On the limits of the political: The problem of overly permissive pluralism in Mouffe's agonism

The critique of depoliticization is a central theme in agonistic political theory. As discussed in the work of Chantal Mouffe and Ernesto Laclau (2001), any principled distinction between legitimate political contestations and indisputable requirements of rationality is deemed problematic (Mouffe, 2000, pp. 135–136). This is because, agonistic theorists believe, the boundaries between the political and non-political realms are always constituted by exclusionary acts of power. Such boundaries seem natural or necessary only until conflicting parties renegotiate them, that is, through hegemonic struggles determining what issues are the intelligible subject matter of political disputes (Mouffe, 2005, p. 18). While the critiqueof depoliticizationbears valuable insights byuncovering hidden formsof contestable power relations, its categorical rejection has controversial implications. In its current form, Mouffe’s agonistic political theory does not provide adequate normative criteria to reject potentially problematic instances of politicization such as science denialism observed in the public debates about climate change, antivaccination movements and the AIDS epidemic (Anderson, 2011; Fassin, 2007; Fitzpatrick, 2004). This is because agonistic political theory does not seem to allow for a prepolitical discursive space in which articulations on facts and truth claims bear genuine objectivity. Since social reality is political all the way down, any antagonistic moment could be viewed as a legitimate conflict that needs tobe containedwithin thepolitical process. Theonly constraints onexcessive politicization are that opposingpositions should peacefully coexist within a certain discursive and institutional order that we call politics, and/or they should be subject to the requirements of “the commonsymbolic space,” that is,minimally sharednormative foundationsof liberal democracies (Mouffe, 2000). I contend that these measures offered by Mouffe do not succeed in filtering out instances of politicization based on pseudo-science for two reasons. First, her conception of the common symbolic space, that is, based on the ideals of freedom and equality, is too narrow to properly limit politicization. Much of the political positions that rely on


INTRODUCTION
The critique of depoliticization is a central theme in agonistic political theory. As discussed in the work of Chantal Mouffe and Ernesto Laclau (2001), any principled distinction between legitimate political contestations and indisputable requirements of rationality is deemed problematic (Mouffe, 2000, pp. 135-136). This is because, agonistic theorists believe, the boundaries between the political and non-political realms are always constituted by exclusionary acts of power. Such boundaries seem natural or necessary only until conflicting parties renegotiate them, that is, through hegemonic struggles determining what issues are the intelligible subject matter of political disputes (Mouffe, 2005, p. 18).
While the critique of depoliticization bears valuable insights by uncovering hidden forms of contestable power relations, its categorical rejection has controversial implications. In its current form, Mouffe's agonistic political theory does not provide adequate normative criteria to reject potentially problematic instances of politicization such as science denialism observed in the public debates about climate change, antivaccination movements and the AIDS epidemic (Anderson, 2011;Fassin, 2007;Fitzpatrick, 2004). This is because agonistic political theory does not seem to allow for a prepolitical discursive space in which articulations on facts and truth claims bear genuine objectivity. Since social reality is political all the way down, any antagonistic moment could be viewed as a legitimate conflict that needs to be contained within the political process. The only constraints on excessive politicization are that opposing positions should peacefully coexist within a certain discursive and institutional order that we call politics, and/or they should be subject to the requirements of "the common symbolic space," that is, minimally shared normative foundations of liberal democracies (Mouffe, 2000).
I contend that these measures offered by Mouffe do not succeed in filtering out instances of politicization based on pseudo-science for two reasons. First, her conception of the common symbolic space, that is, based on the ideals of freedom and equality, is too narrow to properly limit politicization. Much of the political positions that rely on pseudo-science, for example, posttruth debates, are not directly related to disagreements about how to interpret "liberty and equality for all" (Mouffe, 2000, p. 113).
Second, even if we accept a broader view of the common symbolic space including certain epistemic virtues, Mouffe's measures are not capable of accounting for radical misinterpretations of our shared commitments as in the case of science denialism. This is because any satisfactory explanation of how the fundamental notions of the common symbolic space, for example, authority of science, are misapplied requires a prepolitical criterion of correct application. Mouffe's commitment to the poststructuralist social ontology is not conducive to restraining politicization since normative standards prior to the political are inconceivable according to this account.
The lack of distinction between good and bad politicization implies a form of pluralism without substantial limits.
I call this the problem of overly permissive pluralism. Although there are some limits on pluralism in Mouffe's political theory, I contend that they are still too permeable, as a variety of views based on conceptual distortion or empirical inaccuracy infiltrate into the domain of legitimate political adversaries.
The aim of this article is to develop an approach that minimally restrains politicization while still remaining committed to the main insights of the agonistic critique of depoliticization. In achieving this task, I will draw upon resources that are already available in Mouffe's and other agonistic thinkers' writings (Mouffe, 2000;Tully, 1989;Tully, 2008).
These theorists often utilize contextualist aspects of Wittgenstein's philosophy to lay out their contention that there is no unique rational solution to political problems. I, however, focus on how the very same philosophical foundations can deliver a theory of error which would eliminate certain political positions below the threshold of minimal coherence and empirical accuracy. This is possible if agonistic political theory is reoriented toward a more quasi-Wittgensteinian direction at the expense of the antagonistic view of social reality.
I argue that there could be justified limits on politicization without endorsing a rationalist-universalist conception of politics, that is, what agonistic theorists aim to avoid (Mouffe, 2005). I hold that the practice of rule-governed language use constitutes a normative framework that would justifiably constrain excessive politicization through a procedure of immanent critique. There are two steps to my argument. First, I show that a quasi-Wittgensteinian approach allows for conceptual error because there is always the possibility of a gap between agents' interpretation of a meaning and the meaning itself (PI,§198). 1 This insight is more clearly elaborated by recent philosophers like Sally Haslanger (2012) through the distinction between manifest and operative concepts. Manifest concepts are what language users think they mean by using a term, whereas operative concepts refer to actual social-empirical circumstances in which a term is regularly used.
