Toward Political Explanation of Change in Corporate Responsibility: Political Scholarship on CSR and the Case of Palm Oil Biofuels

Corporate social responsibility (CSR) has been recently conceptualized and studied as a political phenomenon. Most debates in this scholarship have thus far focused on normative issues. Less attention has been paid to the explanatory potential of CSR research grounded in political theory and philosophy. In this article, we conduct a pragmatist reading of political scholarship on CSR and seek to deploy existing knowledge for research pursuing political explanation. We argue that the political ontologies that underlie scholarship on CSR can be used to transform normative and descriptive research also for explanatory uses. We show how ontologies vary in terms of potential research objects and scopes of political explanation, and argue that the main types of political-ontological stances adopted in scholarship on CSR, foundational and post-foundational stances, offer explanatory analysis of different schematic guidelines. Our pragmatist reading of previous research and an empirical case illustration of political explanation of change in corporate responsibility in the biofuel industry demonstrate the opportunities, limitations, and challenges different political-ontological commitments provide for political explanation.

Over the last few decades, many of the key academic debates concerning the relationship between business and society have been held under the title of corporate social responsibility (CSR). This scholarship has recently taken a "political turn" (Scherer et al., 2016), which has brought about a variety of new approaches and research objects (Frynas & Stephens, 2015;Scherer et al., 2016;Whelan, 2012). A number of reviews of political scholarship on CSR have been published in the last few years (Frynas & Stephens, 2015;Mäkinen & Kourula, 2012;Scherer, 2018;Scherer et al., 2016;Scherer & Palazzo, 2011). The explicitly "political" scholarship on CSR is "concerned to normatively prescribe, and positively describe and explain, the political duties and activities of corporations" (Whelan, 2012, p. 711). In this scholarship, CSR has been discussed as, for example, ideologically framed business-society interactions (Djelic & Etchanchu, 2017), a new source of (global) governance (Haufler, 2001), and a new role for multinational corporations in the international system (Scherer et al., 2009;Scherer & Palazzo, 2011).
These strands of scholarship demonstrate that corporate responsibility as research object certainly warrants some form of "political" analysis. Political approaches address the societal foundations of CSR and seek answers to fundamental normative and empirical questions such as what kinds of duties the very existence of businesses involves; why some duties, norms, and performance expectations are more pressing than others; what firms must and must not do in different environments and/or in different times; and how firms and other actors influence their environments to provide their own answers to all of the questions above. Most research in the political scholarship on CSR have sought answers to these questions from a normative perspective, typically asking what the concept of CSR or norms and practices described under the concept ought to be like (Scherer, 2018). Less attention has been paid to the explanatory potential of political approaches in empirical research or to the potential of giving specifically political answers to the questions of why certain things are associated with CSR and why they occur.
Political ontology is key to understanding the different underpinnings and stances of normative political research on CSR. Normative political research has provided valuable insights for descriptive empirical research, because its ontological groundings have offered conceptual frameworks for identifying corporate duties and sources of legitimacy. In this article, we argue that political ontology provides helpful insights for pursuing explanatory analysis as well. Understanding differences in political-ontological stances is key for developing empirical explanatory research and avoiding category errors, as different stances provide empirical analysis of politics with quite different research objects and schematic guidelines. Our pragmatist reading of previous research seeks to deploy established theories and concepts to explanatory uses through the political ontologies underlying previous political scholarship on CSR. We show how these underpinnings facilitate explanatory analysis in different ways by delineating different explananda and the explanantia available for political explanation and structuring explanatory research.
Some terminology must be clarified already at this stage. By political ontology, we refer to conceptions that define what kinds of political things exist in the world, what exactly qualifies things as "political," how "political" things differ from other (such as ethical, economic, or corporate) things, and what makes a "political" description, analysis, or explanation meaningful and distinct. We show how different types of political-ontological stances vary according to the distinct things, qualifications, and schematic characteristics they provide for empirical political research. To delineate and differentiate political-ontological stances for empirical research and especially explanatory uses in the specific context of CSR, we will focus on three dimensions defined by such stances: (a) the activity of politics (i.e., what one does when one engages in politics), (b) the locus of politics (i.e., where such activities must take place to qualify as politics over CSR), and (c) the "stuff" of politics (i.e., what kinds of issues, measures, or objects must such activities deal with to qualify as both politics and CSR). By explanantia, we refer to politics as the domain from which explanation is sought-that is, the domain that needs to be addressed as and/or operationalized into (sets of) independent variables for an explanation to qualify as political (and not some other explanation). By explananda, we refer to all events and phenomena that reside within the scope of political analysis and can be explained with the stuff, locus, or activity of politics within a given political-ontological stance.
Our analysis discusses four families of political scholarship which each share similar political-ontological stances. Liberal political ontology revolves primarily around issues of public regulation (stuff) and polities that touch upon the lives of all members of society (locus), and republican ontology addresses issues of legitimation (stuff) and the civic domain (locus) parallel to them. Both families are varied in their conceptions of politics as activity. These two families represent foundational political ontology, which attributes politics to predefined loci, stuff, and/or activities for normative reasons. The two other families represent post-foundational political ontology, which do not predefine or attribute politics to specific stuff, loci, or activities, but only outline the conditions in and qualifications through which arenas, measures, and activities can represent politics. The Neo-Gramscian family sees politics where hegemonic struggles suppress antagonisms to enable ruling, whereas the agonistic family sees politics where antagonisms between social groups are developed into agonisms between adversaries who may oppose one another but who regard each other as holding legitimate views in debate and decision making.
We will describe the types and nature of these stances and their family resemblances and consequences for political explanation in more detail below. We also discuss the opportunities, limitations, and challenges of political explanation grounded in different political ontologies with a sample case illustration of change in corporate responsibility. Our case illustration is focused on the palm oil-based biodiesel production by the Finnish firm Neste Oil, which came to be constructed in politics first as legitimate and responsible and later as illegitimate and irresponsible.

