills itself and makes itself free, disengaged of constraints. It transforms itself in one leap into a community in whose midst work, joy, pleasure, the achievement of needs […] will never be separated". For the Revolution of 1848, he cites Georges Duveau: "[O]ne experiences on the morrow of the proclamation of the republic, an extraordinary impression of freedom, of happiness, of fulfillment". Quite similarly, the crucial turning point of the French Revolution, the session of the National Assembly of August 4, 1789, is described as "a kind of magic" by François Furet (1996) or as a "patriotic delirium" by Michael Fitzsimmons (2003). According to one of the participants, the session "had produced something that twelve centuries of the same religion, the same language, and the habits of common manners had not been able to accomplish, which was the reconciliation of interests and the unity of France toward a single objective -the common good of all" (Fitzsimmons, 2003: 22).
Moments of madness are by no means exclusive to France, as Zolberg at the end of his essay guesses. Turning to a different context, the Iranian Revolution of 1979 provides very similar expressions of collective arousal and joy. When Ayatollah Khomeini returned to Iran in February 1979 after his years in exile, his arrival prompted the final stage of the Islamic revolution. Mass mobilizations and massive desertions of military and police forces led Prime Minister Shapour Bakhtiar to abdicate and to leave the country, as still ruler Reza Shah Pahlavi had already done in January. Heikal et al. describe the Ayatollah's arrival as "an occasion of unbridled religious rejoicing, for which there has probably been no parallel in the modern world.
[…] People were shouting 'The soul of Hussein is coming back!,' 'The doors of Paradise have been opened again!,' 'Now is the hour of martyrdom!' and similar cries of ecstasy" (1981: 177). The former system of legitimation and authority broke down immediately and even communist and nationalist groups, who later would be persecuted by the Islamist revolutionaries, recognized the new leader.
Although exceptional and transitory, these instances of mobilization lay bare the mechanism of plural self-awareness, which allows bridging the gap between the individual and the collective level of motivation and action. Plural self-awareness (conceived as plural bodily-awareness) creates a sense of collective intentional control as it occurs in physically or communicatively assembled masses and, thereby, allows an imagined community such as the people of Paris or the Shia community to be instantiated in manifest collective action. Just as a blow of a whistle turns a soccer team into a collectively acting, unified body, the return of Khomeini converted the demonstrating masses in Teheran into a revolutionary movement and the speeches of two notables turned the National Assembly into the embodiment of the French Nation on August 4, 1789.
The same mechanism, although in a less intense mode, is operative in democratic polities. An imaginatively pre-established collective subject based on ideas of popular sovereignty and partisanship is instantiated by the synchronizing effect of an electoral campaign and the date of the election, which creates a common focus for collective action. Individual voters put on their jackets, walk to the polling station and contribute to the genuinely collective goal of electing a new government. Together with the background awareness of plural selfhood and the focusing event of the election, imaginations of a sovereign demos produce the kind of collective intentionality necessary for overcoming the collective action dilemma ingrained in macro-institutional contexts.

Conclusion
For institutions to bring individuals' motivation to act in line with institutional logics of thought and action, they need to create volition in their members. Ideas are the key to understanding the motivational force of institutions. Ideational institutionalists have asserted that ideas not only constitute 'tracks' of action, but that they are constitutive of the very interests driving action. Up until this point, however, we have not known why this is the case and how ideas exert motivational force on the collective level.
In this article, I proposed a theoretical model for explaining the role of ideas in creating motivation on the collective level. This model relies on theoretical building blocks from social philosophy and motivational psychology, namely on imagination and plural self-awareness, which allow modifying basic parameters of intentionality and motivation. Imagination modifies the structure of intentionality by shifting the intentional subject to the collective level, thereby resolving the problem of self-efficacy in collective contexts. A similar shift is produced by plural self-awareness with regard to intentional control, thereby collectivizing intentions-in-action. As a consequence of this two-fold shift, an imagined We becomes manifest reality. While imagination frees the mind of consistency requirements and perceptual limitations, plural self-awareness allows for action beyond physical, bodily limitations through the transformation of group members into a unified body.
Conceived in that way, not as scripts and typifications imposed on actors, but as empowering tools, ideas provide a link between micro-and macro-levels of institutional analysis. Moreover, the mechanism of imagination brings agency back into institutional theory without recurring to exogenous factors like power or institutional entrepreneurship. Imagination and plural self-awareness rest on the same conceptual bases as phenomenological and ideational institutionalism. The proposed model, therefore, opens up an avenue out of cognitive structuralism and towards an explanation of institutional change without relaxing the ideational paradigm.
Yet, ideas alone are not enough. As historical examples have shown, for manifest collective action to take place, an additional factor is needed: an external focus synchronizing attention and action orientations. Without such a catalytic stimulus, a collective subject remains in a state of latency, like tango dancers without music, a soccer team without a ball or an orchestra without a director. When the music starts, dancing couples emerge out of individual dancers; a whistle turns individual soccer players into an acting group and a sign by the conductor creates a musical body out of individual musicians. In a democracy, the equivalent of a whistle or a conductor's sign is the date of the election, which brings about collective action orientations based on the shared imaginary of a sovereign demos.
Still, without a team of soccer players who know how to play the game and who conceive of themselves as team-members, any whistle whatsoever will remain inconsequential. Equally, even the best director will not be able to produce a symphony without a pre-established orchestra. In the same way, an election alone will not suffice to produce a sovereign demos. What is needed is a shared idea of common goals and common identities as well as a latent sense of belonging and membership on the part of each citizen. Ideas-as-imagination are crucial for constituting this kind of collective actor. They bring about a collective intentional subject beforehand on the level of the mind and provide the cognitive structures necessary to act collectively on a large scale. Without a collective imaginary, an essential element is missing. Institutions without ideational foundations lack the capacity of creating volition in their members. If this is the case, people will no longer ask 'What if everyone did that'. They will cancel their membership in a political party, stop reading the newspaper and stay home on election day because their reasoning is 'What difference does my contribution make?'. When people start asking this question, the motivational force of ideas has disappeared and institutional structures are in danger of collapsing like the ancient regime did during the French Revolution.
Funding Open Access funding enabled and organized by Projekt DEAL.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/.