Second, I argue that this quasi-Wittgensteinian outlook reveals a practice-based understanding of linguistic normativity that can be used to criticize excessive politicization. For instance, certain forms of politicization can be ruled out on the grounds that they rely on the incorrect use of a concept due to language users' improper dispositions which are not compatible with how the concept functions in social practice.
The article proceeds as follows: in the first section, I expound Mouffe's conception of the political and depoliticization, which is based on poststructuralist social ontology. In the second section, I argue that her ontological argument against depoliticization involves overly permissive pluralism, which significantly threatens the normative force of agonistic pluralism. In the third section, I argue that Mouffe's agonistic political theory should be reoriented more toward a pragmatist conception of language inspired by the insights of Wittgenstein and Haslanger. By this means, it is possible to develop limits on politicization that are non-universal but locally justifiable. Section 4 replies to several objections.

THE POLITICAL AND MOUFFE'S CRITIQUE OF DEPOLITICIZATION
Mouffe's political theory employs two fundamental notions which refer to different modes of human interactions: "the political" and "politics." By the notion of "the political," she means "the dimension of antagonism that is inherent in human relations" (Mouffe, 2000, p. 101). The antagonistic nature of human relations implies questions involving "decisions which require us to make a choice between conflicting alternatives" (Mouffe, 2005, p. 10). It is the conflictual social settings that are ineradicable in our interactions. Inspired by Schmitt's (2007) conception of the political, she holds that antagonistic moments are not resolvable by a pursuit of rational consensus among parties with opposing values and interests (Mouffe, 2005, p. 12). Mouffe's point seems to be more than an empirical observation that actual political agents are not capable of reaching a rational consensus. She rather contends that there are "conflicts for which no rational solution could ever exist" (Mouffe, 2013, p. 3). Therefore, the political refers to a metanormative standpoint, which holds that rationality cannot even hypothetically solve conflicts in human communities.
In contrast, politics "indicates [an] ensemble of practices, discourses and institutions which seek to establish a certain order and organize human coexistence in the conditions that are always potentially conflictual because they are effected by the dimension of 'the political'" (Mouffe, 2000, p. 101). This dense definition needs to be unpacked. First, the primary function of politics is "to defuse the potential antagonism that exists in human relations so as to make human coexistence possible" (Mouffe, 2018, p. 91). Second, the conditions of peaceful coexistence are always unstable because the return of antagonisms is an ever-present possibility (Mouffe, 2000, p. 131). In this sense, a seemingly stable social order always faces the possibility of slipping into a spiral of violence or other forms of severe conflict (Cross, 2017, p. 182). Third, as antagonisms can never be conclusively resolved, any political order is necessarily hegemonic (Mouffe, 2018, p. 55). It is hegemonic in the sense that political institutions act as if the interests and values of opposing parties have been rationally reconciled.
Mouffe's agonistic model of democracy aims to strike a proper balance between the political and politics. In other words, the task is to create the conditions of peaceful coexistence without denying the potentially antagonistic nature of social relations. This attempt at reconciliation explains how the transition from antagonism to agonism occurs. Agonism is a mode of social interaction among persons who "share a common symbolic space" despite their ineliminable conflicts about how to organize this space (Mouffe, 2000, p. 13). With the category of the common symbolic space, Mouffe seems to suggest that the members of a political community minimally share certain commitments, conceptual schemes and/or forms of life despite the lack of rational consensus on further political questions (Vasilev, 2015, p. 81). As a result, agonistic democracy aims to find a way for peaceful coexistence among groups with certain commonalities without denying the role of any conflicting party as a legitimate political adversary. Hence, Mouffe's approach allegedly extends the scope of pluralism in a way that would not be possible in the rationalist conceptions of political theory. This is because no particular claim-making about social relations can be rationally expelled from the domain of political contestations.
Mouffe's commitment to extended pluralism connects us to her critique of depoliticization. Her agonistic model of democracy allegedly increases the scope of pluralism because it claims that there are no prepolitical limits to what can or should be politicized. Following this, the critique of depoliticization is a key element in showing that we should recognize marginalized alternatives as legitimate political positions insofar as they are not rejecting the common symbolic space (Mouffe, 1993, p. 6). By depoliticization, I mean rendering a set of social relations and/or regulative ideas incontestable through deeming their alternatives irrational or unreasonable. Mouffe discusses different forms of rendering incontestable in her writings. First, depoliticization is a theoretical deficiency of what she calls the rationalist approach in political theory. For instance, she holds that the Rawlsian and Habermasian political theories have in-built biases excluding certain important questions from the domain of political action and deliberation. Hence, such a theoretical framework dissolves the possibility of contesting particular social relations (Mouffe, 1993, pp. 51-52).
Second, Mouffe sometimes discusses the phenomenon of depoliticization in relation to actual institutional orders.
In her analysis of British realpolitik, from Margaret Thatcher's neoliberalism to Tony Blair's "third way" politics, she contends that political elites' consensus on "the center" gave rise to "a technocratic form of politics according to which politics was not a partisan confrontation but the neutral management of public affairs" (Mouffe, 2018, p. 4). This constitutes an institutional form of depoliticization as the status quo excludes possible alternatives from the domain of public deliberation and legitimate claim-making. 2 Institutional depoliticization can render certain social relations incontestable in at least two different ways. One is that a particular domain of social life, for example, property regimes or family, is framed as a realm in which political interference is categorically unwelcome. Another way could be that an already politicized area is constrained to an artificially small number of options, that is, claiming that there is no alternative.
In her critique of depoliticization, Mouffe mainly exploits the ostensible boundary between the political and the social to discuss the possibility of something non-political. According to Mouffe (2005, pp. 17-18), "the social is the realm of sedimented practices, that is, practices that conceal the originary acts of their contingent political institution and which are taken for granted, as if they were self-grounded." While the conflicts and acts of exclusion are open in the political field, analogous relations of power are invisible in the domain of the social. They are invisible in the sense that ordinary individuals internalize these patterns of interaction in supposedly non-political domains. However, this does not mean that the true nature of the social is non-political. It is something like "false consciousness" in that their contingent "political institution" is concealed from uncritical eyes. What we take to be political is a contingent product of power relations. There can always be new frontiers that distinguish the political from the (non-political) social in novel ways. The exact location of these new frontiers is determined by the balance of power and "renegotiations between social agents" (Mouffe, 2005, pp. 17-18).
