THE PARTICULARIST THESIS AND THE ANTICOSMOPOLITAN THESIS

The particularist thesis can be stated in terms of collective identity, solidarity, or community. It can even be stated in terms of cultures or collective modes of life, as when Anthony D. Smith (1990, 171) asserts that

Claims such as this, about the inherent nature of collective modes of life, speak of a conceptual or metaphysical truth not open to empirical falsification.

Mouffe (2000, 213) advances the particularist thesis explicitly in terms of identity when she says that “collective identities can only be established on the mode of an us/them.” But solidarity and community are related to identity as well. In the ordinary sense of the word, community entails its members' identification with the group that constitutes it and, thus, entails a collective identity; in the stronger, moralized sense of the word, community also entails mutual concern among its members, i.e., a sense of solidarity, a disposition to “give each other's interests some noninstrumental weight in their practical reasoning” (Mason 2000, 27).⁷

And though it is an empirical question, solidarity— at least the kind required for democratic politics—is often thought to depend on a shared collective identity (Cohen 1999, 263). Since it turns out to be the crucial version of the thesis, I focus here on the particularist thesis stated in terms of collective identity.

Insofar as it is true that democratic politics requires, as its motivational basis, a sense of solidarity grounded in a shared collective identity,⁸

the particularist thesis implies that the cosmopolitan dream of a democratically legitimated and just global political order centered on a global human identity is in principle incoherent, or impossible on metaphysical grounds. The anticosmopolitan argument can be reconstructed as follows.

1. Collective identity (and the sense of solidarity centered on it) is inherently particular (i.e., any collective identity presupposes another collective identity that is its external other).
2. Therefore, a global human identity (and a sense of solidarity centered on it) is impossible.
3. Besides shared material interests, democratically legitimated politics requires solidarity centered on a shared collective identity capable of motivating such solidarity and joint political action.
4. Therefore, a democratic cosmopolitan political order is impossible.

Premise 1 is the particularist thesis (in its weak form) and Thesis 4 is the anticosmopolitan thesis routinely drawn as a consequence. Premise 3 articulates the empirical assumption that democracy requires a shared collective identity, or that it requires a sense of solidarity that in turn depends on shared identity. (The entire argument can also be restated in terms of “community” instead of “solidarity.”) I do not concern myself with Premise 3: for the sake of argument I simply take it for granted. My thesis is that, independent of the truth or falsity of Premise 3, the four-step anticosmopolitan argument fails: Premise 1 is false. While individual identity may indeed be inherently particular, collective identity is not, and the arguments for Premise 1 proceed only by equivocating between these two senses of identity. There are undoubtedly many real-world obstacles facing the ancient cosmopolitan dream of global solidarity and a community of humankind. But conceptual incoherence or metaphysical impossibility is not one of them.

HEGEL AND TAYLOR: THE RECOGNITION AND DIALOGIC ARGUMENTS

The notion that identity either conceptually or metaphysically requires an external other has a long history, a history that certainly includes Rousseau's second Discours and Émile. But perhaps the most influential formulation of this idea in the European tradition is Hegel's. I think it best to construe Hegel's argument as metaphysical rather than conceptual, so I want to begin by first getting the conceptual argument out of the way. If Mouffe's (1993, 84) claim that “every definition of a ‘we’ implies the delimitation of a ‘frontier’ and the designation of a ‘them’” is to be understood conceptually, and if the terms “we” and “them” refer to human beings, then the problem with Mouffe's argument is that its premise is false.⁹

We can see the falsity of the premise by simply providing, pace the conceptual argument, a definition of a human “we” that does not designate a human “them.” Definition: Let x refer to all human beings. There is nothing incoherent about this. Now it might be objected that the definition of “we” in terms of humanity begs the question because it includes a reference to “humans,” which is precisely what is in question. This would be to miss the point. The rather simple point is that whatever the definition of “human being,” there is nothing incoherent about defining “we” such that it includes all the individuals picked out by the definition. Alternatively, it might be objected that the truth of particularism is demonstrated by the fact that even a universal human “we” presupposes a nonhuman “them.” And indeed it does, but this has no bearing on the specific particularist thesis that human collective identities presuppose other human collective identities. In fact, the conceptual argument cannot even motivate drawing the line at human beings: that a universal human “we” presupposes a nonhuman “them” is a function of the qualifier human in “universal human we,” not a function of we. There is nothing conceptually incoherent about an identification that goes beyond humanity, to include, for example, all sentient beings. Indeed—Eurocentric humanist intuitions notwithstanding—there is nothing conceptually incoherent about an identification with the entire existing universe, such as that espoused by some pantheists.

