The “Hub-and-Spoke” Metaphor: Early Dyadic Perspectives

Stakeholders are defined as groups or individuals that have a direct or indirect relationship (whether based on interest, interdependence, or power) with a focal actor, generally the firm. Initially formalized through the seminal work of Freeman (Reference Freeman1984; see also Freeman, Harrison, Wicks, Parmar, & De Colle, Reference Freeman, Harrison, Wicks, Parmar and De Colle2010; Freeman, Harrison, & Zyglidopoulos, Reference Freeman, Harrison and Zyglidopoulos2018), early perspectives on stakeholders sought to identify and catalog the major groups of actors that could affect (or be affected by) a firm—for example, customers, employees, suppliers, the media, or governments. This understanding consisted of a “hub-and-spoke” image of ties between the focal firm and stakeholders that was, by construction, dyadic. Subsequent research elaborated on this metaphor, highlighting how, for instance, two actors within the same stakeholder group could potentially have different relationships with the focal firm (Bridoux & Stoelhort, Reference Bridoux and Stoelhorst2014; Shymko & Roulet, Reference Shymko and Roulet2017). Instead of simply defining which stakeholders “counted” (Dawkins, Reference Dawkins2015), management theorists focused on the characteristics that made them salient to a focal actor (Mitchell, Agle, & Wood, Reference Mitchell, Agle and Wood1997; Neville, Bell, & Whitwell, Reference Neville, Bell and Whitwell2011).

Since then, stakeholder theorists have sought to reveal how firms may most effectively react to the pressures they face from stakeholders (Frooman, Reference Frooman1999). For example, instrumental stakeholder theorists propose that organizations build a “relational ethics strategy” that is based on either communal relationships or more transactional- and market-oriented exchanges (Bridoux & Stoelhorst, Reference Bridoux and Stoelhorst2016; Jones, Harrison, & Felps, Reference Jones, Harrison and Felps2018). Both perspectives have implications for where the firm should position itself in its relationship with its constituents for efficient stakeholder management but are still, at their core, based on Freeman’s hub-and-spoke image. The perspectives subsequently fall short in considering how stakeholders connect with or through each other, independently from their link with the focal firm. As Rowley (Reference Rowley1997: 890) notes, “each firm faces a different set of stakeholders, which aggregate into a unique pattern of behavior.” Accordingly, more recent contributions have proposed an alternate metaphor of a “web” (Welcomer et al., Reference Welcomer, Cochran, Rands and Haggerty2003), in other words, a network reconceptualization of stakeholder arrangements.

The “Web” Metaphor of Networks: Conceptual Merits and Limitations

Network research emphasizes the importance of going beyond dyads (Burt, Kilduff, & Tasselli, Reference Burt, Kilduff and Tasselli2013; Kilduff & Brass, Reference Kilduff and Brass2010; Tasselli & Caimo, Reference Tasselli and Caimo2019) based on two observations of stakeholder behavior: 1) the influences of different stakeholders over a focal actor necessarily overlap and interact with each other (Frooman, Reference Frooman1999) and 2) the salience of a stakeholder to a focal organization is not necessarily direct because intermediary parties mediate the relationship with and among stakeholders (Neville et al., Reference Neville, Bell and Whitwell2011; Roulet, Reference Roulet2020). Network characteristics like social capital (Cots, Reference Cots2011; Okazaki, Plangger, Roulet, & Menéndez, Reference Okazaki, Plangger, Roulet and Menéndez2020) are therefore important considerations for understanding stakeholder arrangements: Rowley (Reference Rowley1997), for instance, proposes that the density of the stakeholder network surrounding an organization leads to increased pressure while the organization’s centrality enables more resistance to this pressure. The network perspective therefore offers important explanatory power: Hein, Jankovic, Feng, Farel, Yune, and Yannou (Reference Hein, Jankovic, Feng, Farel, Yune and Yannou2017), for example, consider how organizational goals are potentially achieved not only through resource provision and exchange between the firm and stakeholders but also through facilitating exchanges among stakeholders themselves.

However, while a network image has proven generative for understanding stakeholders as part of a broader web, there are numerous conceptual shortcomings, in particular with capturing the dynamism of stakeholder arrangements. This omission is important for multiple reasons. First, dyadic relationships and structure are interrelated and coevolve. Bothello and Mehrpouya (Reference Bothello and Mehrpouya2018), for instance, illustrate how one intermediary actor shapes the field of urban sustainable development through its dyadic relationships with standard-setters, which carries implications for all other actors within the field. Second, the trajectory of coevolution of dyad and structure may be underpinned by teleological factors that are not captured by a network view, for instance, when actors coalesce to deal with larger problems. In an illustrative case, Haas (Reference Haas1989) shows how cleanup of pollution in the Mediterranean occurred through epistemic communities where actors were, in many cases, disconnected from each other. Third, a network perspective is not geared toward conceptualizing temporal dynamics with stakeholders, such as rhythm and pace, for instance, the frequency of firm announcements to stakeholders or the timing of environmental investments (Bowen, Johnson, Shevlin, & Shores, Reference Bowen, Johnson, Shevlin and Shores1992; Nehrt, Reference Nehrt1996). Fourth, and crucially, understanding network characteristics does not aid in assessing the appropriateness of how stakeholders coordinate to achieve a goal.