It therefore seems that for Mouffe, this distinction between the political and the social is almost a pseudodistinction. The central characteristic of the social is the invisibility of its political nature. Moreover, the depoliticization of the social has pejorative implications in Mouffe's political thinking because it allegedly leads to the artificial restriction of "other possibilities that have been repressed and that can be reactivated" (Mouffe, 2005, p. 18). Once a politically contingent phenomenon is framed as a necessary part of social life, this functions as an obstacle to emancipation which can be achieved through alternative configurations of power relations (Fossen, 2008, pp. 383-384).
For instance, depoliticizing gender inequalities by saying that they are grounded in "biological facts" would unduly limit the ways unjust gender relations could be undone. Given that depoliticization only conceals the political nature of the broader social phenomenon according to Mouffe, this restriction of the legitimate scope of political action would be unjustified. It is unjustified because depoliticization always relies on unsustainable claims, as social reality cannot be truly free from the political (Cross, 2017). Objectivity derived from facts and the norms of rationality cannot properly restrict how we should define the political realm since the social is "never the manifestation of a deeper objectivity exterior" to political practices (Mouffe, 2005, p. 18). 3 As any social relation can be duly politicized and as objectivity claims about truth and facts are produced within certain social institutions, the political seems to be conceptually prior to any authoritative epistemic norm, that is, pertaining to truth, rationality or justifiability, according to Mouffe. This view that social reality is political all the way down comes from Mouffe and Laclau's (2001) poststructuralist ontology, discussed in their Hegemony and Socialist Strategy. Let me briefly lay out the basics of their ontological argument. First, they start with the axiom that "every object is constituted as an object of discourse," in the sense that everything about which we can meaningfully talk is contained within a discursive structure (Laclau & Mouffe, 2001, pp. 107-108). For instance, even if a physical object exists independently of our thoughts, what makes this object a meaningful entity is its position within a discursive formation. A discourse is a "structured totality" in which the relations of difference and identity between objects are constructed (Laclau & Mouffe, 2001, p. 105). Second, they contend that every discursive formation is incomplete because social reality is contingent (Laclau & Mouffe, 2001, p. 106).
A couple of clarifications are needed here. The contingency requirement is that no specific manifestation of social reality is necessary or natural in that there are always alternative ways to see, organize, and conceptualize it (Cross, 2017, p. 181). Further, discursive formations are incomplete in that they always exclude certain possible conceptualizations. This is because there are simply infinitely many ways to categorize objects and properties. Drawing the relations between differences require distinctions between what is included and what is excluded. It is important to note that such acts of exclusion are not determined by a priori principles of rationality. They are rather arbitrary choices which reflect the preferences of the powerful.
Once the incompleteness of discursive formations is established, Laclau and Mouffe explain how this feature is connected to antagonisms. "Antagonism constitutes the limits of every objectivity, which is revealed as partial and precarious objectification" (Laclau & Mouffe, 2001, p. 125). The fact that every discursive formation is based on the exclusion of other alternatives makes objectivity claims vulnerable to the challenges by what is excluded. This precarity of discursively constructed objectivities implies that antagonism is the failure of language as a system of differences (Laclau & Mouffe, 2001, p. 125;Norris, 2006, p. 114).
In other words, the very act of meaning-making is hindered when a discourse encounters a resistance which cannot be resolved through its own logic. Further, as "the subject is constructed through language," the failure of our linguistic paradigms leads to an identity crisis or a crisis of subjectivity (Laclau & Mouffe, 2001, p. 126). Such an identity crisis is formulated in terms of the paradoxical relationship between the subject and the Other. On the one hand, "the presence of the 'Other' prevents me from being totally myself," because my subjectivity which is generated through a precariously objective discursive formation is denied by the Other (Laclau & Mouffe, 2001, p. 125). On the other hand, antagonism is the productive conflict between the subject and the "constitutive outside" that affirms the subject's identity through the us/them relation (Mouffe, 2000, p. 12;Norris, 2002, p. 556). In this sense, antagonism is ineliminable for two reasons. First, it shows the inevitable limits of any objectivity. Second, its conflictual existence has a major impact on the formation of any subjectivity.
What does the abovementioned poststructuralist argument tell us about Mouffe's categorical rejection of depoliticization? It basically says there are no rational limits about how social relations should be politicized or what ideas should be welcomed as the legitimate part of political life. This is because the existence of such limits would presuppose certain objective standards of justifiability that would be binding for opposing parties. However, any such standards are constructed through our discursive formations. As any discursive formation is created by the exclusionary and antagonistic acts of power, there is no prepolitical discursive space whose standards can be authoritatively used to critique certain instances of politicization.

THE BOUNDARIES OF THE POLITICAL AND OVERLY PERMISSIVE PLURALISM
The critique of agonistic political theory has often been centered on two different families of arguments. On the one hand, some critics argue that Mouffe's rejection of consensus-centric political theory is not sustainable as agonistic theory itself relies on the possibility of a minimal consensus (Erman, 2009;Knops, 2007;Vasilev, 2015). On the other hand, others hold that Mouffe and Laclau's poststructuralist ontological argument about the political nature of social reality has relativistic implications as it does not offer any specific normative orientation (Geras, 1988, p. 50;Townshend, 2004, p. 273). According to these scholars, as valid reasons to adopt a political view are only articulated within its own discursive formation, alternative political outlooks would be incommensurable, which suggests some form of relativism.
In this article, I am mainly interested in another kind of critique, which has some similarities to the relativism objection. I call this the problem of overly permissive pluralism. I contend that the lack of principled limits on the political is likely to undermine the quality of our political life as it does not effectively exclude certain problematic claims from the political process. Even if it does not amount to an entirely relativistic position, as one can still oppose these problematic views, Mouffe's agonism is tempted to expand the scope of legitimate political adversaries too much. I will now show how the politicization of science denialism illustrates my contention and explain why Mouffe's replies to this problem are not successful.