Political Ontology and Empirical Research
Political scholarship on CSR is distinct from other strands of CSR scholarship for being grounded in political philosophy and using explicitly "political" concepts to address CSR. Whelan (2012) delineates "political" scholarship on CSR by making a distinction between "political" approaches and stakeholder theory, instrumental, and business ethics approaches. Néron (2010Néron ( , 2013 has classified the contexts in which the term "political" has appeared in CSR research. He finds that "political" approaches have addressed corporations as distributive (societal) agents, political communities and active participants in the political process, and corporate practices and policies as citizenship issues. Frynas and Stephens (2015) have classified various "schools" of political scholarship on CSR with a different type of use of the term "political." What they address as "political" CSR activities is activities where CSR has an intended or unintended political impact, or where intended or unintended political impacts on CSR exist (i.e., impacts related to the functioning of the state as a sphere of activity that is distinctive from business activity). (Frynas & Stephens, 2015, p. 485) In the aforementioned studies, the qualification "political" indicates distinct phenomena and empirical research objects. The studies also provide some kind of conceptual differentiation and specification for "political" approaches to research. These characteristics indicate that the research has been grounded in political ontology. In organization and management studies, ontology has in recent years enjoyed a revival of interest, and various approaches to ontology have been discussed (see Al-Amoudi & O'Mahoney, 2015). In general, ontology defines what things exist and what gives them their meaningful structure. Political ontology deals with what kinds of political things exist in the world, what exactly qualifies things as "political," and what makes a "political" description, analysis, or explanation meaningful. It delineates the scope of a political approach by presenting autonomous qualities of the "political" that are independent of and irreducible to other criteria and domains (morality, ethics, the economic, business, etc.; Marchart, 2007, p. 48).
What makes political ontology distinct from social and organizational ontology is that it deals with the radical contingency and enabling of social orders. Political ontology addresses simultaneously the plane on which the order of an entire society is decided upon and the prima causa, or the possibility, enabling factors and enactment of those orders (Sartori, 1973). Presenting a "political" insight in research inevitably attributes politics to some limited domains, activities, or groups of people, and may thus influence a social order by making (in)visible, highlighting and (de)legitimizing some forms of politics instead of others (Boulay, 1977). In this sense, presenting a political ontology always potentially takes part in politics, which makes political ontology "essentially contested" and similar to theories of power (Lukes, 2005). Thanks to this feature, much attention has been paid in political research to the relations between political ontology and actual (ontic) politics, and political ontologization, or the ways in which political ontology is constructed and presented in different contexts (e.g., Marres, 2013).
In terms of empirical research, political ontologies vary according to the issues, arenas, and activities to which they attribute politics. Political ontology usually deals with one or more of four interconnected times of politics that delineate the scope of empirical research: policy, polity, politicking, and politicization (Palonen, 2003). Policy refers to the regulative aspect of politics. It contains a direction of activities (project, plan, program, doctrine, law, set of principles, etc.), an orientation toward the future that is considered to be a priority over the present state of affairs, and the related actions. Polity refers to the established space, arena, or sphere of politics. It can be regarded as a temporalized space that has been commonly accepted as political and that demarcates activity from that which is not accepted as political. Politicking and politicization refer to two aspects of politics as activity. Politicking asks what should be done and how to do it, and relies on the available power shares to increase the relative advantages in their distribution. Politicization refers to the opening up of things to politics or the act of naming something as political, including the controversies surrounding the acceptance of this naming.
The schematic nature of political research depends on the type of politicalontological stance. Marchart (2007) has discussed three types of politicalontological stances: foundational, anti-foundational, and post-foundational. Foundational stances attribute "political" matters to a priori specified issues, arenas, and/or activities. They seek to address polities and activities that matter for normative reasons, for example, legislatures that enable to create wellordered societies or parliaments that promote democratic decision making. Foundational stances predefine at least some of the times of politics that ought to be followed in empirical analysis. Anti-foundational stances take the opposite view. Put bluntly, they refuse to provide definitions for the "political" or attribute it to anywhere or anything a priori, and maintain that it is not up to the researcher to say what is or is not political. Hence, they adopt a nominalist perspective that avoids taking part in politics.
Post-foundational stances differ from both approaches. They do not attribute politics to any given polity or specified activity, but only provide a set of conditions that must be met for an arena, activity, or issue to become "political." According to post-foundational stances, politics may take numerous forms and occur in multiple arenas depending on what kinds of social and political orders are constituted at different times. For example, post-foundationalists may argue that the emergence of a social order may occur through activities that no political ontology has conceptualized as politics beforehand. However, unlike the nominalist anti-foundational stances, post-foundational stances guide empirical analysis by providing some qualifications for politics: They only refuse to assume that politics takes place in a priori defined arenas or that only certain predefined activities can represent politics. Post-foundational stances typically see politics as the activities that lay the foundations for society and hence embrace the radical contingency of society. What is meant by these activities and foundations depends on the more specific stance.
Some issues have to be clarified to avoid confusion here. First, the three stances described above are specific to political ontology and political philosophy, and should not be confused with the anti-, post-, or foundationalist stances in epistemology or philosophy more generally. They denote ontological stances that tell what a "political" approach entails and deals with, not epistemological stances telling how knowledge on politics can be gathered. Second, grounding research in a particular stance does not necessarily mean that politics is regarded as something good or that politics ought to have a major role in society. Marchart (2007), for example, argues that even though all democratic theories must have some post-foundational qualities, as they need to maintain society open to different foundations, not all research grounded in post-foundational ontology necessarily deals with democratic politics. For instance, Gramscian understandings of politics as hegemonic struggles may be used to describe totalitarian orders just as well as democratic ones.
The available scopes, types, and forms of explanation vary according to the specific political-ontological grounding of research. The type of ontological stance plays some role here. The foundational and post-foundational ontological stances delineate the focal (foundational) or the scope of potentially focal (post-foundational) sources and objects of political explanation, and provide different schemas for empirical research. Research grounded in foundational ontology delineates the scope of relevant polities and activities, and possibly relevant types of policies, and hence guides empirical analysis with firm starting points. Post-foundational ontologies do not take the researcher to any predefined arenas or issues but only provide the criteria (which depend on the more specific ontology) for identifying the political aspects in any social setting or activity. Anti-foundational ontologies guide empirical analysis by urging to focus on the uses and consequences of the term "political" wherever it appears and whatever it may contain. Due to the enormous breadth of this approach, we will focus only on foundational and post-foundational stances in the rest of our article.
A specific political ontology provides the sets of relevant explananda and explanantia that a political explanation can involve (Glynos & Howarth, 2007). In research adopting foundational stances, policies are typically explained with politicking and/or politicization in and around given polities, while research based on post-foundational approaches often seeks explanations from politics for the structure and functioning of polities as well. The available modes of explanation are dependent on both research interests and methods used (Glynos & Howarth, 2007; see also Scherer, 2005), which is why we will discuss explanation in more detail in connection with the more specific approaches to CSR. The type of the political-ontological stance dictates neither research interests nor methods, but nonetheless conditions and limits research. For example, the aforementioned Gramscian approaches typically focus on making visible numerous forms of hegemonic struggle and domination with an emancipatory knowledge interest (Femia, 1981). Given the wide array of knowledge interests and mechanisms of explanation that fall under this scope, various modes of explanation can be deployed, ranging from causal statistical analyses of general phenomena like social reproduction to hermeneutic explanations of dynamics of domination in particular events, times, and places.