These conceptual constraints inhibit exploration of many recent empirical phenomena, especially around how stakeholders coalesce around addressing grand challenges (George et al., Reference George, Howard-Grenville, Joshi and Tihanyi2016). For instance, multistakeholder initiatives (MSIs) like the Roundtable for Sustainable Palm Oil, the Forest Stewardship Council, or the Marine Stewardship Council are constructed around a specific environmental problem and involve a variety of parties that coordinate to collectively create responsible behavior (Mena & Palazzo, Reference Mena and Palazzo2012; Moog, Spicer, & Böhm, Reference Moog, Spicer and Böhm2015). Members of such initiatives are considered stakeholders of an issue rather than of an organization (Saffer, Reference Saffer2019). Yet, in focusing on nodes and ties, a network perspective contains certain blind spots: the moral legitimacy of MSIs, for instance, is treated as the outcome of stakeholders being tightly or loosely coupled to each other (Rasche, Reference Rasche2012), when legitimacy may in fact serve as an antecedent to rather than a consequence of network structure.

These omissions therefore require a different perspective on stakeholder arrangements. A revised conceptualization would ideally retain the strengths of the network view in analyzing the structural arrangement of stakeholders while also incorporating the analytical depth of studying dyadic links. It would also potentially shift away from reifying the organization as the focal actor and instead consider how social actors act as a collective toward solving a social issue (Werhane, Reference Werhane and Phillips2011). We propose that the most promising path is to extend recent work on system perspectives in organization studies (Shipilov & Gawer, Reference Shipilov and Gawer2020).

Developing a Stakeholder Systems Perspective

In his classic treatise Metaphysics, Aristotle (Reference Sachs1999: 8.6: 1045a) discusses how entities are often formed with properties that extend beyond a simple aggregation of individual parts. Von Bertalanffy (Reference von Bertalanffy1950: 417) formalized this simple idea into systems theory, defining a system as “a set of elements standing in interrelation among themselves and with the environment.” He initially applied his insights to account for the functions of biological organisms before extending them to other forms of systems as diverse as computer programs, traffic flows, and economies. The premises of systems theory were particularly useful for explaining organizational and managerial phenomena (Kast & Rosenzweig, Reference Kast and Rosenzweig1972; Parsons, Reference Parsons1951; von Bertalanffy, Reference von Bertalanffy1972), culminating in chaos theory and complexity theory (Allen, Maguire, & McKelvey, Reference Allen, Maguire and McKelvey2011; Anderson, Reference Anderson1999; Anderson, Meyer, Eisenhardt, Carley, & Pettigrew, Reference Anderson, Meyer, Eisenhardt, Carley and Pettigrew1999).

Organizational systems can be understood as arrangements of actors that have purposes but are nonetheless interrelated, coordinating internally and adapting to their environment to achieve a central goal (Lai & Huili Lin, Reference Lai, Huili Lin, Scott and Lewis2017). Beyond organizations, systems thinking has also produced important contributions for business ethics, for example, by “circumvent[ing] what often appear to be intractable problems created by systemic constraints for which no individual appears to be responsible” (Werhane, Reference Werhane2002: 33) and by decentering stakeholder analysis (Werhane, Reference Werhane and Phillips2011).

Importantly, systems thinking is versatile enough to consider collectives of actors at different levels of analysis from the dyad to the entire system—as well as the subsets in between (Kast & Rosenzweig, Reference Kast and Rosenzweig1972; Werhane, Reference Werhane2002). As a consequence (and in contrast to the stakeholder network view), a stakeholder systems perspective considers existing interorganizational arrangements as more than just a set of firm–stakeholder relationships. Systems are by nature adaptive (Werhane, Reference Werhane2008) and therefore have the conceptual bandwidth to simultaneously examine the dynamic properties of both dyadic ties and the overall structure, as well as the recursive relationship between the two.

We define the term stakeholder system as a set of actors who perform various roles and functions but collectively mobilize toward the achievement of a larger social goal.Footnote ³ A systems approach acknowledges the interplay between dyadic ties and overall structure, enabling us to explore the dynamic of this interplay and consider effectiveness across different levels. In considering both components simultaneously, we can also theorize about how the two aspects coevolve. Conceptual and methodological limitations have thus far prevented such an examination, despite stakeholder theory being key to understanding the challenges faced by organizations and society (Margolis & Walsh, Reference Margolis and Walsh2003).