Much more interesting for particularists is Hegel's (1977) argument. The premise of his argument is that selfhood involves the desire for self-certainty, that the self has a fundamental desire for self-consciousness and self-worth. The self first attempts to achieve self-certainty through the appropriation of external physical objects in the world, by asserting dominion over them and integrating them into its plans and projects. But since inanimate objects can only reflect back to the self an image that the self already has of itself, and not the self-certainty it seeks, the satisfaction taken in external objects simply leads to new desires for new objects. This ultimately renders the desiring self-consciousness dependent on new objects, unable to establish its own independence. So self-certainty through the appropriation of nonhuman objects is inadequate for a human being's sense of self-worth: self-consciousness needs an object that does not cease to be an object for it, an object capable of reflecting back to the self its own conception of itself as free. Self-consciousness requires recognition by an other who is also a self-consciousness— another human being. And this recognition must be a mutual recognition, because if the self dominates the other, then the other's recognition will not give the self what it requires, since by dominating the other the self denies to the other the freedom it must have if its recognition is to be worth anything (cf. Wood 1990, 86–92).

Taking a cue from Rousseau and Hegel, Charles Taylor (1994, 32–33) argues that the “close connection between identity and recognition” should be understood in terms of the “fundamentally dialogical character” of human life:

Taylor's point is sound. But anticosmopolitan particularists are mistaken to think that either Hegel's or Taylor's argument provides them any support: for the particularist thesis in the anticosmopolitan argument outlined above tacitly assumes that what is true of our individual identities is also true of our collective identities. This assumption is false. The Hegel–Taylor argument shows why (1) an individual is able to attain self-consciousness only in relation to an external other, dialogically. This is because either (1a) to see myself as a self I need to be recognized by an external other I also recognize as a self (Hegel), or (1b) my sense of self is only developed and clarified dialogically with external others (Taylor), or both ([1a] and [1b]). But it does not follow (from either [1a] or [1b]) that (2) the individual can only identify with a collectivity that excludes some set of individuals. True, if (a) we accept the Hegelian account, the recognition necessary to an individual's sense of self-hood must be external, because none of the constituent parts of the individual self would possess the independent agency needed for a normatively relevant act of recognition. If we grant Hegel's view of self-consciousness, then the self's own internal recognition of itself as a self is dependent on external recognition.¹¹

But the recognition required by a collective identity can come solely from the (other) individuals who make up that collectivity—an option that would not analogously be available for the development of individual identity itself. In other words, a collective identity can receive either the external recognition of an external other or the internal recognition of its own constituent parts, and the latter does not necessarily depend on the former.¹² Similarly, if we reject the Hegelian view of self-consciousness but (b) accept Taylor's less metaphysically demanding view that identity is formed dialogically, then in the case of an individual, dialogue implies an external other with whom dialogue must have been at some point conducted; but in the case of a collectivity, the dialogue might simply take place among its own constituent members. Taylor's argument appeals to the fact that identity formation occurs through a linguistically mediated process of socialization. For an individual, socialization requires interaction with external others. But socializing an individual to identify with a collective identity could, rather obviously, simply occur through interaction with individuals who also identify with it.¹³ The point is not that the constituent members who engage in mutual recognition or dialogue must be paradoxically specifiable prior to the constitution of the collective identity; the point is simply that the constitution of the collective identity via recognition or dialogue is not dependent on the existence of excluded individuals. Neither Hegel's nor Taylor's arguments necessarily commit them to the particularist thesis.

In other words, the arguments in defense of the particularist thesis (that only collectivities with an external other can be the basis for identity) suffer from a fallacy of composition. Whether collectivities with an external other are better able to win identification is an empirical question. There are no conceptual or metaphysical grounds for thinking that communities “just are particularistic,” as Miller asserts, or that “collective attachments are always affirmed in opposition to others,” as Schnapper assumes.

As such, we should also reject Miller's (1989b, 239) assertion that loyalty to the human race is meaningless, because loyalty implies partialism:

Again, the argument suffers from not being attentive to the distinction between individual and collective identity. It fails to distinguish between an “other” external to the collectivity to which I belong and other individual members of the collectivity to which I belong. True, the latter are not “other” insofar as they belong to the same collectivity to which I do, but they are other insofar as they are not me. Loyalty to the human race could very simply mean sacrificing my own narrow personal interests in favor of the interests of the human race as a whole (as when, for example, I make enormous personal sacrifices to ensure that the Earth's air remains pleasant for everyone to breathe). Furthermore, if humanity is in turn constituted by smaller collectivities, and not just individuals, then if I professed loyalty to the human race, I might very sensibly mean that when the interests of humanity as a whole clash with the more parochial interests of some part of humanity to which I belong, I would favor the interests of humanity. To call this nonsensical by definitional fiat is just that— definitional fiat.