Identifying the Political Ontologies Underlying Research
Next, our attention turns to the more specific issues related to the identification of political ontologies underlying the political scholarship on CSR. We focus on the scholarship discussed in four influential reviews in this field: Frynas and Stephens (2015), Mäkinen and Kourula (2012), Palazzo (2011), andScherer et al. (2016). Identifying the political ontologies underlying this body of literature for the purpose of explanation entails an immediate challenge: Very few pieces of research in the field have aimed at explanation. Most research grounded in political ontology has engaged in normative debates on management or politics, or sought to provide conceptual frameworks only for empirical description. For this reason, identifying the political-ontological groundings of previous research and deploying them to new uses require a reading from a position that allows high degrees of (re-) interpretation and further elaboration.
Our vantage point is in the "theoretical eclecticism" required for turning normative political research to empirical uses (see Karppinen et al., 2008). We characterize our reading of previous research as pragmatist and instrumentalist in Dewey's (2015Dewey's ( , 1903 sense, that is, considering that theories should foremost be seen as useful tools in relation to practical problems (see also Kilduff et al., 2011). In this case, the practical problem is related to the conduct of explanatory research in an established field with extensive conceptual frameworks but limited insights on political explanation. Usefulness is defined by the opportunities and limitations for explanation and the schematic guidelines for empirical research provided by a political-ontological grounding.
When applying a theory from this vantage point, a minimal commitment to the theory is needed, "in the sense that theory acceptance does not involve the necessity of believing that the theory is true or that metaphysical unobservables are real" (Kilduff et al., 2011, p. 303). Against this backdrop, political-ontological commitments of research are not evaluated based on "how well they depict 'actual' causal processes" but instead in terms of how they might "facilitate empirical prediction" (p. 302). To be clear, we are not partial to any one of the ontological stances we discuss here. Given our minimal commitment to theories, our approach shares characteristics with Rorty's (1989) figure of the ironist. Even though we selectively subscribe to different ontologies or, in Rorty's terms, "final vocabularies" in different research endeavors, we have doubts about the vocabularies we currently use and realize that they might not be closer to reality than others. The epistemological position of our reading is thus similar to that of "constructive philosophy" as it is premised upon "the linguistic and the pragmatic turn" (Scherer et al., 2016, p. 43).
From this vantage point, the key challenge concerns not the purpose of previous research but the incompleteness of its political-ontological groundings. Whelan (2012) has argued that political scholarship on CSR has rarely been grounded explicitly or consistently in political theory. While this may indeed sometimes be the case, even research that has been explicitly and consistently grounded in political ontology does not necessarily address all times of politics in equal detail. The Rawlsian approaches to CSR (e.g., Mäkinen & Kourula, 2012), for instance, have often located politics of CSR to well-specified polities (i.e., the legislature and its interactions with other actors) and types of policy (i.e., binding regulations), but provided little specification for the activity of politics. Deliberative approaches to CSR (e.g., Scherer & Palazzo, 2007), in contrast, are typically quite specific in their definitions of politics as activity (i.e., deliberation and/or specific type of discourse) and types of policy (i.e., regulation and legitimation), but much less specific in terms of relevant polities. Such gaps are not necessarily due to inconsistency but deliberate choices regarding political-ontological groundings or research questions.
The less complete the political-ontological grounding is (i.e., the fewer the times of politics are addressed explicitly), the more challenging it becomes to delineate the political ontology underlying the study. If one limits the analysis only to research in which all times of politics have been explicitly addressed and delineated, the body of literature included in analysis becomes virtually nonexistent. But if one focuses on research in which only one time of politics is addressed explicitly, the implicit elements of the underlying political ontology have to be somehow construed.
To tackle this challenge, we look for family resemblances between different pieces of research instead of addressing the foundations of each piece of scholarship separately. To facilitate this, we reduce the number of the "times" of politics and make a distinction between only three ontological dimensions: the "stuff" of politics, the "activity" of politics, and the "locus" of politics. The "stuff" of politics includes all matters related to policy and all things politicized, that is, all issues and phenomena that are explicitly or implicitly qualified as "political" and associated with the notion of "CSR" or "corporate responsibility." The "activity" of politics includes all matters related to politicking and politicization as activities, that is, all mobilization, persuasion, communication, and other acts that can ultimately lead to decision making in an arena. The "locus" of politics refers to these arenas by themselves or the polities in which politicization and policy making can take place legitimately (in empirical or normative terms depending on the type of political-ontological stance).
In our following analysis, we have interpreted the political-ontological foundations of political scholarship on CSR as represented by research discussed in the four reviews. We first identified the "stuff" (foundational stances) or qualifications for "activity" (post-foundational stances) of politics, that is, the issues or activities different pieces of research have explicitly qualified as "political" and associated them with "CSR" or "corporate responsibility." We then categorized different characteristics according to family resemblance in relation to this first characteristic.

