In the previous decade, the literature on “relational approach” has burgeoned in the social sciences. Recently, a “relational turn” in political science was called for in a symposium in this journal (January 2011). The participants perceived a promising path for such a “turn” by the introduction of social network analysis (SNA) into political science. This call is informed by a conviction that the central concept of political science—that is, power—is “at its very core, relational” (McClurg and Young Reference McClurg and Young2011, 39). Considering this viewpoint, I contend that there are two different understandings of the connection between the qualifier “relational” and the concept of power. I refer to them as “Anglo-American” and “Continental” relationalism. The labels designate certain ideal types of arguments, not geographical locations (although the obvious connotation with a similar division in philosophy is intentional). With no intention to downplay either the contribution of SNA or the importance of the symposium, I nevertheless argue that the participants presented only the Anglo-American perspective on relationalism and that introducing the Continental understanding would add value for political science.
Whereas the symposium was largely about methodology, this article contributes mostly to political ontology (cf. Hay Reference Hay, Goodin and Tilly2006), considering epistemological and methodological issues only briefly. Both contributions are equally important for a comprehensive discussion of “relational political science” (McClurg and Young Reference McClurg and Young2011). Therefore, closing the gap of clarifying the two alternative ontological underpinnings of “relational political science” might contribute to the development of such an approach more generally. Hence, I am not only introducing the respective views of relationalism but also describing the potential for a dialogue between them.
To distinguish between the Anglo-American and the Continental understandings of relationalism, I adapt to political science the vocabulary of the most-cited meta-theoretical paper on “relational sociology” (Emirbayer Reference Emirbayer1997) and a more recent contribution (Dépelteau Reference Dépelteau2008). Both scholars borrow from Dewey and Bentley (Reference Dewey and Bentley1949) and distinguish among three views of social action: “self-action,” “inter-action,” or “trans-action.” I argue that partly the first but especially the second view are characteristic of Anglo-American relationalism. The “trans-actional” view is specific to the Continental perspective.
After outlining a potential route for a dialogue between these two perspectives, I conclude that neither is superior per se and that each has a legitimate scope of application.
THE ANGLO-AMERICAN PERSPECTIVE
Anglo-American theorists of power are Hobbesian “methodological individualists.” They perceive social action as “self-action,” presuming that social entities operate according to their own powers that “exist outside or prior to social relations” (Dépelteau Reference Dépelteau2008, 60). Various perspectives on power—ranging from international relations theory to political geography—that treat it as a “resource” or “potential” are self-actionalist.
Perceiving social relations as “inter-actions” presumes that “the relevant action takes place among the entities themselves. Entities remain fixed and unchanged throughout such interaction” (Emirbayer Reference Emirbayer1997, 285–6). Such an inter-actionalist perspective on power typically is characterized as “relational” by Anglo-American authors. It also is captured by McClurg and Young (Reference McClurg and Young2011)—the editors of the symposium mentioned previously (and tacitly or explicitly shared by all of the participants Footnote 1 )—who draw mostly from the “faces of power” debate. For them, a starting point is Dahl’s (Reference Dahl1957) understanding according to which power consists in someone getting someone else to do something that the latter would not otherwise do. For McClurg and Young (2011, 39), Dahl’s understanding makes a clear case “that power is relational”—that is, it “only exists when considering interactions between and among individuals and groups.” They even claim that although “political scientists have challenged Dahl’s simple definition and formulation of power, the concept has always remained relational” (McClurg and Young Reference McClurg and Young2011). Thus, for them, Bachrach and Baratz’s (Reference Bachrach and Baratz1962) two-dimensional and Lukes’s (Reference Lukes2005) three-dimensional understandings of power represent “relational view[s] of power” (McClurg and Young Reference McClurg and Young2011, 40). Consequently, Lazer, the most cited author in the symposium, can argue that “power is intrinsically relational: it flows from the capacity to affect other actors” (2011, 66).
This usage of the word relational differs from the Continental usage, the origins of which lead back to Marx and Simmel and were later developed by the schools of Elias, Foucault, and Bourdieu. Continental relationalism perceives social action as trans-actional.