THE PECULIAR PARTICULARITY OF NATIONAL SOVEREIGNTY

None of this is to suggest that there are no forms of collective identity that necessarily presuppose an external other. There very well may be. Indeed, I would like to suggest that the particularist thesis that collective identity, solidarity, or community necessarily presupposes an external other has seemed so self-evidently true to so many, despite being based on a patent fallacy, because our intuitions on this matter have been misled by the modern conceptions of sovereignty and nation.¹⁵

F. H. Hinsley (1986) has argued that the modern notion of sovereignty developed in tandem with the modern state within the context of two problématiques: the nature of legitimate authority within a political community and the regulation of relations between separate political communities. First, the modern notion of sovereignty came to involve the idea that there is a final and absolute political authority within the political community (26). This is what has come to be called “internal sovereignty,” which Hedley Bull (1977, 8) defines as “supremacy over all other authorities” within a particular territory and over a particular population (cf. Krasner 1988, 86). But second, the modern notion of sovereignty asserts that no other authority exists outside of the political community (Hinsley 1986, 26). This furnishes the state's “external sovereignty,” “by which is meant not supremacy but independence” from “outside authorities” (Bull 1977, 8). As such, via the application of reciprocity, the notion of sovereignty was deployed in practice to regulate interstate relations through the mutual recognition of each other's internal sovereignty over its own territory and population (Hinsley 1986, 158).

As constructivists have noted, this reciprocal recognition has become a constitutive feature of the modern practice and institution of state sovereignty (Ruggie 1993, 162). As Alexander Wendt (1992) once argued,

If this is right, then at least state sovereignty would follow a particularist logic of external mutual recognition. (Indeed, this is Hegel's view).¹⁷ The implication would be that collective identities centered on the modern sovereign state presuppose an external other.

And there is a sense in which this is right. The meaning that sovereignty has acquired for us is, in part, constituted by its role in the Westphalian era of regulating interstate relations. This explains why Anthony Giddens (1987, 281) puts the matter in conceptual terms: sovereignty, he claims, “only has meaning in the context of a reflexively regulated system larger than any one state.” So if sovereignty has come to mean two things conjoined, that is, internal sovereignty and the reciprocal recognition of this internal sovereignty among states, then sovereignty has come to be defined in a way that conceptually presupposes more than one sovereign power. But the combination of “internal sovereignty” with “external sovereignty” is itself a contingent, historical, not conceptually necessary, development.¹⁸

The modern ideology of state sovereignty is so pervasive, of course, that the contrary assumption, that internal sovereignty somehow entails external sovereignty, is almost as widely held as the particularist thesis itself (e.g., Philpott 1995, 357). But to say that there is, in the political community, a final political authority requires conceptually that there not be others outside of the political community who can override that authority; it does not require conceptually that there be others outside of the political community who cannot override that authority. Sovereignty has come to mean, for us, both internal and external sovereignty, precisely because it has been deployed in the Westphalian era as a principle of legitimation regulating interstate relations.¹⁹ But a “world sovereign,” sovereign in the sense of being an ultimate political authority, is not conceptually incoherent; it would simply strip sovereignty of the additional meaning it has for us as an institutionalized way of regulating interstate relations. Indeed, Claude Lefort (2000, 33) argues that in France the innovation of the Middle Ages was precisely the delimitation of sovereignty: inscribing sovereignty on a limited, bounded territory in formulae such as “empereur dans son royaume” was initially deeply paradoxical, for such formulae combine “the idea of imperium [previously] reserved for the Sovereign of the world with that of a power exercised within the limits of a country.”

As Stephen Krasner (1999, 3–4) points out, the different senses of sovereignty both are conceptually independent and have been routinely disaggregated in practice. Indeed, the two components can be disaggregated further: for Krasner, there are four “meanings of sovereignty.” (1) The claim to legitimate authority and effective control within a territorially circumscribed polity (what Krasner calls “domestic sovereignty”) is distinct from (2) the capacity to regulate movement across the circumscribed borders (“interdependence sovereignty”). Even (3) “the exclusion of external actors from authority structures within a given territory,” which he calls “Westphalian sovereignty,” is distinct from (4) what he calls “international legal sovereignty,” or “practices associated with mutual recognition” between territorial entities with formal judicial independence. Krasner demonstrates not only that, in practice, states have possessed sovereignty in each of these senses without the others, but that in some cases the possession of one kind of sovereignty can actually undermine another. It is only sovereignty in the sense of international legal sovereignty that conceptually presupposes recognition by another sovereign. Sovereignty in the sense of border control requires an other, but not necessarily another sovereign—just other territory. And Krasner's “domestic sovereignty” and “Westphalian sovereignty” do not conceptually presuppose the existence of an external other at all.

Despite the fact that, in practice, states have possessed the various kinds of sovereignty in different combinations and degrees, the modern ideology of sovereignty nevertheless combines all four senses and construes them as mutually entailing.²⁰

Given this (historically contingent) combination—given what sovereignty has come to mean ideologically—any collective identity that is linked to this modern notion is no doubt prone to reproduce its requirement of external recognition. This is not to say, of course, that alternative ideologies of (say, divided or dispersed) sovereignty have been completely crushed—far from it. The point is, rather, that the ideological combination of these distinct attributes has powerfully shaped the terms of the debate.