Political Ontology in Scholarship on CSR
We identify four families of political ontology in political scholarship on CSR. The two foundational families, the liberal and republican families, are labeled primarily on the basis of the "stuff" of politics they dealt with. This means that their understanding of the political aspects of CSR represents liberal or republican thought, while being more varied in terms of activity and/ or locus of politics. The foundational stances could be identified based on relatively clear specifications of the aspects of CSR and/or of the locus of politics. Examples range from very general delineations of political aspects of CSR as government-set regulations (e.g., Detomasi, 2007) or duties to engage in specific types of activities to gain legitimacy (e.g., Scherer & Palazzo, 2007, to very specific definitions such as "engagement with collective issues and public goods, that facilitates positive and impedes negative business contributions to society" (Scherer, 2018, p. 394) or "impacts related to the functioning of the state as a sphere of activity that is distinctive from business activity" (Frynas & Stephens, 2015, p. 485).
The two post-foundational families, the neo-Gramscian and agonistic families, were identified and are named primarily on the basis of the activity of politics they described, because their understanding of stuff and loci of politics is contingent upon the activity of politics. Post-foundational stances could be identified most importantly for treating CSR as an "empty signifier" that may appear in multiple terrains as a frame for "doing" politics (e.g., Banerjee, 2014;Djelic & Etchanchu, 2017;Fougère & Solitander, 2009;Levy & Kaplan, 2008;Newell, 2005). Common to such works was also to highlight the plurality and contested nature of the research objects that it can entail (e.g., Okoye, 2009).
Below we will present the four families in more detail. In addition, we will discuss the explananda and explanantia offered by different families, and how each family might conceptualize and be deployed to explain cases of change in corporate responsibility for the purposes of our case illustration.

The Liberal Family
We call the first foundational family liberal. Most works in this family rely directly on different strands of liberal political philosophy, most notably the works of John Rawls. Conceptions of CSR based on varieties of political liberalism have been extensively discussed by Mäkinen and Kourula (2012) and Mäkinen and Kasanen (2016). Liberal political philosophy in general locates politics in legislative institutions constituted by free citizens (Singer, 2015). In liberal thought, politics by definition has to potentially touch upon the lives of every member of society, and only free, private, individual citizens with real personhood can legitimately decide upon the constitutional orders that warrant decision making over such measures. Liberal approaches are foundational in the sense that they conceptualize politics through normative principles of a just (Hsieh, 2009) or well-ordered (Wettstein, 2009) society whose institutions ought to, among other things, be impartial, maximize individual liberty, and prevent politics from becoming an economic commodity (Freeman, 2011). Common to all "members" of the family is to regard as political the laws and regulations (stuff) produced in constitutionally mandated political systems (locus), such as representative parliamentary institutions. Beyond this commonality, much variation exists within the family.
In the liberal family, businesses are primarily seen as economic actors and/ or as private contractual ventures with only legal personhood and who operate in the societal "basic structure" laid out in politics. Most, but not all, liberal conceptions consider direct firm engagement in politics conceptually impossible because only individuals can or have the right to engage in politics. For this reason, the liberal family does not share the view that the erosion of state power over global business (see Scherer & Palazzo, 2007) or the ostensible blurring of boundaries between business and politics (Cutler et al., 1999) implies that firms should engage with politics (see Mäkinen & Kasanen, 2016). Instead, the family could contend that if boundaries that guarantee individual liberty and a just basic structure of society have become unclear, then a liberal political order with clear boundaries must be re-established, albeit not necessarily at the level of existing states (Mäkinen & Kourula, 2012, pp. 665-666).
There are two "family feuds" within the liberal family regarding the boundaries and scope of political ontology. First, there is much disagreement on what kind of assembly of political and business institutions qualifies as liberal. More precisely, the liberal family assigns firms the responsibility to respect the societal "basic structure" like any private actor (Esquith, 2011), but much variation exists on the nature of an appropriate basic structure. Mäkinen and Kourula (2012) and Mäkinen and Kasanen (2016) have used the notion of "political systems of the division of moral labor" to contrast different strands of liberal thought in the context of CSR. Some scholars limit the CSR of firms to not causing harm (Hsieh, 2009) or to compensation for ill-gotten gains from activities that go against the prevailing moral, legal, and political norms (Esquith, 2011). In this view, CSR has little to do with politics beyond compliance with regulations imposed by formal political actors.
Other scholars see that private actors, including companies, have a general duty to support the functioning of political institutions (Mäkinen & Kasanen, 2016) or just social institutions more generally (Wettstein, 2009) via public action (but not necessarily politics). In this broader political-ontological view, politics of CSR is located not only within legislatures but also between legislatures and private actors, and the "stuff" of politics may involve policies and practices of firms and not only legislators and regulators.
The second "family feud" concerns the activity of politics. Politics as activity has rarely been clearly delineated in the liberal family. Most research has focused on the appropriate roles of political and business actors, not the activities through which these roles are debated and enacted in practice. However, the literature cited in the liberal political scholarship on CSR suggests that the conception might be highly varied. The prevalent conceptions range from what can be called "high" political liberalism that embraces more deliberative, communicative, and argumentative understandings of politics to classical and economic liberalism that embraces more interest-aggregating, representative, and economic (calculative) understandings of politics (Freeman, 2011).
In political-ontological terms, liberal explanations revolve around (but are not limited to) issues of public regulation, because it is politics and public regulation, not business and private regulation, that is supposed to safeguard a just or well-ordered society for free individuals. For this reason, political explanation based on liberal political ontology must be primarily sought from a polity that touches upon the lives of all members of society. Some liberal scholars may add that duties of businesses may extend beyond the regulations crafted in such polity especially if the quality and functioning of the polity is compromised (Mäkinen & Kasanen, 2016). The conception of politics as activity and, hence, the explanantia from which explanation is sought, vary within the family. The liberal explanation is schematically linear: If there is a change in CSR in terms of stuff of politics, it must come from political institutions and be caused by some kind of activity occurring within them-be it coalition-building between interest-aggregating parties within a parliament or deliberation among parliamentary representatives and external stakeholders.