Considering the abstract nature of stakeholder systems and the potentially limitless number of variables that could be considered to understand them, we turn toward the use of metaphors to select key dimensions. We do not aim at being exhaustive but rather use those dimensions to illustrate the value of a stakeholder systems approach. Metaphors offer some flexibility in our thinking because of the “plurality and openness in meaning” and interpretation they enable (Cornelissen, Reference Cornelissen2005: 753). Standard theory building largely draws on metaphors from natural and social sciences, as well as from the humanities, with some of the most valuable contributions coming from powerful parallels with the performance arts (Biehl, Reference Biehl2017). Thus we propose that examining the rich artistic material of ballet choreography—through the lens of dance theory—can reinvigorate accounts of stakeholder engagement.

Motivating the Metaphor of Dance and Ballet Choreography

Given the diversity of systems that exist (von Bertalanffy, Reference von Bertalanffy1968), a stakeholder systems view requires a specific metaphor for us to demonstrate its generative value. As such, we proceed to explore the potential of dance—and ballet choreographies more specifically—to fill that role. Dance has already been used to explain organizational phenomena: Harrison and Rouse (Reference Harrison and Rouse2014), for instance, explore how creativity and elastic coordination occur in modern dance groups. With respect to stakeholder dynamics, scholars have drawn passing parallels to dance. The interaction between parties has, for example, been labeled as a “tango” (Kraan, Hendriksen, van Hoof, van Leeuwen, & Jouanneau, Reference Kraan, Hendriksen, van Hoof, van Leeuwen and Jouanneau2014; Tasselli & Caimo, Reference Tasselli and Caimo2019): in tandem with stakeholders, the focal organization coordinates and aligns action to produce a form of collective performance.

We propose that this metaphor has the potential to extend stakeholder theory for multiple reasons. First, dance contains temporal patterns, such as pacing and rhythm, that may be highly relevant for understanding how firms and stakeholders interact (Barnett, Reference Barnett2007; Bowen et al., Reference Bowen, Johnson, Shevlin and Shores1992). As Hanna (Reference Hanna1987: 30) observes, dance has a “motor” comprising elements like accent (stress/intensity), duration (length of time for movements), meter (grouping of beats/accents), and tempo (rate or speed of sequential movements). Another key distinguishing feature is that choreography is a system where action is purposive and purposeful rather than routine: Kurath (Reference Kurath1960: 234) notes that “out of ordinary motor activities, dance selects, heightens or subdues.” Third, dance contains a judgment of aesthetic value that complements the conventional focus on economic value in stakeholder studies (Freeman et al., Reference Freeman, Harrison and Zyglidopoulos2018): the quality and competence of a dance must be determined by some referent group, which allows for inclusion of norms and values around what stakeholder actions can be considered appropriate. In a broad sense, then, choreography informs the coordination of actors in social and physical spaces (Harrison & Rouse, Reference Harrison and Rouse2014; Kolo, Reference Kolo2015), thereby offering a fruitful domain of inspiration.

Ballet specifically is a rich metaphor compared with other forms of dance for generating new conjectures on the dynamics underlying stakeholder systems: the aesthetic value is based on dyadic dependence between dancers—the dance of one individual is affected by and affects the movement of others—as well as an overall choreographic arrangement (Hanna, Reference Hanna1987). More pragmatically, given its cultural hegemony in the West, the various subgenres of ballet have, for better or worse, been explored in dance studies more extensively than other forms, such as swing or hip-hop. With ballet, we are therefore furnished with conceptual and empirical material to distinguish highly diverse formats and rules—or lack thereof (Anderson, Reference Anderson2018)—which allows us to contribute to stakeholder theory.

The implications for understanding the dyadic relationships between actors and the dynamic of the system as a whole are not captured by other social systems—even those in the arts. For instance, orchestras may provide insight into characteristics like harmony and tempo, but how one performer is expected to play is largely predefined, with few dyadic exchanges among musicians and no spatial separations. By contrast, choreography acknowledges that the spatial positions of actors and their actions are relational to others’ and that this apparatus as a whole is what produces aesthetic value. In this sense, ballet offers a “representation that form[s] approximations of the target subject under consideration and that subsequently provide[s] the groundwork for extended theorizing” (Cornelissen, Reference Cornelissen2006: 1579).

Dance Theory as a Bridge between the Ballet Metaphor and Stakeholder Theory

Our metaphor (Cornelissen, Reference Cornelissen2005, Reference Cornelissen2006) enables us to characterize the dynamics experienced by systems of stakeholders and the dimensions across which those systems can be compared and understood. The material of ballet choreography provides a unique source that can be transferred toward the target domain of stakeholder theory, thereby enabling a dialogue between dance choreographies and stakeholder theorizing. We propose that dance theory (Hanna, Reference Hanna1987) acts as a bridge, facilitating interpretation of the first domain and its application to the second. Specifically, dance theory allows us to interpret and characterize a variety of ballet choreographies, applying those characteristics to understand stakeholder systems.