The lack of principled limits on what should be politicized is likely to justify the most scientifically inadmissible policy positions as a legitimate political adversary within a polity. Consider the public debate about the alleged causal link between MMR vaccination and autism. Despite the widespread scientific agreement that there is no evidence for the existence of such causal link, we observe an actual confrontation between the scientific community and policy-makers on the one hand, and antivaccine parents and pseudo-scientists on the other (Fitzpatrick, 2004). There are several ways we can characterize this conflict as political, that is, something more than an ordinary epistemic disagreement.
First, it inevitably pertains to what course of policy-making is desirable. Second, some scholars show that antivaccine attitudes among parents are shaped within "a social context of group identity" (Attwell & Smith, 2017, p. 189). This also seems to fit the picture of antagonistic conflicts among identities which Laclau and Mouffe often emphasize. Further, the antivaccination demands have been politicized in a partisan context when right-wing parties started working on legislation to drop the mandatory vaccination policy in Italy (Giuffrida, 2019).
These cases are clearly a politicization of some kind that would cohere with Mouffe's definition. However, following her poststructuralist social ontology, if there are no prepolitical standards of justifiability, how can we possibly critique such instances of politicization, which rely on false empirical beliefs and incoherent ideological convictions? If we say that these positions are to be recognized as legitimate political adversaries within the political process, limits on the relativistic tendencies would not be drawn properly. Accepting such problematic instances of politicization as legitimate adversaries would significantly deteriorate the quality of public deliberation and the performance of political institutions. A substantial amount of time and effort would have to be allocated to dealing with overly irrational convictions in political life. I believe this degree of pluralism is unhealthy for any functioning political system including contemporary democracies. For instance, in the climate change debate, when the media "represents industry-funded sources as on a par with scientists," it clearly hinders the prospects of policies based on the best available evidence (Anderson, 2011, p. 154). Further, I believe shared commitments to the basic epistemic values such as scientific truth are as essential for the functioning of human communities as the normative values (i.e., freedom and equality) of Mouffe's common symbolic space. The former are indispensable for a sense of orientation in most instances of knowledge production that are essential for the coordination and maintenance of contemporary social life.
Rejecting science denialism as a legitimate political adversary does not mean that we should not address the social-political roots of such unacceptable collective beliefs. Further, this rejection does not simply prescribe that we should entirely ignore what denialists say and never attempt to understand their mindset. Denunciation of bad politicization only suggests that the defenders of improper politicization should not be given access to the channels of political claim-making and public deliberation that legitimate political actors are. For instance, this kind of exclusion would mean not giving pseudo-scientists as much media coverage as scientists, and criticizing media outlets when their editorial policies are determined on the basis of profit-maximization rather than journalistic principles such as accuracy and truthfulness. Exclusion of the denialists from the political process is indeed compatible with communicatively engaging with hesitant individuals by organizing local forums of discussion where they are given opportunities to raise their concerns and the scientists make an effort to communicate the best available evidence.
Let me clarify that I am not defending a vision of science as a field that is entirely above the political. I accept the possibility that the scientific norms themselves could be a product of a sedimented hegemony reflecting the past record of power relations within the scientific community. Hence, there are certainly very good ways to raise political questions within our scientific enterprises as they often rely on contestable but implicit normative commitments. However, what I am suggesting is that there are also wrong ways to politicize certain social institutions. The attempts to politicize science are problematic when they are detached from the reality of how scientific practices operate. Hence, I believe the proper politicization of the scientific hegemony should primarily take place within the scientific community itself instead of a vulgar politicization making scientific norms vulnerable to the electoral pressures of democratic politics. 4 There are two main replies Mouffe presents against the dangers of extreme pluralism. In The Democratic Paradox, Mouffe (2000, p. 20) argues that extreme pluralism is not compatible with agonistic political theory because of the former's "refusal of any attempt to construct 'we' , a collective identity that would articulate the demands found in the different struggles against subordination." As extreme pluralism would imply incommensurability of alternative views, it would make it impossible to achieve a required level of hegemonic homogeneity created through the construction of collective identities. Similarly, her definition of politics also requires certain commonalities so that the peaceful coexistence of various social groups will be possible (Mouffe, 2005, p. 21). Since any political order presupposes this kind of homogeneity to properly function, Mouffe seems to believe that extreme pluralism fails to understand the political nature of human communities.
There are two shortcomings of this reply. First, Mouffe conflates the practical impossibility of accommodating extreme pluralism with justifiability of a particular hegemonic project. It might be true that political orders are necessarily hegemonic because certain alternative configurations of power relations are always off the table. As a result, one can plausibly say that extreme pluralism can never be realized in an actual political order. However, this does not say anything about what hegemonic projects political agents ought to engage in among several alternatives.
As the inevitability of exclusion does not justify a particular type of exclusion, Mouffe does not offer a criterion distinguishing a science denialist populist project from better options. Second, as collective identities are multiple, the necessity of constructing a "we" identity that would guarantee a healthy communal unity does not undermine the possibility of conflicts smuggling overly irrational sub-identities into the political field. Clearly, antivaccine movements do not jeopardize the future of Western democracy. Social identities associated with the denialist movements coexist with the fact that the same individuals share other, bigger identities that are crucial in the functioning of the political system. Hence, this objection does not give us a strong reason to dismiss the problematic instances of politicization like science denialism. Agonistic theory seems to allow extreme pluralism insofar as it stays within the domain of policy-making without harming the foundations of political unity.
Mouffe's second reply would be that certain instances of politicization can be justifiably excluded from the legitimate domain of politics on the grounds that they do not belong to the common symbolic space of the community (Mouffe, 2000, p. 13). Although this reply sounds similar to the previous one, it is more about finding a normative ground to filter out certain instances of politicization. The common symbolic space refers to a particular normative substance found in political communities. For instance, liberal democratic societies are bounded by "the ethicopolitical values of liberty and equality for all" (Mouffe, 2005, p. 32). Such shared commitments are "identified as the common good of the liberal democratic society" (Rummens, 2009, p. 379). Following this commonality in the basic normative framework of liberal democratic societies, Mouffe believes that agonistic political theory does not lead to extreme pluralism. Its scope of pluralism is limited by those who share a common symbolic space, despite their internal disagreements about how to organize it (Mouffe, 2000, p. 13).