The Republican Family
We call the second family republican due to its common features of addressing public regulations as well as civic legitimacy in connection with politics, and locating politics into constitutional polities as well as the civic domain. The republican family resembles the views of the advocates of broader political-ontological approaches within the liberal family, but has a broader understanding of the stuff of politics, as it regards legitimation not as secondary to but parallel to regulation, and discusses public as well as private regulation as political. In this family, all civil society actors, including firms, are seen to have a duty to act politically to ensure public legitimacy and promote civic virtue that enables civic self-determination and makes possible (although not guarantees) the proper functioning of public polities. This suggests that firms ought to communicate and deliberate with the broader society to find politically legitimate solutions to social problems, be they enforced through public regulation or private self-regulation.
The republican family draws especially on works of Scherer, Palazzo, and others of the "Habermasian camp" discussed by Frynas and Stephens (2015), whose research has been perhaps the most influential and the most debated in business and society literature, and literature on "extended corporate citizenship" (Matten & Crane, 2005). We acknowledge that not all works within this family explicitly subscribe to republican political philosophy and that the group is quite heterogeneous in its understanding of politics as activity. We use the label republican here to highlight the family resemblances only in terms of the locus and stuff of politics. Indeed, republican political philosophy locates politics both in the constitutional arenas that shape an entire society and in civil society, which influences these arenas (e.g., Dryzek, 1996). Following Habermas, these can also be called the "strong" and "weak" publics (Baynes, 2002). The strong publics refer to the domain of "politics proper" (Sartori, 1973), the political system whose outcomes are applicable to all members of society. But politics can also be found in weak publics, or the institutions and arenas warranted by or independent of the constitutional political system. Weak publics are necessary for facilitating the civic virtue of deliberation and for ensuring civic self-determination, which in turn is necessary for strong publics to exist and function.
Most works that belong to this family understand the activity of politics broadly as communication, albeit in quite different ways and with different weights on issues such as representation and deliberation (e.g., compare the two interpretations of Habermas in Scherer & Palazzo, 2007). For example, Scherer et al. (2009) specify the activity of politics as communication and discourse as the process of forming and transforming preferences (p. 340). Politics occurs when "people collectively regulate their social conditions and decide on the direction they wish to take -regardless of the various motivations the actors have to enter this process" (Scherer et al., 2009, p. 339). Evidently, the actual conduct of communication and discourse is qualitatively different in a parliament with formal procedures including party and regional representation and in informal citizen groups. Policies, too, are different in terms of their form, regulatory force, and scope if they are formal public policies with supervisory bodies and sanctions or informal results of stakeholder dialogue. But these all are included in political analysis.
A political explanation drawing on republican ontology offers a broad set of explananda and an elaborate (but varied) understanding of the activity of politics as potential explanans in the context of CSR. It can address a broad array of formal and informal regulations and other communications that can be explained, because its stuff of politics is based on a broader understanding of policies than in liberal approaches. Moreover, it allows to seek explanation from formal political and civic domains with the activity of politics. However, this also makes republican explanations potentially more complex than liberal ones: They have to address various loci of politics to provide a coherent explanation.

The Neo-Gramscian Family
The political scholarship on CSR that draws on post-foundational politicalontological stances includes two families: the neo-Gramscian and agonistic families. The two families are close to each other in the sense that both draw on Gramsci's concept of hegemony as their most significant starting point. Both families also see politics as contingent upon foundations laid by temporary hegemonic social orders (e.g., Levy et al., 2016;Moog et al., 2015;Tregidga et al., 2018). In both families, antagonisms that might prevent peaceful decision making and commitment toward it have to be suppressed for any kind of discourse or decision-making to qualify as politics instead of, say, mere public debate or ruling. The differences between agonistic and neo-Gramscian families emerge from different interpretations of the scope of the autonomy of politics and qualifications for the activity of politics.
The neo-Gramscian family comprises of studies that are somewhat faithful to the Marxist primacy of dynamics related to the economic base. For neo-Gramscians, the distinct characteristics of politics lie in power. In such accounts, the economic base determines the formation and reproduction of hegemonic blocs, but not the outcomes of the struggles between them. The qualifications for politics as activity involves a range of characteristics related to politicking and "power politics" within and between the blocs, including alliance formation, negotiation, resistance, and compromisewhatever is needed to suppress antagonisms and establish hegemonic orders. The existence of polities in a hegemonic order is dependent on the will of hegemons. A hegemonic order does not necessarily involve a meaningful political system, but hegemonic powers may also provide very broad mandate for representative democracy.
Out of the few pieces that comprise the family, most neo-Gramscian approaches to CSR discuss CSR as expressions of hegemonic orders in the form of polities (e.g., Djelic & Etchanchu, 2017;Levy, 2008). For example, CSR has been seen to represent an undemocratic and "privatized" form of neoliberal global governance (Levy & Kaplan, 2008), and multi-stakeholder initiatives (MSIs) with a CSR framing have been seen as tools for serving the interests of a dominant coalition rather than a more general interest (Levy et al., 2010;Moog et al., 2015).
Deploying neo-Gramscian insights for political explanation is challenging given the narrow body of existing literature. Post-foundational insights make it possible to address multiple polities and policies, like the republican family, and also include polities as their explananda. Yet, few studies have sought to explain particular policies or polities associated with "CSR" through neo-Gramscian politicking. Such political explanation would start by paying close attention to hegemonic struggles -focusing either on more moderate accommodations by the hegemonic blocs or on more radical shifts in the economic base that determines the composition of those blocs -and to the role of the signifier of "CSR" in the power politics within these struggles.