Dance theory is the body of frameworks to codify, produce, and interpret dance performances (Hanna, Reference Hanna1987). In other terms, it is a “frame of reference” composed of “the laws, forms and power structure” that give sense to those performances, specifically the articulation between bodily movements, space, and narrative (Brandstetter, Reference Brandstetter2014: 199). Yet, dance theory also emphasizes as much “fluid[ity] and open[ness] in our thinking as possible” (Burt, Reference Burt2000: 130); there is therefore no formal theorization as we understand it in the field of management, with hypotheses or propositions. Hanna (Reference Hanna1987: 187) refers to dance theory as an assessment of aesthetic value, or the appropriateness of what is staged when it involves interpreting characteristics of performances, of what makes them remarkable, and of what connects them. As new concepts and aesthetics have emerged, giving rise to new ways of rendering artistic performances, the evaluation and new ways of appreciating them evolve in parallel, whether for public audiences, architects, or performers (Brandstetter, Reference Brandstetter2014) Nonetheless, dance theory, despite its differences with theorization in other fields, is not only for dance scholars but also targeted at creating a “discursive context” around the dancing praxis (Burt, Reference Burt2000: 125). It thus serves as a way to label a material practice in order to better understand it, reproduce it, and apprehend it in its broader social context.

Dance theory has been used outside the field of cultural production and has found many applications in management science, business, and organization studies. Among the most important contributions, the work of Biehl (Reference Biehl2017) explores the power of dance theory to derive takeaways for collaboration or unpack relational leadership. A significant body of work bringing together dance theory and organization explores questions of individual embodiment (e.g., Slutskaya & De Cock, Reference Slutskaya and De Cock2008). Other pieces explore group interactions: for example, Rowe (Reference Rowe2008) uses dance as a metaphor to inform the mechanisms of team learning. Kolo (Reference Kolo2015) explores the potential contributions of choreography to organizational research, noting its focus on living together in time and space.

Setting the Stage: Ballet and Its Various Subgenres

Ballet is a performance dance that originated in fifteenth-century Italy before spreading to France and Russia (Homans, Reference Homans2013; Paskevska, Reference Paskevska2005). It combines both music and dance and was initially performed in chambers at royal courts by and for nobles. Strongly influenced by the schools of dance across those different geographies, the richness of ballet stems from a process of striking evolution where subgenres and stylistic approaches branched off without replacing prior styles (Sgourev, Reference Sgourev2021). For instance, the emergence of neoclassical and contemporary ballet did not erase the legacy of classical and Romantic ballet. Subgenres coexisted, informed each other, and addressed overlapping audiences (Sgourev, Reference Sgourev2015). Nowadays, ballet performances usually rely on a set of professional dancers who have undergone expert training, and performances of classical and Romantic ballet are still sought after as much as newer subgenres.

We cover a range of seven paradigmatic ballet performances across four subgenres: classical (Swan Lake, The Nutcracker), Romantic (La Sylphide, Giselle), neoclassical (Apollo), and contemporary (Clytemnestra, Isadora). (Anderson, Reference Anderson2018). We select those performances because they represent a diversity not only in the style of dance but also in the structure and dynamics of choreography. Though the different subgenres have borrowed characteristics from each other, the manner in which they combine different attributes in each case produces a distinct style of dance that informs subsequent theorization. In supplementary materials published online, Appendix A provides illustrative photographs of one ballet in each subgenre, and Appendix B provides links to videos of four ballets exemplifying the dynamics discussed in the article.

The most well-known type of ballet is also the first to emerge in Italy, France, and Russia: the classical ballet. This subgenre is therefore considered traditional and based on an established repertoire of movements, choreography, and story lines, supported by elaborate costumes and sumptuous sets (Sgourev, Reference Sgourev2015). Swan Lake and The Nutcracker are considered as the epitome of classical ballet and are characterized by large ensembles and a rigid narrative structure. The aesthetic value of this subgenre is derived from the technical execution of moves, with dancer interactions that tend to be frequent and symmetrical (Anderson, Reference Anderson2018).

Romantic ballets, on the other hand, despite having numerous overlapping elements with their classical counterparts, are distinct in their focus on the role and the performance of the dancers, with emotions as a source of aesthetic. These ballets emerged in the mid-nineteenth century, almost two centuries after the first academy of ballet, which formalized classical ballet as a professionalized art, was founded in Paris (Homans, Reference Homans2013). As with classical ballets, Romantic ballets are driven by linear storytelling and focus on technique but are distinguished by their emphasis on more episodic interactions between the characters. The Romantic subgenre highlights the prima ballerina, and the narratives often revolve around spirit characters, many of whom are female. La Sylphide is often considered as the first Romantic ballet, with a plot revolving around a young Scotsman’s infatuation with a forest spirit and his tumultuous relationships with his fiancée and best friend. Giselle is the second Romantic ballet we review and is the story of a woman who is revived by female forest spirits after dying of heartbreak.