So it is rather unsurprising that national identity has come to be conceived in this way as well. After all, a typical feature of the modern phenomenon of the nation is the political claim to sovereignty over some territorially defined space.²¹

Benedict Anderson (1991, 6) defines the nation as “an imagined political community” that is “imagined as both inherently limited and sovereign.” Since the modern ideology of sovereignty construes the sovereign as necessarily one sovereign among others, to say that the nation is imagined as sovereign is already to have said that it is imagined as limited. So it is no surprise that national identity as we know it today is, like the modern notion of sovereignty, necessarily particularist. Jonathan Rée (1998, 83) has even argued that identification with a nation presupposes the existence of other nations: Rée's argument is that to transform local sentiments of attachment, which find their sources in the personal experiences of face-to-face interaction, into the necessarily imagined realm of the nation, one must imagine one's nation as bounded and in contrast to other nations, because of the notion's embeddedness in an imagined international order. But contrary to what the particularist might suggest, this does not establish that individuals can come to identify with an imagined, non–face-to-face community only if that community is particular and bounded. Rather, the point must remain specific to national identity: identification with the nation, as opposed to just any form of imagined community, presupposes the existence of other nations. What the particularist thesis falsely claims for collective identity as such might be true of some collective identities.

To sum up, then: even the apparently necessary particularism of national identity depends on the nation having become invested in the modern notion of sovereignty. With the rise of the idea that sovereignty inheres in the nation, if the nation just is, by definition, a community that is imagined as sovereign, and if sovereignty refers to an institutionalized way of regulating external relations with other sovereigns, then national identity is necessarily particularistic. Again, this concatenation of ideas is historically contingent. Liah Greenfeld (1992, 8–14, passim) has in fact argued that the conception of the nation as a particular community, i.e., one whose uniqueness requires a contrast with other nations, is a late, eighteenth-century, development. If she is correct, then it is no surprise that this development coincides with the rise of the doctrine of national sovereignty.

SCHMITT: THE POLITICAL ENEMY ARGUMENT

Does any of this show that a collective identity centered on “humanity” could in principle be the basis for a democratically legitimated cosmopolitan political order? Such a conclusion might still be denied by urging that, even if collective identities as such need not presuppose an other, specifically political collective identities do. This is the thrust of Mouffe's (2000, 43–44) Schmittian claim that humanity “is not a political concept”; it is also what motivates her endorsement of Schmitt's conclusion that “political democracy cannot be based on the generality of all mankind” (Mouffe 2000, 40). Indeed, the reach of the Schmittian objection is not restricted to specifically cosmopolitan projects: the objection, if valid, would show that my argument about collective identities has no bearing on the political sphere at all.

What I want to demonstrate is that Carl Schmitt's argument, which centers on an analysis of the nature of sovereignty, the modern state, and “the political” sphere, far from providing a wholly novel justification for the particularist thesis, instead provides a brilliant illustration of how the modern ideology of sovereignty helps to mask the fallacy, already noted, behind the particularist thesis: Schmitt's “political” argument in defense of the particularist thesis, and the anticosmopolitan conclusion he draws from it, mirrors the same fallacious structure of the recognition and dialogic arguments for particularism—it rests on a systematic conflation of the necessity of internal and external “others.”

For Schmitt (1996, 19, 47), the modern sovereign state is “a specific entity of the people,” “an organized political entity” that seeks to maintain internal peace, is “territorially enclosed,” and is “impenetrable to aliens.” At least for the early, pre-Nazi Schmitt (1985b), sovereignty inheres, within the context of a Rechtsstaat, in the one who “decides what constitutes an exception” in the application of general rules to particular circumstances (43).²²

This presupposes, with Jean Bodin, a distinction between laws and exceptional decrees or measures and, thus, between ordinary times and exceptional or emergency situations.²³ As Schmitt puts it elsewhere, “Sovereign is he who decides on the emergency situation [or exception, or state of emergency (Ausnahmezustand)]” (1922, 8; 1985a, 5).²⁴ The nature of political emergency arises, for Schmitt (1996), from the fact that the political sphere is inherently constituted by the distinction “between friend and enemy” (26). “The concept of the state,” he asserts, “presupposes the concept of the political” (19) and the political is always haunted by the spectre of existential struggle: As Gopal Balakrishnan (2000, 148, 192) points out, Schmitt later abandons the sharp distinction between laws and emergency measures and, by the time he becomes a Nazi, completely jettisons the theoretical apparatus of the Rechtsstaat: the “total” sovereign legitimately rules by decree. What remains constant, however, is Schmitt's commitment to viewing the sovereign as the one who decides matters of existential struggle. Rechtsstaat or no, sovereignty is a question of who “decides the extreme case (Konfliktsfall) and determines the decisive friend-and-enemy grouping” (Schmitt 1987, 43; 1996, 43).

Schmitt's (1996) anticosmopolitan conclusion derives from an argument whose structure is similar to the anticosmopolitan argument I considered earlier, though it begins with a stronger premise:

Schmitt thus introduces two modifications to the anticosmopolitan argument. First, he glosses the particularist thesis in its stronger, adversarial form. Second, he embeds the particularist thesis within a specifically “political” context. Schmitt's political version of the anticosmopolitan argument can be reconstructed as follows.