The Agonistic Family
The agonistic family is typically grounded in the post-Marxist works of Mouffe (1999Mouffe ( , 2005 as well as Laclau and Mouffe (2001). In them, hegemony and counter-hegemony are seen not as inevitably dependent on economic base but as radically contingent and based on discourse articulation. The agonistic family sees politics as activity that requires mutual recognition and legitimacy. Unless antagonisms between social groups are developed into agonisms between "adversaries" who may oppose one another but who regard each other as holding legitimate views, there will be no commitment to common discourse. Without these agonistic qualities, politics cannot enjoy radical autonomy or, to set social orders, political systems and policies that its members will follow.
Many, but not all (e.g., Ehrnström-Fuentes, 2016), works belonging to the agonistic family draw on Mouffe. These include Whelan's (2013) "dissensual CSR," studies focused on agonistic co-optation and/or resistance (Burchell & Cook, 2013a, 2013b, and a number of discourse theoretical approaches to CSR (Laasonen et al., 2012;Tregidga et al., 2018). This research has focused on the role of powerful "empty signifiers" or "nodal points" in establishing and changing a hegemonic order by constructing a "chain of equivalence" that unites different societal demands and power resources. "CSR" (e.g., Sorsa, 2013) and "sustainability" (e.g., Brown, 2015) have both been seen as such empty signifiers that stand in for all societal aspirations for a business activity, as they make a broad variety of (environmental, ethical, social, etc.) concerns "equivalent" in such a way that powerful parties agree that more "CSR" or "sustainability" is needed. A dislocation of a hegemonic discourse and its chain of equivalence may be possible when a counter-hegemonic movement targets the nodal point of that discourse, treating it intentionally as a floating signifier and giving it a particular meaning, perhaps through another powerful (empty) signifier that might set up a chain of equivalence in its own right.
Even though agonistic approaches do not attribute politics to any specific loci a priori, a significant share of works in the agonistic family have focused on partnerships and/or MSIs. Especially, Cook (2013a, 2013b) have discussed the so-called "co-optation" dynamics of partnerships as agonistic relations. For example, while some civil society groups may follow a relatively cooperative, passive strategy toward their relationship with business, other groups may frame these relationships within a context of ongoing confrontational external activities and push companies toward engagement that they may otherwise have avoided (Burchell & Cook, 2013b). Typically, the policies emerging from such agonistic politics involve self-regulations, statements or principles, or changes in the form of business-community dialogues.
As is the case with neo-Gramscian explanations, an agonistic explanation would first need to identify a hegemonic struggle. However, the scope of explanation with the activity of politics is much broader in research that draws on agonistic political ontology. For example, a change in the stuff or locus of politics, such as legal banning of a product or inclusion of nongovernmental organizations (NGOs) to corporate decision making, might be caused by a counter-hegemonic movement organized by legitimate actors, not only through change in the economic base or power struggles between hegemonic blocs. Such movement may partially succeed to challenge the hegemonic bloc discursively and force it to go beyond a symbolic accommodation to the critique. In such cases, the explanation focuses on the success factors of the counter-hegemonic movement in terms of discourse articulation and formation of chains of equivalence.

Toward Political Explanation: The Case Illustration
In the following case illustration, we show how different political-ontological groundings can be deployed to explain change in corporate responsibility. Our case study should not be regarded as a typical case of politics of CSR, but only as an illustrative example with specific characteristics that allow valid deployment of all the families discussed above. Unless a legislature produces regulations that indicate binding duties, liberal groundings are unlikely to find relevant stuff of politics in the first place. If there is no deliberation over preferences in civil society or legislature, no such thing as politics occurs according to republican groundings. Hegemonic struggles must be witnessed to be able to find neo-Gramscian or agonistic politics. The case includes all these characteristics. The case illustrates how different political-ontological groundings will shift attention to explaining somewhat different changes or provide somewhat different puzzles in respect to explaining similar changes.
The case illustration is focused on the palm oil-based biodiesel production by the Finnish firm Neste Oil, which came to be constructed in politics first as responsible and later as irresponsible. Supported by the Finnish energy subsidy scheme, Neste (called Neste Oil at the time) started using its NexBTL biofuel technology with the raw material feedstock of the highest possible yield-palm oil. By 2007, Neste Oil was one of the largest buyers of palm oil globally, and the first oil company to have joined the MSI Roundtable on Sustainable Palm Oil (RSPO) in 2006. It aimed to buy all its palm oil from plantations certified as "sustainable." To a large extent, Neste Oil's palm oil-based biofuel was validated as renewable and sustainable at the time. It was further incentivized by two influential policies meant to reduce greenhouse gas (GHG) emissions in Europe: (a) the EU's 2009 Renewable Energy Directive (RED) and its target of having "10% of the transport fuel of every EU country [to] come from renewable sources such as biofuels" (European Commission, 2018), and (b) the 2010 Finnish Act on the Biofuels Distribution Infrastructure, aiming for biofuels to reach a 20% share of the Finnish transport sector by 2020 (Finnish Ministry of Transport and Communication, 2013).
In parallel with these validations, a number of NGOs fighting against the deforestation incurred by oil palm expansion started to target biofuels, and particularly Neste Oil, as a key culprit of deforestation in the areas where forests and peatlands are converted into oil palm plantations (e.g., Friends of the Earth International [FoEI], 2009;Greenpeace, 2013). The NGOs listed the explosion of GHG emissions, biodiversity loss, habitat loss for endangered species, losses of livelihoods for local communities, and land conflicts as adverse consequences of palm oil biofuels. These perspectives gave rise to a significant public debate over whether it is "corporately responsible" to produce this type of biofuel, a debate involving institutions such as the EU Parliament and Commission, RSPO and its member organizations (including large buyers and producers of palm oil and large international NGOs like WWF and Oxfam), as well as civil society organizations refusing to be members of RSPO such as FoEI, the Rainforest Action Network (RAN), and Greenpeace.
The provisional outcome of this debate at which we end our illustration was the adoption of a report by an EU Parliament Committee on March 9, 2017, "urging the European Commission to phase out the use of palm oil as a component of biodiesel by 2020 'at the latest'" (Michalopoulos, 2017), followed by an overwhelming EU Parliament vote (640-18) in early April recommending the ban of biofuels made from vegetable oils including palm oil by 2020 (Neslen, 2017). At the time of writing, the EU Commission has not yet implemented this recommendation in EU legislation, and the largest palm oil-producing countries Indonesia and Malaysia as well as some large corporations with stakes in biofuels are working to counter the recommended ban. Yet, at this stage, it appears that what was once broadly validated as a socially and environmentally responsible good has now become broadly understood as socially and environmentally irresponsible.