With the passage of time, and at the turn of the 1910s, ballet productions started becoming more adventurous and experimental, in particular under the influence of Sergei Diaghilev and the Ballets Russes (Sgourev, Reference Sgourev2015). As with many cultural productions during this era, such ballets were expressions of opposition to the Romantic movement (Anderson, Reference Anderson2018) and were aimed at reviving and reinterpreting elements from traditional ballet. This new subgenre became known as neoclassical ballet and included spectacles like Apollo by George Balanchine, arguably the most well-known neoclassical choreographer. In contrast to Romantic ballets, the scenery and the costumes of neoclassical ballets are muted or in some cases absent, while narratives are more abstract and nonlinear. The dancers rely more on core body movements, with interactions that are physically asymmetric, combined with emergent connections that break away from formalities common in classical and Romantic ballets. In such ballets, dancers interact more through continuous opposition than through communion (Paskevska, Reference Paskevska2005). In the case of Balanchine’s Apollo, the story is oriented on the Greek god Apollo and three muses.

Contemporary ballets, finally, combine elements of classical ballets with more recent aspects of dance derived from jazz and mime (Paskevska, Reference Paskevska2005). They also rely on physical performances but are faster, reintroducing narrative centrality and incorporating acting during the performance. Clytemnestra by Martha Graham is thus characterized by energy and borderline violent dance moves, supported at the same time by a conspicuous narrative. The last performance we mobilize is Kenneth McMillan’s Isadora, a tribute to the life of dancer Isadora Duncan. Isadora is often seen as an act rather than a ballet, oriented around the life of the dancer and her tumultuous relationships with her partners.

The seven ballets we selected are emblematic of their subgenres; they were dominant at different epochs yet now coexist as iconic performances. They capture, in our view, the diversity of ballet choreographies; the striking differences across styles can be used to generate theoretical insights into the nature of stakeholder arrangements, as well as into organization–stakeholder relationships. In the following section, we proceed to translate those key characteristics into multiple dimensions that can apply to stakeholder engagement.

Structural and Dyadic Characteristics of Ballet (and Stakeholder) Systems

Structural Dimensions

Polyphonic versus unified narrative. The classical and Romantic ballet performances are typically choreographed as a unified narrative, generally oriented around a principal dancer; effectively, this style organizes the rest of the dance troupe and can be readily communicated to the audience. The Romantic ballets of La Sylphide and Giselle, for instance, follow the story of a principal female spirit character and their relationships with a human male. Alternatively, neoclassical ballet (in our case, Apollo) relies on a multiplicity of intricate narratives, often in the background, as the performance relies less on a unified plot than on a confluence of distinctive and competing narrative threads. Thus a dance can also be choreographed as a polyphonic narrative, with a “multiplicity of independent and unmerged voices and consciousnesses … each with equal rights and its own world [that] combine, but do not merge, into the unity of an event” (Bakhtin, Reference Bakhtin1984: 208). In those pieces, the aesthetic value is derived from following multiple perspectives on the same event.

We conceive these as two ends of the same spectrum: applied to a stakeholder system, this dimension reveals the extent to which multiple stakeholders share a narrative with a focal actor, such as a firm. At one end of the spectrum, we have a unified narrative, where the entire stakeholder system approaches an issue from the same perspective, with consensus around definition, diagnosis, and/or goals. This can be driven by a focal organization that leads how sense making occurs around an issue; for instance, Bothello and Mehrpouya (Reference Bothello and Mehrpouya2018) discuss how a nongovernmental organization (NGO) named ICLEI created a framing of sustainable urban development that other NGOs eventually adopted. However, a unified narrative can also emerge in the absence of direct ties: Haas (Reference Haas1989, Reference Haas1992) illustrates how stakeholders mobilized toward regulation of ozone-depleting substances and Mediterranean pollution through similar understandings of cause-and-effect relationships, despite being disconnected from each other.

At the other end of the spectrum, when causal attributions are more complex, multiple divergent perspectives within a stakeholder arrangement can emerge, leading to a polyphonic narrative (Belova, King, & Sliwa, Reference Belova, King and Sliwa2008; Hargrave, Reference Hargrave2009). Different pockets of social actors within the stakeholder system perceive an issue differently, and their own perspective evolves independently from other narratives. Reinecke and Ansari (Reference Reinecke and Ansari2016) highlight how polyphony is endemic to these so-called wicked problems as they observe the diverging perspectives of various actors in the context of conflict minerals in Democratic Republic of Congo. These actors not only possess differing ideas of prescriptions to a problem but also disagree on the diagnosis of the problem itself.

Prompted versus emergent coherence. The second dimension articulates a key aspect common to both choreography and stakeholder systems, namely, the motivation that drives the arrangement to coalesce in the first place. Classical and Romantic ballets tend to be highly deterministic and scripted; appropriateness is based on technical proficiency, where dancers’ moves are programmed and propelled by the narrative. In contrast, the neoclassical and contemporary ballets are based more on an organic evolution rather than on adherence to a scripted routine; the appropriateness of the choreography is contingent on reactions to and from other characters. For instance, in neoclassical ballets like Apollo or contemporary pieces like Clytemnestra, the appeal of the performance is partly derived from physical and emotional spontaneity, stemming from proximity to other dancers.