I claim that this reply does not successfully tackle the problem of overly permissive pluralism either. First, the common symbolic space seems to be too narrow to sufficiently deal with the relativist tendencies displayed by denialists.
Most post-truth arguments in political life do not really engage in attempts to refuse the values of "liberty and equality for all." However, we can expand our conception of the common symbolic space to include certain epistemic virtues that are widely endorsed in contemporary liberal democracies: truthfulness, authority of science, etc. By this, agonists might expel denialism from the legitimate domain of the political.
Nonetheless, this move does not really solve the problem. This is because the mere existence of shared commitments is not sufficient to detect the cases in which these fundamental values are radically misapplied due to heavy ideological distortions or empirically false beliefs. Consider some of the right-wing populist arguments that aim to legitimize Islamophobia through the use of an ostensibly liberal discourse (Akkerman, 2005). Such arguments often appeal to the values of gender equality and tolerance to demonize the Muslim community by claiming that the submission of women and violent denial of diversity is the essence of Islamic identity. However, it does not seem possible to denounce such negative politicization of certain religious identities on the grounds that racist politics does not belong to the common symbolic space. They often proudly declare that they are committed to the fundamental values of liberal democracy.
Similarly, in the cases of science denialism, the problematic instances of politicization do not necessarily reject the value and authority of science, which is widely celebrated in liberal democratic societies. They rather hold empirically false beliefs about what counts as true science. Hence, it is not possible to rule out denialists' claims on the grounds that they openly resist the common symbolic space. The best one can say is that denialists misunderstand or misapply certain important concepts of the common symbolic space. However, as there is no prepolitical standard of justifiability, truth or rationality, agonistic theorists suffer from a lack of normative criteria to check if an instance of politicization can genuinely be derived from the defining commonalities of the political community. As a result, the problem of overly permissive pluralism is not defused because there are still many plainly false political takes that can be "legitimized" in relation to the common symbolic space.

AGONISM WITHOUT OVERLY PERMISSIVE PLURALISM
In this section, I argue that agonistic political theorists could adopt a strategy to identify unacceptable interpretations of the common symbolic space. I previously explained that inhabiting a common symbolic space is not sufficient to defuse the problem of overly permissive pluralism as there could be certain normative positions which radically misapply the basic concepts of this space. Illegitimate politicization of science is an example of this. I hold that, by relying on a quasi-Wittgensteinian conception of language, certain implausible instances of politicization can be ruled out on the grounds that they are based on incorrect uses of concepts. As a result, categorical rejection of depoliticization would be replaced by a more nuanced position according to which there are prepolitical standards to draw the boundaries of the political. However, the price one pays for this improvement is that Mouffe's poststructuralist social ontology is rejected. I will first review how a quasi-Wittgensteinian conception of language offers normative criteria in the uses of concepts. Second, I will clarify how such an approach can be used to criticize implausible instances of politicization.
Lastly, I will reply to some objections regarding the suitability of my revision for agonistic political theory.
Wittgenstein is not an alien philosophical figure in agonistic theory. In The Democratic Paradox, Mouffe (2000, p. 60) draws on the late Wittgensteinian philosophy of language to illustrate the kind of theoretical foundations her non-rationalist political theory is based on. According to Mouffe (2000, p. 62), Wittgensteinian insights offer a way to construct a context-dependent political theory. The contextualist approach basically suggests that liberal democratic political norms are neither universally valid nor uniquely rational. They are rather a contingent product of the cultural-historical context of Western civilization. As their emergence is tied to the development of certain types of social practices and cultural legacies, the rationalization of liberal democratic principles is only possible within its own context. As a result, normative systems of ideas and practices are deemed defensible within a particular "language game," that is, rule-governed activity of language use within a social setting. Also, there are multiple normative solutions to political problems, as "the criteria for the application of a term are not determinate. . . There is always a field of possible reasonable redescriptions" (Tully, 2008, p. 28).
Further, language games are not reducible to conceptual schemes. As they are embedded in regular social practices, it implies that uses of concepts can be defended by virtue of the fact that language users are already participating in these "forms of life," that is, "patterns in the fabric of human existence and activity on earth" (Pitkin, 1972, p. 132). The Wittgensteinian framework shows that what we practice as a member of a collectivity (e.g., linguistic community) is prior to what is rational to believe.
Mouffe is quite positive about the potential of Wittgenstein's insights to contribute to agonistic political theory.
However, she largely focuses on the non-rationalist aspects of his thought, ignoring some other essential features.
I will now expound on another aspect of Wittgenstein's philosophy, limits imposed by rule-governance, and present how this could help alleviate the abovementioned problem of overly permissive pluralism.
The rule-governance in language use indicates that our subjective interpretations of what a term means are not necessarily accurate. Individuals' conceptions of what a term means are merely interpretations, which can always be challenged by another interpretation (PI,(198)(199)(200)(201). However, what is objectively binding for community members is how a certain notion is regularly used within the communicative network of individuals (PI, §198). This sense of regularity creates the linguistic environment whose customs enable us to assess whether a member obeys or deviates from the rules of language use. Kripke (1982, p. 74) relates such rules following activity to what he calls justification conditions. This emphasizes that language use is inherently a normative process where the subjective linguistic experiences are to be assessed by the standards of the linguistic community.
It is important to note that such public standards of warranted language use are not constituted by the community members' shared beliefs about the definitions of concepts (PI, § §239-242). Providing the correct definition of a concept and having an understanding of how to properly use that concept are two different things. Grasping a rule and correctly applying it is a practice rather than having the right kind of belief or interpretation about the meaning of terms (McDowell, 1984, p. 339). Hence, the rules of language use are derived from the social practices corresponding to particular conceptual categories rather than our semantic beliefs, that is, what we think we mean by using a term.