The Liberal Explanation
A political explanation of change drawing on liberal political ontology will focus primarily on how the regulatory arrangements that initially incentivized, supported, and validated palm oil biofuels came to change over time.
Assuming that the EU apparatus counts as a liberal polity, the relevant locus here is the European Parliament. Depending on the more specific particular conception of politics, explanation is sought from the aggregation of interests, deliberation, or some other activity within parliamentary work. One line of inquiry that draws on the "high" liberal tradition might explore parliamentary committee perceptions, especially the question whether it started to regard industry self-regulation as such insufficient (i.e., an appropriate division of moral labor was perceived to be lacking). Otherwise, the explanation boils down to whether a majority that is willing to deem palm oil biofuel as "renewable" or not is formed, because the legislature had already produced a duty for states to adopt renewable energy sources. Given the overwhelming consensus of the European Parliament vote in 2017, it is hard to maintain that there was a major shift in the interests since 2009, as no commensurate shifts had occurred among the electorate. Hence, the explanation must focus on the change in the perceptions of benefits and harmfulness of palm oil production in the parliament.
Two processes determined the change in perception: (a) the RED's operationalizations of what qualifies as a "renewable" biofuel, and (b) the EU Parliament Committee that worked on this particular issue. The RED first qualified and later disqualified palm oil biofuels as renewable. At the core of this process was the concept of "Indirect Land Use Change" (ILUC). In the RED, ILUC entailed calculations of what can be considered a "renewable" and sustainable biofuel. Through it, the RED refused that lower GHG emissions from EU countries can be done at the cost of much higher emissions elsewhere. The ILUC concept draws on the observation that "when biofuels are produced on existing agricultural land, the demand for food and feed crops remains, and may lead to someone producing more food and feed somewhere else" (European Commission, 2012). As biofuels are meant to lead to lower emissions, the fact that they cause heightened emissions in other parts of the world-according to some studies, 17 to 420 times more emissions than what is saved by using the biofuels in traffic in Europe (e.g., Fargione et al., 2008)-cannot be overlooked, as civil society actors reminded the parliamentary committee. This concept of ILUC thus led to the disqualification of palm oil biofuels as "renewable" in the parliamentary committee, hence justifying the parliament's recommendation of a ban.

The Republican Explanation
A republican political explanation will expand the liberal focus on the regulatory shift as explanandum to self-regulation and informal legitimacy of corporate activities, and focus on deliberation in both the formal political institutions and within civil society. It also expands the explanans to preferences in relation to regulatory activities in both domains. A republican explanation would also pay attention to discourse and deliberation within the European Parliament in a similar manner to the "high" liberal explanations. In light of these starting points, the main puzzle that requires explanation is the strong divergence of conclusions of deliberation in different publics and their relation to the nearly unanimous strong publics.
If one had paid attention to the outcomes of certification and deliberation within the RSPO, in 2013 one would have concluded that the palm oil biofuel of Neste Oil-the first recipient of RSPO Supply Chain Certificatewas confirmed and strongly legitimized in civil society deliberation. Neste Oil was also appraised as responsible by a number of NGOs (see, for example, Sullström, 2016). However, a bit before the release of the supply chain certification, the RED had also clarified its ILUC criteria, partly as a result of pressure from adversarial civil society organizations such as FoEI, Greenpeace, and RAN. It was eventually these, and not the principles developed in connection with the RSPO certificate, that the EU parliamentary committee adopted in its positions. So why did the strong legitimization provided by the RSPO not translate into similar outcomes in the European Parliament?
One line of explanation could focus on the communicative practices adopted in different arenas. For example, one could argue that the RSPO was insufficiently deliberative and representative (see Schouten et al., 2012), and hence civil society communication and discourse did not legitimately qualify as politics in the eyes of the parliament. In contrast, the RED, with its openness to dissenting voices, was able to facilitate the civic virtue of political deliberation better and thus showed a better example to the parliament. A related line of explanation could focus on the hierarchy between relevant weak and strong publics. With this line of argument, the explanation could focus on the perceptions over appropriate division of regulative responsibilities in the EU policy process and also expand to relevant international treaties and conventions, such as the United Nations Framework Convention on Climate Change (UNFCCC), which provides principles for the regulation of emissions globally.

The Neo-Gramscian Explanation
The two families of post-foundational political ontology can be deployed to explain politically a variety of changes in the Neste Oil case, including change in regulation, shift of the political arenas, or shift in power relations within each arena. The neo-Gramscian explanations may start by looking into the RSPO or EU directives as hegemonic projects attempting to lay the polities and policies for the corporate responsibility of palm oil buyers. Each of the two can be interpreted as polity to represent all the stakeholders and work through consensus, and as attempts to hegemonize the meaning of "sustainable palm oil" and the policies through which it can be enacted (certificates or laws). From such starting points, different interpretations of the change in corporate responsibility would be made by the two families.
The neo-Gramscian explanation could focus on the exercise of power by industry blocs. The RSPO could be readily interpreted as an instrument for sustaining the interests of the "hegemonic bloc" (i.e., large industry producers and buyers of the palm oil value chain), which is meant not to leave much role for politics within the initiative. The neo-Gramscian explanation could also focus on the status of Neste Oil in the hegemonic blocs in the context of the European Parliament. Here, it is apparent that Neste Oil was on the wrong side of the "fuel" versus "food" division of industry interests, which ultimately led to its demise. Neste Oil belonged exclusively to the "fuel bloc" unlike all the other, larger and more diversified palm oil buyers (Unilever, Nestlé, etc.), who belonged and successfully mobilized through the "food bloc."

The Agonistic Explanation
The agonistic explanations could expand neo-Gramscian explanations by analyzing how the coalition of adversarial voices involving FoEI, Greenpeace, RAN, and many more local NGOs in palm oil production areas succeeded to shift the policy-making arena. These actors succeeded in promoting an adversarial counter-hegemonic discourse relying on a powerful enough nodal point ("land-use change") to achieve a discursive chain of equivalence between different struggles (about climate, biodiversity, local livelihoods, land rights, etc.) against oil palm expansion in arenas that hold ample power resources (the EU apparatus). While the adversarial actors abandoned the civil society initiatives and negatively affected their legitimacy as a political arena, they remained committed to and participated in the working of the EU apparatus through lobbying and influencing the parliamentary committee. In this sense, the vital chain of equivalence was established eventually in the encounters of hegemonic and counter-hegemonic forces within the formal political system, not the civil society initiatives.