These characteristics inform our second dimension of coherence. While the previous dimension of narrative focuses on the process of ongoing sense making and storytelling, coherence is about whether actors are “pushed” or “pulled” together. Applied to stakeholder theory, actors may be pushed together or prompted to form an arrangement, for instance, when a crisis threatens to harm an organization and/or its constituent stakeholders (Pearson & Clair, Reference Pearson and Clair1998). The COVID-19 pandemic, for instance, spurred the creation of the COVAX facility to ensure equitable distribution of coronavirus vaccines, an initiative involving Gavi (an alliance of for-profit vaccine producers), the Coalition for Epidemic Preparedness Innovations, and the WHO (McAdams, McDade, Ogbuoji, Johnson, Dixit, & Yamey, Reference McAdams, McDade, Ogbuoji, Johnson, Dixit and Yamey2020). Alternatively, an arrangement can cohere in the absence of such prompts, forming more organically through common interests that “pull” together to address an important—albeit not necessarily urgent—problem. These emergent stakeholder arrangements can still be facilitated by external actors to catalyze problem solving: for example, we observe how supranational actors, such as the UN, play an essential role in “orchestration,” creating platforms and fora for stakeholders to self-organize around a specific issue (Abbott & Snidal, Reference Abbott and Snidal2010; Andonova & Levy, Reference Andonova and Levy2003). We conceptualize these systems as the other end of the spectrum, where an arrangement coheres in the absence of a shock (Aakhus & Bzdak, Reference Aakhus and Bzdak2015; Allen et al., Reference Allen, Maguire and McKelvey2011; Anderson, Reference Anderson1999). This is a common process in the case of MSIs, which usually come together to solve specific issues, such as climate change (de Bakker, Rasche, & Ponte, Reference de Bakker, Rasche and Ponte2019), even if there is no short-term threat.

The two structural dimensions that we outline here are distinct yet complementary: the narrative dimension captures the evolving discursive aspects of the system, while coherence focuses on the triggers driving the emergence of the system. We observe a range of combinations: a stakeholder system can, for instance, emerge organically without forming a unified narrative; some initiatives around addressing a social issue feature a diversity of frames or viewpoints that do not need to be reconciled over time (Reinecke & Ansari, Reference Reinecke and Ansari2016). On the other hand, a prompted system can create a unified narrative despite the relevant actors coming from diverse backgrounds. Table 2 summarizes the four structural conditions of a stakeholder system and showcases illustrative examples of each.

Table 2: Structural Dimensions of a Stakeholder System

Dyadic Dimensions

In contrast to the first pair of dimensions, which focus on structural characteristics, our second pair relates to the dyadic ties between the focal organization (principal dancer) and its related stakeholders (supporting dancers). Contingent on the arrangement, these insights potentially extend to relationships between supporting dancers themselves. In the seven ballet performances we included, we focused on the properties of the interactions between individual dancers. In particular, the “importance of movement (the fact of bodily action) and motion (illusion and residual action resulting from the kind of movement produced)” lies in the way they connect individual performers (Hanna, Reference Hanna1987: 37). Within the discursive space created by the performance (Brandstetter, Reference Brandstetter2014), the interaction between dancers also signals the balance of power. Such interactions not only determine the immediate performance but also contribute to the unfolding of the narrative (Burt, Reference Burt2000; Roulet, Reference Roulet2020). We use dance theory to outline these dimensions, then proceed to apply them to firm–stakeholder relationships as well as to linkages between stakeholders.

Continuous versus episodic interactions. The first of these dimensions is based on the continuity of interactions: relationships may be episodic and sporadic, or they can be continuous. Classical ballets, such as The Nutcracker, follow the same duo of characters—Clara and the Nutcracker—and make them the focal point of the entire performance. In La Sylphide, the interactions between the main characters—the Sylph and the Scotsman—are more intermittent as they are brought together and taken apart several times throughout the ballet. Part of the aesthetic value of these performances is contingent on the rhythm, pace, and frequency of these interactions. By contrast, Clytemnestra and Isadora feature complex webs of characters who interact intensely but irregularly with each other.

Similarly, the frequency of firm–stakeholder relationships can be mapped as two ends of a continuum. Episodic interactions of dancers are likely to be ad hoc, triggered by unpredictable issues that emerge in the relationship and irregularly increase stakeholder urgency (Mitchell et al., Reference Mitchell, Agle and Wood1997), for instance, product recalls or supply chain breakdowns. Accordingly, the episodic nature of these relationships implies transactional and market-based collaborations (Jones et al., Reference Jones, Harrison and Felps2018), as those episodes usually call for swift resolutions—or at least attempts at them. At the other end of the continuum, in the same way that two dancers remain entangled for the whole performance, we can imagine firm–stakeholder relationships that are continuous. Those are more likely to be labeled as working relationships (Welcomer et al., Reference Welcomer, Cochran, Rands and Haggerty2003) that are stable and consistent over time, often based on norms of mutual reciprocity (Jones et al., Reference Jones, Harrison and Felps2018).