THE CONTINENTAL PERSPECTIVE
The trans-actional view acknowledges the “primacy of process” (Dépelteau Reference Dépelteau2008, 62) in the social world. Therefore, the “unit” of political analysis should be conceptualized as a “complex joint activity, in which it makes no sense to envision constituent elements apart from the flows within which they are involved (and vice versa)” (Emirbayer Reference Emirbayer1997, 289). For instance, one cannot intelligibly speak of a subordinator/power wielder (A), and a subordinated/power recipient (B), and their relations as something “added” to or “characteristic” of those “creatures.” It is exactly this aspect that often is bracketed in the Anglo-American approaches that also present themselves as “relational,” but that view social relations in terms of inter-actions or self-actions.
The difference becomes clear when the Anglo-American view is contrasted with that of Clegg (Reference Clegg1989), who based his argument on various Continental thinkers (e.g., Michel Foucault, Ernesto Laclau and Chantal Mouffe, Michel Callon, Bruno Latour, and others). Unlike inter-actional and self-actional views, a starting point for analysis for Clegg would not be agency but rather “the relational field of force in which power is configured and in which one aspect of this configuration is the social relations in which agency is constituted” (Clegg Reference Clegg1989, 207). The fundamental question is how those “actors” and their “agency” are constituted in a “relational field.” The Anglo-American relationalism focuses almost exclusively on inter-actions or self-actions of already constituted actors. Nevertheless, contestants in the “faces of power” debate, on whom the Anglo-American relationalists base their respective understandings, have different potentials for a dialogue with Continental relationalism. The next section focuses only on the classic participants in this debate—Dahl, Bachrach and Baratz, and Lukes rather than later developments in the schools revolving around those thinkers.
Initiating a dialogue between the Anglo-American and the Continental relationalism perspectives can assist both in overcoming their mutual misunderstandings and contributing to the development of “relational political science.”
TOWARD A DIALOGUE
Initiating a dialogue between the Anglo-American and the Continental relationalism perspectives can assist both in overcoming their mutual misunderstandings and contributing to the development of “relational political science.” In this treatment of the “faces of power” debate, my argument is conceptual rather than historical: I move from the perspective with the smallest potential to that with the largest for a dialogue with the Continental/trans-actional view that is within the “faces” framework.
The “Faces of Power”
Lukes’s contribution—historically the most recent and often depicted as the most developed—has the least potential for such a dialogue because his “third dimension/face” of power is self-actional, consisting in A’s securing B’s compliance by controlling B’s “thoughts and desires” (Lukes Reference Lukes2005, 27). In the second edition of his most eminent book, Lukes moves even further away from this potential by identifying the “exercise fallacy” (a notion he borrows from Peter Morriss [Reference Morriss2002, 16–18]). It is committed by those who view power as existing only to the extent that it is exercised (ibid., 70) and ignore the idea that power refers to “an ability or capacity, which may or may not be exercised” (Morriss Reference Morriss2002, 109). However, in the trans-actional/Continental perspective, imagining unexercised power is like speaking of winds that do not blow (cf. Elias Reference Elias1978, 112). Hence, it is no surprise that Foucault, for instance, is firmly committed to “exercise fallacy”: “Power exists only when it is put into action” (Foucault Reference Foucault1982, 788). In this sense, Foucault allies with the earlier “faces” theorists: Dahl, and Bachrach and Baratz.
Although Dahl sought to develop conceptual tools for expressing “the amount of an actor’s power” that “can be represented by a probability statement” (1957, 203), his perspective nevertheless has important aspects that are promising for a potential dialogue with Continental relationalism: that is, its presumption that power is a “relation among people” (Dahl Reference Dahl1957), not merely a property or potential. Although he believed that this relation can be analyzed by breaking it down into pieces of self-action or inter-action, power existed for Dahl—as for Foucault—in practice, to the extent that it is exercised. This brings us to the strongest potential for a dialogue with Continental relationalism found within the “faces” debate: more explicitly than Dahl, Bachrach and Baratz were committed to the “exercise fallacy.”
The most celebrated work of Bachrach and Baratz (Reference Bachrach and Baratz1962) is a contribution of enduring relevance even more than 50 years later. However, it is their later paper in which they explicitly criticized self-actional (i.e., Hobbesian) understanding of power as a “possession which enables its owner to secure some apparent future Good” (Bachrach and Baratz Reference Bachrach and Baratz1963, 632). Their staging of the problem is strikingly similar, even in its details, to a host of Continental relationalists who came decades later (cf. Elias Reference Elias1978, 74; Emirbayer Reference Emirbayer1997, 291; Foucault Reference Foucault1980, 88). These scholars would all subscribe to Bachrach and Baratz’s negative contention regarding “the fundamental relational attribute of power: that it cannot be possessed” (1963, 633). However, their allegiance to Bachrach and Baratz’s positive conception of relationalism would be considerably weaker. Bachrach and Baratz presented their conception in terms of three “relational characteristics” of power (i.e., a, b, and c).