1. Any political entity presupposes an adversary.
2. An entity's adversary is other than the entity itself (i.e., it is an other to the entity).
3. Therefore, any political entity presupposes an adversary who is its other.
4. The sovereign state is a political entity.
5. Therefore, the sovereign state presupposes an other (another sovereign state who is its adversary).
6. Therefore, a sovereign global state is impossible.

Premise 1 is a political version of the adversary claim and Premise 2 reflects the exclusion claim. Thesis 6 is Schmitt's version of the anticosmopolitan thesis. Steps 1–5 constitute Schmitt's argument for his political and adversarial version of the particularist thesis (Thesis 5), construed in terms of sovereignty. I take Premise 1 for granted for now in order to notice the (fallacious) structure of Schmitt's argument for his version of the particularist thesis. My thesis is that by embedding the particularist thesis within an argument framed in terms of sovereignty, Schmitt helps himself to an equivocation between the internal and external aspects of sovereignty; in other words, the argument conflates internal and external “others.”

More particularly, the derivation of Thesis 5 (that the state presupposes an other state) depends on an equivocation between two senses of “political entity” and/or between two senses of “other.” “Political entity” can refer to the state as a whole, that is, the entire population, territory, institutions, etc., within its bounded space— call this the comprehensive sense of political entity. Or it can refer to a component part of the state, for example, the sovereign power within it—call this the partial sense. Similarly, an entity's “other” can refer to another entity that shares no constituent parts with it— call this the external (or strict) sense of “other.” (This is the sense in which I have used the word so far.) Or it might refer to another entity whose constituent parts are not fully coterminous with it, i.e., an entity that is not identical to it—call this the expansive sense of “other.” So defined, the external (or strict) sense of other is a species of the expansive sense. One example of a given political entity's other in the expansive but not strict sense might be another political entity some and only some of whose citizens (or territory, institutions, etc.) are shared with it: for example, England and Scotland share some citizens and institutions (the British Parliament) but not others (the Scottish Parliament). Call this an overlapping other. Another kind of other in the expansive, nonstrict sense is one that is simply a part of the first entity: call this the encompassed sense of other. For example, Scotland is Great Britain's encompassed other. The converse would be the encompassing sense of other.

The plausibility of Thesis 2 (that an entity's adversary is other than itself) depends on construing “other” in the expansive sense; the restrictive external sense of “other” would implausibly rule out the possibility that the adversarial relation is between an entity and one of its constituent parts. Canada's “adversary” may be Russia (its external other) or, as arguably often is the case, it may be Quebec (its encompassed other). But once “other” in Thesis 2 is understood in this expansive sense, if Thesis 3 is to be derived from Theses 1 and 2, then it must also be understood as leaving open the possibility that the adversary is an internal adversary, i.e., a constituent part of the political entity, an “other” in the encompassed sense. Hence the only way to construe, in Thesis 3, “other” in the strict, external sense is to assert it independently of Theses 1 and 2. But even if it is so asserted, then my point can be made in a different way: if “other” is construed in the sense of external other, then Thesis 3 is plausible only if “political entity” is understood in the partial (and not comprehensive) sense—as the part of the state that is sovereign, for example. For the sovereign's adversary, as Schmitt knew perfectly well, could be a domestic enemy.²⁵

The upshot is that Thesis 3 (that any political entity presupposes an adversary who is its other) only makes sense on one of two assumptions: either “other” is to be understood in the expansive sense or “political entity” is to be understood in the partial sense (or both). The problem is that in neither case can Thesis 5 be logically derived from the premises. For the derivation of Thesis 5 from Theses 3 and 4 either requires construing “other” expansively, in which case its claim that “the sovereign state presupposes an other” can refer to an encompassed other—for example, a part of the state who is not sovereign—or “sovereign state” (in Thesis 4) must be equated with the partial sense of political entity, as not referring to the entire state, in which case its strict other may be a nonsovereign within the political entity. The outcome in both cases is the same, and in neither case does it follow that, as Thesis 5 claims, the sovereign state presupposes another sovereign state. A sovereign state simply presupposes something over which it is sovereign. So even if “the political” depends on the friend–enemy distinction, it does not follow that a sovereign state's enemy must be other territorially organized states. That state violence is always a present possibility, in other words, does not imply the existence of “others” external to the jurisdiction of the legitimated use of violence.

The fallacious structure of the Schmittian political enemy argument for the particularist thesis mimics, of course, the structure of appeals to the recognition or dialogic arguments: while the latter appeals fail to see that a collective identity can be “recognized” both by an external “other” in the strict sense and by its own constituent parts, the Schmittian argument fails to see that a sovereign state may find an adversary both in an “other” state (i.e., an external other) and domestically. The role that “sovereignty” plays here in masking this equivocation between the strict and expansive senses of “other” is rather clear: the point is that Schmitt's argument in The Concept of the Political rests on a systematic conflation of internal and external sovereignty—which of course he believes go together.