Symmetric versus asymmetric power. Across the different ballet performances, we consistently observe characters influencing the movement and behavior of others (Daly, Reference Daly1994). For instance, Clara frees the Nutcracker from a spell, while the Sylph in La Sylphide entrances the Scotsman, drawing him into the forest. By and large, though, Romantic and classical ballets usually involve highly symmetrical performances, with the dancers engaging in reciprocal interactions. By contrast, neoclassical and contemporary ballets often rely on asymmetrical dancing that often includes physical and symbolic dependence; Apollo, for instance, features the main dancer hoisting and holding aloft a muse on his back for several minutes, while in another scene, he “grants” space for the muses to take center stage and express themselves in turn.

When transferred to the realm of stakeholder relationships, we argue that this dimension of balance captures the extent to which one party exercises power and influence over another in a relationship (Dawkins, Reference Dawkins2015; Hillman, Withers, & Collins, Reference Hillman, Withers and Collins2009), usually through control over crucial material, human, financial, and symbolic resources. As Mitchell et al. (Reference Mitchell, Agle and Wood1997) note, certain stakeholder arrangements may be characterized by power asymmetry that is either firm dominant or stakeholder dominant, where the actor providing resources uses this ability to control the relationship and dominate decision-making. Yet there also exist instances of firm–stakeholder relationships in which both parties are mutually dependent on each other (Ahmadsimab & Chowdhury, Reference Ahmadsimab and Chowdhury2021; Ashraf, Ahmadsimab, & Pinkse, Reference Ashraf, Ahmadsimab and Pinkse2017). This symmetry within a relationship usually accompanies a bidirectional sharing of resources.

Taken together, our theorization of the two dimensions of dyadic relationships within stakeholder systems acknowledges and integrates a wider range of relationship dependencies and dynamics beyond what is covered in prior studies (Bundy, Vogel, & Zachary, Reference Bundy, Vogel and Zachary2018). Importantly, the nature of these dyadic characteristics applies to both firm–stakeholder relationships and stakeholder–stakeholder linkages. Our conceptualization, therefore, does not necessarily consider the firm as the center of the system but rather as one actor within a larger constellation of actors oriented around a particular issue. Table 3 summarizes the four conditions of dyadic relationships within a system and illustrates those four situations with relevant examples.

Table 3: Dyadic Dimensions of Stakeholder Relationships

Key Assumptions to Identify Dynamics of Stakeholder Systems

Although we distinguish between structural characteristics and dyadic relations of stakeholder arrangements, a systems perspective also highlights how the two sets of dimensions are interrelated. Lawrence (Reference Lawrence, Greenwood, Oliver, Sahlin-Andersson and Suddaby2008), for instance, suggests that influence and force—which are individual-level expressions of power—condition and are conditioned by institutional-level forms of power, such as domination and discipline. A systemic perspective therefore acknowledges the importance of feedback loops (Levasseur, Reference Levasseur2004) as changes in the subpart of the system affect the system as a whole. Conversely, dynamics of the system trigger change at the micro-level. We can imagine the interaction between structural and dyadic dimensions to be recursive: structural dimensions would affect dyadic elements, which would in turn affect structural dimensions, and so on. This recursive cycle might well contribute to the disruption of the system but also potentially to its maintenance (Islam, Zyphur, & Boje, Reference Islam, Zyphur and Boje2008).

Accordingly, we proceed to theorize how the dynamics in structural characteristics and those at the dyad level potentially co-constitute each other. On one hand, dyadic ties are affected by structural characteristics of the overall arrangement—for example, more ties can be created when the overall system is larger. On the other hand, some higher-level system characteristics can be derived from the attributes of components interacting (von Bertalanffy, Reference von Bertalanffy1972: 411)—this implies that an aggregation of dyadic changes may result in changes to structural characteristics. Although there may be a multitude of potential paths linking structural and dyadic dimensions of stakeholder systems, we focus on what are potentially the most salient connections. Accordingly, the insights we develop are based on the characteristics of two specific dynamics of stakeholder systems. They are meant, not to be exhaustive or to capture all potential evolutions of stakeholder systems, but simply to illustrate the explanatory power of our model to formulate relevant scenarios.

Illustrating the Dynamics of Stakeholder Systems

We proceed to highlight two potential dynamic paths that stakeholder systems may take (Figure 2): the first illustrates how structural and dyadic dimensions are interrelated and coevolve in an emergent system, while the second depicts the same process, but in a prompted system. In doing so, we can point to the recursive relationships between dyadic and structural dimensions in our model, while outlining how both evolve over time.