We can disregard the first term (a) that “there is a conflict over values or course of action between A and B” (Bachrach and Baratz Reference Bachrach and Baratz1963, 635) because it is not specific to either the Continental or the Anglo-American perspective.
This apparently is the case with the third term (c) as well: “a power relation can exist only if one of the parties can threaten to invoke sanctions” (Bachrach and Baratz Reference Bachrach and Baratz1963, 633). However, Bachrach and Baratz specified that the availability of sanctions gives A power over B only when four additional conditions apply. Essentially, those conditions establish that power over B is not possible if B does not respond properly to A’s threats—basically, if she does not understand them and/or does not take them seriously (Bachrach and Baratz Reference Bachrach and Baratz1963). We see here, an Anglo-American inter-actionalism with a certain self-actional flavor because both A’s and B’s identities, values, and rationality are perceived as fixed prior to their power relation, and as being something that can be possessed. For instance, they maintained that “in a situation involving power, B is rational in the sense that he chooses compliance instead of defiance because it seems the less of two evils” (638).
A similar tendency seems to be even more pronounced when we move to “relational characteristic” term (b), which stipulates that “a power relationship exists only if B actually bows to A’s wishes” (Bachrach and Baratz Reference Bachrach and Baratz1963, 633). However, the reason that I consider this condition last is that it is the most “Continental” of the three because the authors added immediately: “if B does not comply, A’s policy will either become a dead letter or will be effectuated through the exercise of force rather than through power” (ibid., italics added). This second part of the formulation opens the possibility for a Foucauldian insight that relations of violence and power are antithetical. This is the debate that actually revolves around the issue of how to conceive the relation between power and resistance. It is explicitly present in Foucault’s oeuvre but more implicitly in that of Bachrach and Baratz.
Power and Resistance
What is the difference between power and force? One aspect by which Bachrach and Baratz distinguish them is that force “in some circumstances…is non-relational” (1963, 636). Basically, this means that, in its extreme, A’s successful act of force requires no action from B. In making this remark, Bachrach and Baratz are similar to Foucault, for whom two aspects are necessary for the existence of power relations: (1) thorough recognition of the power recipient “as a person who acts”; and (2) opening up of “a whole field of responses, reactions, results, and possible inventions” (Foucault Reference Foucault1982, 789).
Thus, Bachrach and Baratz, as well as Foucault, seem to assume that the possibility of resistance is a crucial condition for the existence of power relations. What, then, is the difference between those perspectives? Basically, it is the difference between conceptualizing power and resistance in either self-actional and inter-actional terms or in trans-actional terms. The latter presumes perceiving power/resistance relations as processes the elements of which are intelligible only in view of “the flows within which they are involved (and vice versa)” (Emirbayer Reference Emirbayer1997, 289). Resistance, as Foucault stated, “is never in a position of exteriority in relation to power” (1978, 95). From the inter-actionalist/self-actionalist perspective, this would not make sense because the subjects who either resist, consent to, or exercise power are presumed to be given. From that viewpoint, stating that resistance is not exterior to power is comparable to stating that there is no longer any “possibility of people being more or less free from others’ power” (Lukes Reference Lukes2005, 108). For Foucault, however, this would be “misunderstanding the strictly relational [i.e., trans-actional] character of power relationships” (1978, 95). The latter’s “existence depends on a multiplicity of points of resistance” that “are present everywhere in the power network” (1978, 95).
This point makes it unavoidable to highlight an aspect of the Continental perspective that—for the sake of expository clarity—has been bracketed thus far: unlike with self-actions and inter-actions, trans-actions of A and B cannot be considered in isolation but rather always as a part of wider networks of interdependent relations (i.e., with C, D, E, and so forth that also are constituted in these networks)—themselves considered as trans-active. The notion of interdependent relations is important because it is a core notion for all network approaches to politics, ranging from theories of governance and policy networks and actor–network theory (ANT) to Manuel Castells-inspired theories of network society. However, this notion also holds true for SNA. Because it is an approach that is presumed (in the symposium mentioned previously) to be the key “[f]or a truly relational political science to succeed” (McClurg and Young Reference McClurg and Young2011, 41), the next section discusses its specificity and potential routes for a dialogue with Continental approaches.