Sally Haslanger's (2012) philosophy of language further develops this Wittgensteinian insight into a novel explanation of how language users can (in)correctly use concepts. By following the distinction between what individuals think they mean and what the term really means, her semantic externalist approach to meaning introduces the notions of manifest and operative concepts. Operative concepts are the ones "that actually determines how we apply the term to cases," whereas manifest concepts are what "users of the term typically take (or took) themselves to be applying" (Haslanger, 2012, p. 92).
In other words, the operative concept specifies the social-empirical circumstances in which language users regularly employ a term. In this sense, the meaning of a term "is determined not simply by intrinsic facts about us but at least in part by facts about our environment" (Haslanger, 2012, p. 395). This external environment is primarily a combination of repetitive social practices that shape the customs of language use (Haslanger, 2012, p. 224). By pointing out the gap between manifest and operative concepts, Haslanger (2012, p. 390) presents an explanation of how certain language users suffer a particular type of error. According to this, an individual's language use would be unwarranted to the extent that their beliefs and utterances are not coherent with the operative concepts, which depend on the patterns of linguistic behavior in their community. The relative independence between our beliefs and the rules of language use creates the possibility of error in the former.
I will now argue that the abovementioned quasi-Wittgensteinian outlook enriched by Haslanger's contribution can be utilized to filter out problematic forms of politicization. The problem posed by extreme pluralism is that Mouffe's critique of depoliticization is too permissive about the range of legitimate political positions. As the norms of rationality are not prior to the political, an instance of politicization cannot be ruled out on the grounds that it is fallacious or unreasonable. Even with the existence of a common symbolic space, there are simply too many interpretations of its central concepts, which allow for an unmanageable degree of pluralism. I hold that the quasi-Wittgensteinian approach is capable of narrowing the scope of pluralism by filtering out instances of politicization that rely on unwarranted interpretations of common symbolic space. 5 By common symbolic space, I mean its expanded version which includes the relevant epistemic values, for example, authority of science, accuracy etc., that are the backbone of liberal democratic societies.
For instance, science denialism politicizes health policies in their defense of the view that there is a causal link between MMR vaccination and autism (Fitzpatrick, 2004). They present an alternative "scientific" narrative and introduce a new conflictual dimension in politics. Denialists do not seem to question the authority of science as such. They rather assert that what they communicate is a true empirical claim demonstrated by "marginalized" scientists, that is, pseudo-scientists. Mouffe could attempt to criticize science denialism by claiming that respect for science is an integral part of the common symbolic space within contemporary liberal democratic societies. However, the core of the problem is about whether denialists have a proper understanding of what science is, rather than if they openly reject the validity of the liberal common symbolic space. Hence, we need a test assessing if a group properly employs the fundamental notions of the common symbolic space such as science, truth, rationality, etc. Keep in mind that I am not defending any armchair or positivist definitions of these concepts. Instead, I am testing denialists' use of concepts in relation to the socially constructed meanings of such terms.
I will now focus on the notion of "scientific truth," as much of the political conflict regarding denialism is about how to make sense of this term. I argue that the notion of "scientific truth" is radically misapplied in science denialism. This could be better understood by appealing to Haslanger's operative understanding of concepts. Denialists' use of these concepts does not accurately capture how these terms function in wide-ranging social practices. They seem to implicitly assume that scientific truth is tracked by what they personally find rational or convincing. This kind of attitude is explained by the notion of epistemic individualism, which suggests that certain individuals overestimate their cognitive skills to engage with complex problems (Levy, 2019). They also appeal to "epistemic authority" of "experts," that is, pseudo-scientists. However, their criterion of being a reliable expert similarly depends on what they find persuasive, trustworthy and/or compatible with their political/social identities (Kahan, 2015). One such example of science denialism is when the South African government supported narratives that downplayed the causal link between HIV and AIDS in favor of the view that scientific truth on the subject was contaminated by geopolitical considerations (Fassin, 2007, p. 7). This is different from how our notion of "scientific truth" operates in most domains of social life. Scientific truths are generated by an institutionalized epistemic division of labor. Hence, science is more than a systematic aggregation of justified empirical statements. It is "a system of social rules and related objects which, existing independently of any given person, constrains his behaviour" (Hartung, 1951, p. 44). The norms of proper scientific activity, such as peer reviewing, replicability and statistical rigor, are enforced and crystallized in organizations such as universities and research institutes. Our daily interactions in many aspects of social life from modern economy to education are governed by the presumption that there are institutionalized and authoritative practices responsible for generating scientific truths. Therefore, denialists' use of the term "scientific truth" is fallacious as it is based on what they personally find rational to believe. The notion of "scientific truth" has never been systematically applied to the non-specialized and individualistic matters of persuasion in our social practices. Institutionalization and epistemic division of labor are the central characteristics of the operative concept of "scientific truth." One might say that emphasizing the institutional nature of science is not sufficient to conclude that denialists misapply the concept because pseudo-scientists can also organize into similar institutional schemes, imitating real scientific institutions. However, what matters is not institutionalization per se but the type of practices embodied in institutions. Scientific practices are ultimately governed by certain cognitive values such as empirical accuracy, predictive success, and generality of application. Even if there is some room for subjective attitudes and disagreement in deciding how to rank these values, disagreements are rationally solvable once a particular cognitive value is prioritized (Laudan, 1986, pp. 31-49). These values normatively govern scientific practices, indicating what proper standards and methods need to be followed to realize cognitive ends. Pseudo-scientists' organized attempts fail to exhibit proper research practices that are adopted as the test of evidential acceptability by the mainstream scientific community (Pennock, 2011, p. 196). What does this say about denialists' misuse of the term "science"? I believe their use is unwarranted for the simple reason that they ignore the normative structure of scientific practice. The notion of science describes not only certain organizational forms but also a type of normativity inherent to the practices in those organized bodies.
It is not important if ordinary language users are really knowledgeable or not about the inner working of scientific practices. Their manifest concept, that is, their view of what science means, could be relatively independent of what practices the term objectively categorizes. Adopting the quasi-Wittgensteinian strategy proposed by Haslanger, the operative concept is deemed independent of language users' beliefs or knowledge. It is basically a social fact that the term is used to categorize activities with certain characteristics regardless of our awareness of what these characteristics are.