Figure 2: Potential Dynamic Paths in Stakeholder Systems

In situations of emergent stakeholder systems, stakeholders may be brought together organically through, for instance, shared interests or geographical and political proximity. Initiatives with voluntary participation and proximity in interests and understanding are likely to bring together actors that share at least some common understanding of the issue, indicating a unified narrative at the system level. Early on in such a system, we propose that this results in balanced relationships and continuous interactions. However, this dynamic may not necessarily persist. Numerous studies in the area of private and transnational governance highlight a process where stakeholders convene to address a social or environmental problem but, over time, fragment into competing interests with different views of the problem (Bothello & Mehrpouya, Reference Bothello and Mehrpouya2018). While this may not necessarily happen, an emergent stakeholder system coupled with a unified narrative may risk a fragmentation spiral. If the stakeholder system does not converge toward potential solutions, differences between stakeholders’ perspectives are more likely to emerge (Gutiérrez, Márquez, & Reficco, Reference Gutiérrez, Márquez and Reficco2016), leading to the fragmentation of the narrative. In turn, such fragmentation is likely to generate imbalance in dyadic relationships as positions become more entrenched and interactions more episodic. Recursively, those degraded dyadic relationships further fragment the narrative into polyphony.

There are several empirical accounts of such a spiral. Fransen (Reference Fransen2011; Fransen & Conzelmann, Reference Fransen and Conzelmann2015) reveals how, for instance, the fragmentation of private sustainability standards in forestry and clothing partially stemmed from regulatory actors that became too stringent—an issue which hints at dyadic power imbalance between the regulator and the regulated. In this situation, we witness a fragmentation spiral triggered by the effect of deteriorating dyadic relationships on structural dimensions of the stakeholder system. Yet the reverse can also occur, where narrative plurality leads to power imbalance: Marques (Reference Marques2019) demonstrates how transnational governance settings that are highly fragmented can, in some cases, result in regulatory capture by industry and trade associations. This therefore shifts the power balance in dyadic relationships, asymmetrically favoring business interests.

In the case of a prompted system, on the other hand, stakeholders may be cast together (perhaps unwillingly in situations of crises) and obliged to interact and coordinate. Different actors struggle to make sense of the issue or phenomenon, specifically with respect to cause–effect relationships and potential solutions (Reinecke & Ansari, Reference Reinecke and Ansari2016). In such situations, actors produce different perspectives on a problem and lack the experience of productive dialogue and exchange (Daudigeos, Roulet, & Valiorgue, Reference Daudigeos, Roulet and Valiorgue2018). As a result of this plurality of narratives, we expect dyadic relationships to be, at this early stage, episodic and asymmetric; this may in turn reinforce polyphony at a system level.

Over time, though, the repeated interactions between stakeholders in this newly formed system may enable mutual learning and thus homogenization—about both the issue at stake and others’ perspectives. Hargrave (Reference Hargrave2009), for instance, discusses how moral imagination can facilitate collective action toward an overarching goal, even among competing interests. As stakeholders coordinate in this prompted system, they may compromise or rethink their initial positions (Hill & Jones, Reference Hill and Jones1992), facilitating convergence on a unified narrative (de Bakker et al., Reference de Bakker, Rasche and Ponte2019). We argue that, recursively, this unified narrative will positively influence stakeholder relationships. First, this will increase the continuity of the relationship as partners develop a common language, perspective, and purpose (Mena & Palazzo, Reference Mena and Palazzo2012). Second, because of this unifying narrative, stakeholders can learn, coalesce, and reorganize themselves to minimize their dependencies and increase their negotiating power (Gutiérrez et al., Reference Gutiérrez, Márquez and Reficco2016). In turn, those more frequent interactions caused by their embeddedness in a developing system will further the unification of the system’s narrative.

This convergence spiral is one of the potential dynamics of a prompted stakeholder system. As the stakeholder system is forced into collaboration and suffers from polyphonic narratives, the dyadic level is characterized by episodic interactions and asymmetric interactions. At this stage, stakeholders necessarily struggle to understand each other’s position and rationale and what they have to offer to each other. As the narrative becomes more homogenous, and stakeholders learn from each other, interactions become more frequent and balanced.

An example of this process can be observed in the stakeholder systems that work toward solving climate change (Ansari, Wijen, & Gray, Reference Ansari, Wijen and Gray2013; Schüssler, Rüling, & Wittneben, Reference Schüssler, Rüling and Wittneben2014). Ansari et al. (Reference Ansari, Wijen and Gray2013), for instance, highlight how a disparate set of actors with distinct logics were brought together to address the problem of climate change. Through a series of field-configuring events, these actors collectively created an overarching consensus of the problem in addition to solutions, such as the Kyoto Protocol. Although coming from diverse perspectives, they learned from each other, compromised, and accepted each other’s system of rationality. Thus they converged toward a unified narrative, also benefiting dyadic relationships, which recursively furthered a common systemic understanding and approach.