SNA and Continental Relationalism
Network approaches have different interpretations of interdependency. Most are inter-actional; some are trans-actional (e.g., ANT, Elias-inspired figurational approaches, and some poststructuralist/Foucauldian approaches to governance and governmentality); and some are clearly self-actional (e.g., rational-choice approaches to governance and policy networks). The majority of SNA is inter-actional.
Essentially, this means adopting the picture of elements and their relations that was previously problematized from the viewpoint of trans-actionalism: there are As and Bs as actors (i.e., nodes), and their relations (i.e., ties) are something “added” to them.
It is a truism that the explicit aim of SNA is to distance sociopolitical analysis from the self-actionalist focus on attributes and to set the “primacy of relations over atomised units” (Marin and Wellman Reference Marin, Wellman, Scott and Carrington2011, 22) at the forefront of analysis. However, it is equally true that SNA is “motivated by a structural intuition based on ties linking social actors” and “draws heavily on graphic imagery”—to name the two most specific characteristics of the four listed by Freeman in the most influential history of the movement (Freeman Reference Freeman2004, 3). Essentially, this means adopting the picture of elements and their relations that was previously problematized from the viewpoint of trans-actionalism: there are As and Bs as actors (i.e., nodes), and their relations (i.e., ties) are something “added” to them. There is no presumption of mutual (re)constitution of those elements within the very relations conceived as dynamic processes (i.e., as trans-actions). Of course, this might be perceived as a pragmatic methodological choice: mathematical methods are at the heart of the approach and there is no ontological presumption of a world composed of countable bits and pieces (i.e., nodes and ties). This may be the case, but there recently was an important attempt to articulate the underlying theory (not only methodology) of SNA in the most canonical outlet on the topic by authors led by one of the most cited scholars in the field. Borgatti and Lopez-Kidwell (Reference Borgatti, Lopez-Kidwell, Scott and Carrington2011) argue that there are basically two fundamental models of social reality (ontologies, in my term) of which major theories in SNA are variations: the network-flow model and the network-architecture model. The two models have differences but, metaphorically speaking, they presume the ties (i.e., relations) to be akin to either (1) “pipes,” “through which things flow” from one actor to another; or (2) bonds “that bind the network together” (Borgatti and Lopez-Kidwell Reference Borgatti, Lopez-Kidwell, Scott and Carrington2011, 47). Relations, therefore, are uniting or mediating actors, not constituting them. Relations (i.e., ties) are not processes but rather “things” that the actors (i.e., nodes) “have.” To have the “primacy of process” in the sense envisioned in trans-actional perspectives, both nodes and ties (or actors and their relations) should be considered as mutually constitutive processes, rife with contingencies and indeterminacies. In this view, SNA is definitely inter-actional. Is it a problem per se? Not necessarily.
However, I contend that this approach to networks must be supplemented with other approaches if the aim is to grasp power as a process of mutual constituting of social actors and/in networks. This aim does not necessarily need to contradict the “graph-theoretically” informed inter-actionalism or Anglo-American relationalism of SNA.
I return to the notion of power to explain this viewpoint. In formulating his “circuits of power” model, Clegg engaged thoroughly with the Anglo-American “faces” debate for articulating an element of his model that he called “episodic power” and the corresponding “episodic agency.” The paradigm for “episodic power” was, for him, none other than Dahl’s understanding that McClurg and Young (Reference McClurg and Young2011) characterized as clearly “relational.” For Clegg, the Dahlian “one-dimensional conceptions of agency and power may tell us something about the nature of power relations between an already constituted A and B” (Clegg Reference Clegg1989, 208–209). This is exactly the notion of power and agency suitable for the inter-actional SNA type of analyses as demonstrated by contributions to the symposium (Lazer Reference Lazer2011; Perliger and Pedahzur Reference Perliger and Pedahzur2011; Siegel Reference Siegel2011; Sokhey and Djupe Reference Sokhey and Djupe2011). Yet, this does not have to be the entire explanation. In addition, we can move on to “the constitutive nature of the relational field in which A and B presently are” and to “how this privileges and handicaps them respectively, in relation to those resources that are constituted as powerful” (Clegg Reference Clegg1989, 208–209). For this, we need the trans-actional perspective on network that would conceptualize ties (and also nodes) as “unfolding, ongoing processes” (Emirbayer Reference Emirbayer1997, 289). What the proper extant methodology is for such an understanding is an open question that cannot be answered here, only briefly highlighted.