Once the rules of language use are established by the systematic use of a concept in a particular way, a warranted use of the concept would amount to compliance with the custom. However, denialists' use of the term "science" violates this as pseudo-scientists' activities objectively do not bear the properties that we find in the systematic use of the term in our social practices.
As a result, the politicization of the MMR vaccination can be ruled out by employing a type of immanent critique based on the sociolinguistic standards of the community. Since denialists employ a false conception of science, one cannot plausibly say that their concerns are the proper subject matter of politics which scientific authority is not capable of resolving. Scientific practices are not impartial or divided between two sides of the debate, hence creating space for legitimate political contestations. As there is no serious evidence in favor of the denialists' empirical claims, scientific authorities legitimately reject such claims about the causal link between the MMR vaccination and autism. Thus, there is no genuine pluralism of scientific takes about the issue at hand, which would transfer the disagreement into the domain of politics. 6 This shows that not every instance of politicization is equally justified or should be accepted as a legitimate adversary in the political process. There are certain prepolitical standards that determine the boundaries of the political vis-à-vis other aspects of social life. 7 Mouffe's insight about the necessity of the common symbolic space is essential here. Without such shared background, it would not be possible to defuse the problem of overly permissive pluralism because differences across groups would be unbridgeable. However, unless one does not openly reject the common symbolic space, Mouffe's approach is not sufficient to eliminate an improper degree of pluralism. This is because she does not offer a testing mechanism that enables us to say when a certain instance of politicization relies on a radical misinterpretation of the common symbolic space by employing unacceptably incoherent or false convictions. The quasi-Wittgensteinian approach I defend is instead a way to limit the political without endorsing a universalist or rationalist style of political theory. By providing a rule-governed understanding of language, certain individuals' political convictions are deemed unacceptable with reference to the socially binding norms of concept use.
While my solution defuses the extreme pluralism charge in that it constructs a threshold of minimum acceptability for legitimate politicization, it also demands revision in Mouffe's agonistic theory. This is because drawing boundaries for the political is incompatible with the poststructuralist ontology she endorses. According to the latter, social reality is inevitably political due to the constitutive role of antagonisms in social life. Switching to a quasi-Wittgensteinian conception of sociolinguistic community makes the elimination of some-but not all-antagonisms possible by virtue of our shared practices and dispositions. In a sense, shared forms of life give us a general sociolinguistic framework on which many competing discourses all together rely.
I believe the key political lessons of agonistic theory can still be maintained after abandoning the poststructuralist argument. Retaining the insights of agonistic theory means that we can still frame politics as primarily a conflict-driven field of human life. And for those multiple political positions that are not excluded from democratic debate, we can still preserve the claim that no rational consensus is possible. I will now show how this is possible in my replies to the potential objections.

OBJECTIONS AND REPLIES
I will discuss three objections to my quasi-Wittgensteinian solution to draw boundaries for the political. First, one might argue that the rule-governed approach to language does not give us a strict normative manual regulating every aspect of our linguistic practices. As Wittgenstein puts it, the use of a word "is not everywhere circumscribed by rules; but no more are there any rules for how high one throws the ball in tennis, or how hard; yet tennis is a game for all that and has rules too" (PI, § 68). Thus, although the use of language is governed by custom, there is always the room for innovative uses that have not been explained by the previous record of language use. Further, as linguistic communities are never perfectly homogenous, the diversity in the forms of life that different subgroups inhabit implies that there could be new uses of concepts corresponding to different social practices.
My reply to this objection is twofold. First, even if indeterminacy in the rules of language use makes it possible to use concepts in novel ways, this does not mean that interpretations which explicitly contradict the existing rules can be authorized (Godden & Brenner, 2010, p. 76). In the case of science denialism, the unwarranted use of the concept is not an innovative and unforeseen application. It is rather a subjective view of the concept which clearly conflicts with the established norms of scientific practices endorsed in society at large. Second, assuming that denialists have their own forms of life that is isolated from the mainstream segments of the linguistic community would imply that their own concept of "science" is a completely different one, reflecting how the term operates in the social practices of this subgroup. In an extreme case like this, it is true that a Wittgensteinian approach would not be capable of attributing any linguistic error to denialists. However, it seems implausible to hold that denialist groups are so isolated from the rest of the linguistic community. The recent empirical evidence suggests that denialists are indeed responsive to the findings of scientific consensus when pseudo-scientists' claims are effectively challenged in the public debate (Schmid & Betsch, 2019). Hence, their social world seems to be common with the rest of society to the extent that the former's errors can be corrected by the latter's standards.
The second objection is that my proposal to employ sociolinguistic normativity to limit the political undermines the key message agonistic political theory aims to deliver: the centrality of conflict in political life. According to this, as criticizing an incorrect application of common symbolic space presupposes that there is some correct interpretation, it falls into the rationalistic or truth-conducive paradigm where marginalized positions are simply deemed unreasonable. Hence, conflicts in politics are made invisible. In contrast, I believe that my position can be made fully coherent with the belief in the centrality of conflict in political life. Claiming that prepolitical truths necessarily destroy the agonistic vision relies on a narrow understanding of how conflicts can be of utmost importance in politics. First, a minimal threshold of acceptability does not amount to elimination of conflicts for those who are above the threshold.
Therefore, there can still be ineliminable conflicts among legitimate political adversaries. Second, rather than holding that there are no prepolitical truths, one can argue that politics inherently functions under severe time constraints and sociopsychological obstacles, which make convergence on these truths impossible. Therefore, the political importance of truth and rationality can still be considered secondary, not because of a poststructuralist assertion but since the nature of politics is institutionally unsuitable to generate consensus (Rossi & Sleat, 2014). Following this, agonistic political theorists can retain their conflict-centric conception of politics without endorsing the poststructuralist vision of truth or rationality. Lastly, just like the antagonistic conception of social reality, the Wittgensteinian view is similarly antiessentialist. Hence the revision I propose does not run the danger of falling into a universalist conception of normativity.