Conclusion and Discussion
Political scholarship on CSR has revolved around large-scale normative questions such as what the division of labor between the public and private sectors should be like. What has been somewhat lacking in this scholarship is the development of research that seeks to explain phenomena associated with the notion or idea of CSR with politics. In this article, we have shown that the established political-ontological foundations of scholarship on CSR can be deployed to explanatory uses. We have shown that different political-ontological stances adopted in research provide different types of research objects and, depending on the type of the stance, different schemas for seeking explanation. Our sample illustration shows how different "political" approaches deal with different issues and activities, and shift attention to quite different sets and scopes of issues in explanation.
Of course, we have here only taken first steps in a long process toward explaining CSR-related phenomena with politics. Much more is needed for developing this field, including the identification of available modes of explanation, mechanisms involved in different modes of explanation, the factors that activate and use these mechanisms, and more generally theorybuilding based on various kinds of empirical research settings. However, thanks to the distinct characteristics of "political" research, one simply cannot avoid the question of ontological grounding when doing any of these. Different political ontologies will inevitably shift attention to different things and attribute politics to different arenas and/or activities. For this reason, we suggest that the necessary development work described above must remain highly sensitive to ontological foundations. This does not mean that development work should be limited within specific schools of thought, but only that one has to acknowledge and understand the different scopes of analysis that different ontological groundings offer. To end the article, we will briefly reflect upon the opportunities, limitations, and practical challenges that different ontological groundings provide for explanatory research.
The key opportunity that liberal ontology provides for political explanation lies in its apparent simplicity and clarity brought by the primary stuff and locus of politics. Given the increase of regulation of and government policies on CSR (Knudsen & Moon, 2017), the liberal understanding of the stuff of politics is likely to remain a popular and relevant grounding in research. At the same time, the narrow focus is the main limitation of liberal political ontology. Put bluntly, explanation grounded in liberal ontology may be able to politically explain very few things in the world, if its political institutions and activities are not liberal. More conceptual work is certainly needed to say when a polity qualifies as liberal and when not. In our case illustration, we interpreted the European Parliament to qualify as a liberal legislature. Without this interpretation-which is far from obvious-little politics could be observed to occur in the case before some liberal national parliaments started to discuss the matter. If this was the case, liberal explanation could have focused only on events that took place after major disputes and controversies. Yet, perhaps the main challenge of the current stage of scholarship is the lack of specification of politics as activity. Without further elaboration on that, too much weight in explanation will be put on who the political actors are or what they represent and too little on what they in fact do.
The key opportunity provided by republican ontology is its broader scope of the stuff and locus of politics and the broader scope of explanation this brings. Republican understandings of the stuff and loci of politics in connection with CSR allow to include in political analysis a broad variety of legitimacy-related communications and civic arenas that liberal ontology would not readily qualify as political. The republican family has also been more specific in its understanding of politics as activity than the liberal one, which makes it highly useful for empirical research. Indeed, due to the maturity of normative and descriptive research on politics grounded in republican ontology, grounding research in this ontology can provide relatively clear schemas, conceptual tools, and scopes of relevant issues for political explanation. However, a number of research gaps somewhat blur the picture. In particular, there are still relatively few studies that address politics of CSR in the "strong publics" despite increasing focus on the issue in state politics (Knudsen & Moon, 2017). Without extensive research on the nature and functioning of strong publics, it remains unclear how common republican politics (e.g., certain types of deliberation or discourse) is in them and, consequently, what the scope of republican explanation can be in the strong publics. The relations between weak and strong publics also remain somewhat underexplored (Scherer et al., 2016).
Evidently, all foundational political-ontological stances can provide broad and valid political explanations only in a world whose political institutions correspond with the ontological commitments of research. Because they focus on predefined political issues, arenas, or activities, foundational political-ontological stances may look for political explanations for phenomena from few sources. The promise of post-foundational stances is to avoid these problems. If the prevalent political systems are not founded on liberal or republican principles and maintained by hegemonic powers, then one should not simply consider them as relevant loci of politics for explanatory research. And if any one thing that post-foundational political-ontological stances qualify as political is not occurring, then one should not consider a political explanation in the first place. To counter such promises, the post-foundational stances also have their pitfalls.
The explanatory opportunities provided by agonistic ontology are similar to those of republican ontology in the specific context of CSR. The main difference is that agonistic explanations are more flexible in terms of explananda and explanantia. For example, they can explain the emergence of polities such as parliaments or MSIs with the activity of politics (instead of focusing on the activity of politics contained within those polities). They can also consider "CSR" as relevant for creating hegemonic orders, specific polities, or policies coming from specific polities. It all depends on the case. The problem here is that politics may occur in numerous forms and terrains. If one cannot predefine any locus or stuff, where should one then go and what to observe to find something that might count as politics?
Explanatory analyses grounded in neo-Gramscian ontology somewhat differ from agonistic analyses. On the one hand, the empirical scope of explanation with politics is more limited in the neo-Gramscian family, as the understanding of politics as activity (i.e., politicking between power blocs defined by the economic base) is narrower than in agonistic ontology. On the other hand, neo-Gramscian ontology enables political explanation wherever the use of power to suppress antagonisms is found-basically everywhere in modern societies. But if political explanation can be deployed to explain almost anything social, what added value does a "political" explanation bring to broader social explanation?
A grand challenge for explanatory research that draws on either of the two post-foundational families is the underdevelopment of methodology. It remains very unclear how and with what kind of data one can identify relevant loci, stuff, and activities of politics in connection with the label "CSR." Doing this would require a clear delineation of a hegemonic discourse (agonistic) or hegemonic social order (neo-Gramscian) and identification of all the terrains in which struggles can take place. Few insights on methodology for doing something like this have been presented apart from certain cases of business-driven MSIs (see Levy et al., 2010;Moog et al., 2015). Perhaps for this reason, many phenomena remain empirically unexplored, including the processes in which enemies turn to adversaries (agonistic research) or actors fall from hegemonic blocs (neo-Gramscian research) in CSR contexts.
To conclude, when deploying the current state of the art in political scholarship on CSR for political explanation, the choice between foundational and post-foundational stances currently appears to involve some trade-offs between validity and feasibility. Moreover, different ontological groundings within each type of stance can offer different scopes for political explanation and thus different degrees of relevance. Such concerns are perhaps not exclusive to matters of explanation. In analogy, the use of political strategies prescribed by normative political theory in a world whose actual politics are different may lead to very counterproductive results (Sabadoz & Singer, 2017). Put bluntly, grounding actions on liberal or republican ontology in a nonliberal or nonrepublican world may point toward irrelevant arenas for those who wish to influence specific issues. Meanwhile, depending on the data and methods used, research grounded in post-foundational political ontology may lead to discovery of highly deterministic or highly symbolic politics, which may prove just as disappointing for those who wish to change socio-material realities.