The upshot of my critique as it stands, however, is rather grim for the cosmopolitan thesis that Thesis 6 opposes: I have refuted the derivation of Thesis 6 by arguing that the adversary—by which Schmitt means enemy—of the sovereign state may be an internal adversary. And a Schmittian might be willing to concede this: as Balakrishnan (2000, 110) points out, at the time of writing The Concept of the Political in the 1930s, there

So in principle the enemy may be an external enemy (other nation-states) or it may be an internal enemy (the proletariat or the capitalists), and “the question of which of these two conceptions was the politically decisive one could be determined only historically” (Balakrishnan 2000, 110–11).²⁶ But if indeed a political entity presupposes an adversary, then it appears that a cosmopolitan political order taking the form of a world state could be erected only on the backs of internal enemies whom the state dominates and seeks to crush.²⁷ A world state would be a world domination of one part over another—a worldwide dictatorship of the proletariat, for instance. The world state would be—to raise the old bogeyman once again—a terrifying nightmare. And if not a world state, then a cosmopolitan order would be an imperialist order à la the British Empire, imposed by a world hegemon who dominates and lords over its “domestic” adversaries. A cosmopolitan order perhaps, but not a democratically legitimated one.

The grim conclusion depends, of course, on taking Premise 1 (that any political entity presupposes an adversary) for granted. And this premise is the adversary claim that forms the second component of the stronger version of the particularist thesis. So if we expand the term “entity” to encompass collective identities, and if we assume that the adversary must be an external adversary, then Premise 1 itself turns into a strong and politicized version of the particularist thesis: a collective political identity requires not just an external other, but an external adversary. The implication is that my critique of the particularist thesis is politically irrelevant. This is the point of saying that humanity is not a political concept. By contrast, if we assume that the adversary can be an internal adversary, then Premise 1 yields the conclusion that any appeal to “humanity” as a collective identity can work politically only insofar as it serves to legitimate and mask the domination of one part of humanity over others in the false name of humanity itself. (This is how Schmitt suggested “humanity” is in fact used.)²⁸

A “solidarity” centered on such an identity could hardly play the required role in democratic legitimation, for it would be an oppressive, hierarchical identity serving to crush some who are at one and the same time humanity's “members” and lorded over by their “fellows.” We are talking about an imperialist sort of cosmopolitan “solidarity” à la the white man's burden.

So Premise 1 seems to imply either (A) that there can be no politically relevant collective identity at the global level—a political version of the particularist thesis—or (B) that such an identity would be terribly oppressive. Neither option allows for a democratically legitimated cosmopolitan political order based on a genuine sense of human solidarity.

The question is why we should grant Premise 1 in the first place. Of course if we define “the political” in particularist or adversarial terms as Schmitt does—if the political presupposes by definition a distinction between “friend and enemy”—then it is hardly surprising if “political” collective identities are necessarily particularistic or oppressive. But to glean a substantive argument requires that we abandon the tautological helps. Schmitt's stated motivation for defining the political in these terms is to provide a definition independent of the state: first, in order to avoid the circularity of defining the political and the state in terms of each other and, second, in order to acknowledge the possibly political dimension of society (Schmitt 1996, 20–22; cf. Balakrishnan 2000, 103). With Schmitt's goal in mind—to determine the relation among the political realm, modern state, and society—the political and the state should indeed not be defined in terms of each other; but with our goal in mind—to determine the relation among the political realm and the particularist thesis and the adversary claim—then the political should not be defined in terms of the particularist thesis or adversary claim. It is, as G. E. Moore might have said, always an “open question” whether the political is particularist or adversarial, and not something that can be settled by definitional fiat.

Schmitt's substantive argument for the decisiveness of the friend/enemy distinction, and thus for the adversary claim that every political collectivity must be constituted antagonistically in relation to an other (whether external or domestic), is Hobbesian: Schmitt (1996) appeals to the ineradicable possibility of existential struggle to the death, “the real possibility of physical killing” (33). This possibility of killing necessarily insinuates itself into the very nature of politically constituted collectivities: the possibility of war constitutes a collective political entity essentially as a combative collectivity, a “kämpfende Gesamtheit von Menschen.”