An important clarification is needed here. I am not suggesting that Wittgenstein offers an epistemic account of truth. In the traditional sense of the term "truth," semantic grounding "is not true, nor yet false" (OC, § 205). 8 This is because the ground is where the exchange of arguments is exhausted. His linguistic norms are derived from the foundations of meaning-making practices, and evaluate individual linguistic behavior by these standards. Hence, by truth or rationality I mean semantic truth or rationality, that is, the rules of proper linguistic conduct subject to socially determined customs.
Lastly, one might object to my argument by holding that the quasi-Wittgensteinian approach I defend suffers from a status quo bias. By prioritizing the established meaning-making practices over marginalized interpretations of common symbolic space, my approach would allegedly harm the prospects of counter-hegemonic politics. Developing counter-hegemonic alternatives is an essential feature of agonistic political theory (Mouffe, 2018, p. 79). However, this would not be possible if radical alterations in the way we organize the common symbolic space are not permissible. Hence, the objection would conclude that the Wittgensteinian emphasis on sociolinguistic customs should be dropped because it considerably narrows the scope of contestability in the common symbolic space.
I believe that my solution is not in tension with counter-hegemonic projects in general while it clearly does not offer automatic support to every counter-hegemonic position. The claim that social constructionist conceptions of meaning have an inherent status quo bias stems from a mischaracterization. If one takes linguistic rule-governance only as a matter of imitating the past record of connecting words with objects, there is no doubt that my proposal would be too conservative. Yet the proper use of language is more complicated than that. In her interpretation of Wittgenstein's work, Pitkin (1972, p. 183) suggests that the correct use of normative concepts such as justice is determined by the standards of judgment in addition to how the term is already applied to particular cases in wide-ranging social practices. For instance, when we learn how to use the concept "justice," the very normative nature of the concept requires us to learn how to implement certain standards of judgment in specific contexts. This knowledge cannot be simply extracted from the past record of the concept that categorizes certain practices as just. One further needs to consider the specific facts of the situation at hand and reflect on how to make a judgment in relation to these facts and endorsed standards of judgment: fairness, equality of opportunity, etc. Hence, the correct use of the concept depends on how that concept is linked with other related concepts. As the context, that is, facts of a situation, changes in new cases, the way we draw such interrelationships between concepts are largely shaped by our subjective interpretations of externally determined standards, which enables one to effectively depart from pro-status quo convictions.
This understanding of rule-governed language could be quite counter-hegemonic in that the particular judgments it allows for may lead to interpretations of the common symbolic space that are considerably different from mainstream positions. In other words, counter-hegemonic interventions can be made possible by employing the standards of judgment in alternative ways when we discuss the normative foundations of the common symbolic space. However, this is different from, for example, the way far-right politicians apply certain notions of the common symbolic space such as (gender) equality in Islamophobic arguments. I would like to discuss the example of anti-Islamism because, in contrast to science denialism, its error fundamentally relies on the mischaracterization of the object (Islamic identity) of judgments regarding political values (e.g., gender equality) rather than a misinterpretation of what the alleged value (i.e., gender equality or scientific truth) means. My quasi-Wittgensteinian approach not only regulates the ways we use the epistemic notions of the common symbolic space, but also tests whether the social categories to which we apply political values are accurately understood. In cases of racist politics, the source of misuse of the common symbolic space is premised upon an empirically unsustainable view of Islam as a "unified, static and essentially fundamentalist" religious identity (Akkerman, 2005, p. 347). Hence, such political positions can be ruled out on the grounds that they are incompatible with the reality of Islamic religious identity, which is heterogeneous and non-essentialist in our social practices.

9
Of course, Islamic identity itself is not an element of the common symbolic space as the latter consists of the most basic normative foundations of a community, for example, equality, freedom, and tolerance. It also does not belong to the expanded version of the common symbolic space where certain epistemic values would be included such as truthfulness, authority of science, coherence, etc. However, empirically inaccurate views of Islam and/or science still lead to incorrect applications of the common symbolic space for two reasons. First, they obviously violate the norm of truthfulness in that they are not able to provide sound interpretations by the shared semantic standards of the community (Sleat, 2007). Second, racist politicians' particular judgments about equality or tolerance are also incorrect by virtue of the fact that the object of such judgments, that is, Islamic identity, is ill-defined.  (2007) for taxonomies of depoliticization. 3 One can find the term "political realm" controversial as Mouffe's interpretation of the political implies a quality of social relationships rather than a distinct location. However, she sometimes talks of the political as a domain (Mouffe, 2018, p. 87).
One possible interpretation is that the political as a realm is the subset of social life in which its political quality is widely recognized. 4 Indeed, even social constructivist thinkers like Bruno Latour (2018), who is the archenemy of positivism, is highly skeptical of the second type of politicization. Second, even when social movements can justifiably politicize science, as in the case of contesting normatively controversial assumptions of economics, there is a sense in which such attempts should be in alliance with dissenting subgroups within the scientific community. This is because contemporary societies seem to grant scientific institutions a degree of autonomy within their domain of activity. 5 I use the term "unwarranted" to denote that a particular notion of the common symbolic space is used in a way that is not compatible with the linguistic conventions of the community. cates of science. However, this does not undermine the possibility that some groups might openly challenge the authority of science as such due to other normative considerations like religious freedom. 7 I admit that ruling out certain ideas on the basis of alleged prepolitical standards could be in tension with Mouffe and Laclau's defense of the historical contingency of social formations. While I agree that our sociolinguistic traditions might be contingent, the options that are epistemically available to us still seem to be limited due to the path dependencies created by our intellectual and institutional legacies. Hence, it might not be entirely arbitrary to rely on our local standards. 8  . On certainty. (G.E.M. Anscombe & D. Paul, Trans.). Oxford: Blackwell. 9 The critique of any religion is welcome to the extent that critics do not conflate religion as social practices of a particular community with religion as a theological belief system (Spinner-Halev, 2005). Hence, the only constraint is that one should not apply criticisms of a religious doctrine to the actual religious populations who do not necessarily follow theological prescriptions in their social practices.