Some authors (e.g., Harrison White) within SNA started as inter-actionalists and subsequently took a clearly trans-actional route (see Mische Reference Mische, Scott and Carrington2011 for an overview). They can be perceived as “bridges” between the two approaches to networks. In addition, the increasingly influential exponents of “interpretive approach” (Bevir and Rhodes Reference Bevir and Rhodes2006) that presume the provision of “narrative explanations” (analogous to “thick descriptions” in a Geertzian manner) to be an adequate approach in political science have several affinities with Continental classic scholars. Although rarely if ever using the SNA-type vocabulary (e.g., nodes and ties), the SNA mainstream could learn from them how to use the trans-actional notion of power for constructing “governance stories” as both explanations and interpretations of political outcomes. Their methodological precautions are based on the ontology of power akin to Foucault, Bourdieu, and Elias, who speak about network character of power or field of power rather than power as something “flowing” in the network (or field). Therefore, it is crucial for all of the approaches discussed herein that the analysis cannot concentrate only on the synchronic power configurations or even comparisons of different such configurations but also must engage in diachronic “process tracing”—hence, the overwhelmingly narrative form of analysis prevalent in the Continental approaches. However, this does not mean that once one adopts Continental/trans-actional ontology, there are only qualitative methods available for conducting research. Trans-actional ontology does not preclude “freezing” the process (methodologically) with synchronic (or comparative synchronic) SNA techniques for obtaining insight into the patterns of relations in which the actors are enmeshed. However, if further questions about the very constitution of these relations and actors are addressed, then additional historical or ethnographic research or in-depth interviews or discourse analyses may be useful for studying the be(com)ing of subjects, meanings, and networks (cf. Mische Reference Mische, Scott and Carrington2011).
It cannot be claimed that process-oriented trans-actional Continental relationalism is more in line with reality “as it is” than the more static Anglo-American inter-actional approach. Therefore, in their research, political scientists should not be exclusively committed to one or the other but rather be aware of the “ontological commitments” involved in incorporating the elements of either or both (Jackson and Nexon Reference Jackson and Nexon1999, 292).
CONCLUSION
This article distinguishes between two understandings of the “relational” approach to power: the Anglo-American and the Continental perspectives. I do not argue for the superiority of either perspective. In fact, I contend that both approaches benefit mutually from engagement with the other. The deeper reason for this may be evident in the simple precaution that “[s]ocial scientists should never be so bold as to claim they know what reality is” (Jackson and Nexon Reference Jackson and Nexon1999, 292). It cannot be claimed that process-oriented trans-actional Continental relationalism is more in line with reality “as it is” than the more static Anglo-American inter-actional approach. Therefore, in their research, political scientists should not be exclusively committed to one or the other but rather be aware of the “ontological commitments” involved in incorporating the elements of either or both (Jackson and Nexon Reference Jackson and Nexon1999, 292).
The argument in this article is similar to the one that Bachrach and Baratz made about the first “face” of power as decision making in their most influential paper titled “Two Faces of Power” (Bachrach and Baratz Reference Bachrach and Baratz1962). They did not contend that power is not about decision making but instead pointed out that, in addition, it is about non-decision making. Therefore, power has two faces. Similarly, I argue that there are two understandings of what it means that power is relational—and that, in one sense, they are two faces of a possible “relational turn” in political science.
ACKNOWLEDGMENTS
I thank many people for reading and commenting on earlier drafts of this article (including regular participants in the doctoral seminars in the Institute of Political Science and Governance at Tallinn University and those at the Institute of Advanced Social Research at the University of Tampere). I especially thank Ali Qadir, Antti Gronow, Esa Reunanen, Risto Heiskala, Pertti Alasuutari, Martin Mölder, Jari Aro, Andreas Ventsel, Nelli Piattoeva, Jorma Sipilä, Hannele Mäkelä, Marjaana Rautalin and the three anonymous reviewers of PS: Political Science and Politics.