The question is why one possibility among others— and a possibility is always only one among others—should be determinative for political relations. Jacques Derrida (1994, 105–6) has argued that in seeing the possibility of an event as essentially constitutive of the political, Schmitt is forced to portray the possibility as an event itself. Schmitt's language constantly betrays this “slide from possibility to eventuality…and from eventuality to effectivity”:²⁹

It might be replied, on Schmitt's behalf, that not all possibilities are equal, that some possibilities do indeed structure reality, and that killing is one of these structuring possibilities. We can get a sense of the kind of argument required to transform mere possibility into effectivity by returning to the Hobbesian roots of Schmitt's argument. In order to show that the possibility of killing necessarily leads to war in the state of nature, Hobbes's ([1651] 1996) argument requires at least three premises. (a) First, Hobbes assumes that by nature human beings' overwhelming motive for action is self-preservation. (b) Second, Hobbes assumes the truth of epistemological skepticism vis-à-vis the external world. The consequent epistemological uncertainty implies that, in the state of nature, we simply cannot know whether the other whom we confront is benign or a threat to our survival (and the latter is always a possibility). (c) Third, Hobbes structurally defines the state of nature as a state where there is no overriding common sovereign power who can guarantee one's security against potential aggressors (through the enforcement of common rules). Hobbes's conclusion is that (a) the motive for self-preservation, combined with (b) our epistemological weakness and (c) structural lack of sovereign power, renders each of us “diffident” and thus prone to engage in preemptive strikes. This, then, is how the possibility of killing infallibly transforms itself into a reality for Hobbes.³⁰

The problem for Schmitt is twofold. First, the conclusion does not quite follow from the premises: when confronted with an other being, my epistemological weakness in fact faces three logical possibilities, not two. The other may be benign, or a threat to my survival, but for all I know the other might also be a willing and cooperative aid to my survival—indeed, his or her benevolence may turn out to be indispensable to it. So it is not clear why epistemological uncertainty should not lead to hope, rather than diffidence. This is presumably why Hobbes adds another assumption about human nature to get humans on a war footing: he needs to assert that (d) by nature humans seek glory. The second problem for Schmitt is that once we abandon any of the four premises, Hobbes's conclusion does not follow. To say nothing of Hobbes's assumptions about human nature, and the skepticism that Schmitt does not share, the structural assumption (c) that Hobbes needs to make demonstrates the way in which the effectiveness or realization of a possibility is dependent on contingent features of empirical reality.

In other words, in order to construe the possibility of killing as an eventuality, if not inevitability, we need to engage in two reifications. First, we must explain human behavior in terms of question-begging assertions about human nature. Such assertions, if themselves based on observations of human behavior or psychology, are question-begging because, as Rousseau famously pointed out in his second Discours, the characteristics we attribute to human nature may be the effect, not the cause, of sociopolitical phenomena.³¹

Second, we would need to take for granted precisely those contingent structural features of reality that should be the object of empirical investigation. By transforming a contingent, empirical possibility into a metaphysically determinative necessity, Schmitt has simply reified the contingent causes of antagonism. Not only does such reification illegitimately constrain normative theory by confronting it with fake metaphysical barriers, it also hampers the development of realistic empirical theories seeking to explain the contingent causes of antagonism and war. The question of when and under what circumstances the possibility of existential struggle is transformed into actual combat, or even relations of antagonism, is an empirical one.³²

MOUFFE: THE DERRIDEAN DIFFERENCE ARGUMENT

Chantal Mouffe provides a more promising argument for the political version of the particularist thesis and adversary claim espoused by Schmitt. She argues that (a) “any social objectivity is constituted through acts of power” and (b) “has to show the traces of the acts of exclusion which govern its constitution” (Mouffe 2000, 21). The point applies to a collective identity insofar as it has “objective” social existence, i.e., insofar as it manifests itself in social effects. Elsewhere Mouffe makes what is apparently a related point: “[a′] Politics aims at the creation of unity in a context of conflict and diversity; [b′] it is always concerned with the creation of an ‘us’ by the determination of a ‘them’” (101). The point seems to be that (a′) in the face of conflict, (a) collective identities are able to unify individuals only via the exercise of power. And (b) this exercise of power itself necessarily does “violence” to (some?) individuals, for such power can never be fully subject to rational justification. As such, from any given forged unity a remainder of exclusionary violence inevitably escapes: “the impossibility of establishing a rational consensus without exclusion” (45) follows from the fact that “difference is the condition of the possibility of constituting unity and totality at the same time that it provides their essential limits” (33). The upshot is, apparently, that (b′) every collective identity simultaneously and necessarily excludes some individuals from its constitution.

I find the points that Mouffe makes in (a), (b), and (a′) quite plausible. The problem lies with (b′), the upshot. I see two ways of reading this upshot.

(A) The first way assumes that the “them” required in the construction of “us” refers (1) to actually existing particular individuals who are (2) denied membership. It assumes, in other words, that Mouffe is a neo-Schmittian seeking to defend Schmitt's version of the particularist thesis.

That Mouffe does indeed subscribe to Schmitt's position here is clear. She explicitly follows Schmitt in claiming that antagonism “is inherent to all human relations,” that this antagonism is the defining feature of “the political” (Mouffe 2000, 101), and that the process of collective identity formation can only be understood in light of this fact:

While here Mouffe (1993) suggests that the enemy-relation is a possibility, elsewhere she claims that it is in fact a necessity: “to construct the ‘we’ it must be distinguished from the ‘them’, and that means establishing a frontier, defining an ‘enemy’” (69). Mouffe's argument, in other words, reproduces precisely the same equivocation we have already seen in Schmitt. And lest there be any doubt that the “enemy” whom Mouffe considers necessary for collective identity formation refers to concrete, actually existing human beings, she notes that “it is useful to remember with Carl Schmitt that the defining feature of politics is struggle and that ‘There always are concrete human groupings which fight other concrete human groups…’” (113; quoting Schmitt 1985b, 67). She does, however, make one modification to Schmitt's use of the term “enemy.” She distinguishes between two kinds of enemy: an enemy proper, whom one antagonistically seeks to destroy, and what she calls an “adversary,” which is an enemy whom one agonistically combats but whose existence is deemed legitimate (Mouffe 1993, 4). The fact that an enemy can take the form of a legitimate adversary makes democracy possible; but the fact that it is “impossible” to form collective identity except in relation to a concrete other qua enemy makes cosmopolitan democracy impossible: Schmitt “is right to say that a political democracy cannot be based on the generality of all mankind, and that it must belong to a specific people” (Mouffe 2000, 40).

If (b′) is read in this light, then Mouffe is advancing the Schmittian particularist thesis: the enemy necessary for collective identity formation must be an external other consisting of concrete individuals excluded from membership. This would explain why democracy belongs, as Mouffe says, to a “specific”–i.e., particular— people.³³

If this is the correct reading, then my point is simply that conclusion (b′) does not follow from Mouffe's argument. Some might, for example, be sympathetic to (a), (b), and (a′) for the reasons that Paul Ricoeur gives for what he calls the “tragedy of all human action”: beginning with the pluralist premise that there exist a plurality of incommensurable ends—which implies that “the historical realization of one value cannot be secured without harm to some other value”—he concludes that “in action, one must choose, hence prioritize, hence exclude.” The upshot, for Ricoeur, is that every constitution involves exclusion, which consists in the compromise of some binding values in order to secure others.³⁴

Now, one might argue in defense of Mouffe that “exclusion,” whether of the value-conflict sort Ricoeur has in mind or not, will invariably be targeted at particular individuals. And this may be true; but exclusion need not consist in the denial of membership to particular individuals, which is what the particularist thesis claims. The exclusionary violence to which Mouffe refers in (b) might, for example, take the form of including the targeted individuals as members by forcibly excluding (i.e., eliminating, suppressing, privatizing, reeducating, etc.) characteristics that fail to fit into the mold—the way that non–Parisians in France were included and assimilated into the French nation, for instance (Weber 1976). (Yes this occurred against the backdrop of “Spaniards” and others who were not members; my point is simply to illustrate what I mean by “exclusion” in nonmembership terms.) The Schmittian particularist thesis does not follow from Mouffe's argument.

(B) Of course it might be replied that if every forging of a unity inevitably produces exclusions, individual targets of exclusions are effectively denied full membership, even if formally included: some ostensible “members” invariably must be incarcerated, socially marginalized, or perhaps simply misrecognized. This glosses Schmitt's Premise 1 (that every political entity presupposes an enemy) in terms of domestic oppression, and it provides the second way of reading the upshot (b′) of Mouffe's argument. While the first reading assumed that the enemy Mouffe deems necessary for collective identity formation must be an external other consisting of concrete individuals excluded from membership, the second reading assumes that the enemy can also be an enemy internal to the collective itself. We have already sighted this spectre of a domestic enemy in Schmitt's work; on this reading, collective identity presupposes an other, but it is an other in the expansive sense of the term.

What is common to both the strict and the expansive readings of Schmitt and Mouffe, however, is that the enemy is supposed to consist in an empirically specifiable group(s) of actually existing concrete individuals. And therein lies the problem. We can begin to see the problem by turning to Mouffe's more explicit argument for the adversary claim and the necessity of the other-relation (whether in the strict or expansive sense). Mouffe seeks to defend her Schmittian conclusions by appeal to an intellectual tradition whose point of departure is not Schmitt, but Derrida's engagement with Saussure and Nietzsche: she appeals to the related Derridean categories of “difference” and the “constitutive outside.”

Derrida's (1967, chap. 2) notion of the constitutive outside emerges in the context of his critique of structuralist linguistics. Mouffe has sought to generalize Derrida's argument about language to identity formation, and she is not the only one to read Derrida in this way (e.g., Wenman 2003, 60; cf. Laclau 1996). What is particular to Mouffe is the argument that “the notion of the ‘constitutive outside’, borrowed from Derrida, can help us elucidate” her explicitly Schmittian theses. “One of Derrida's central ideas,” Mouffe (1993) claims,

Thus the notion of the constitutive outside helps explicate The constitutive outside is supposed to map onto the “them” that Mouffe thinks is constitutively necessary for a collective “we,” which in turn supposedly maps onto the Schmittian friend/enemy relation. The move from Derrida to Schmitt requires four steps: (1) identity presupposes difference, which means that difference is constitutive of identity; (2) therefore, any identity constitutively presupposes an other different from it; (3) this relation to an identity's constitutive other is inherently characterized by antagonism; and (4) in the case of collective identity formation, the categories of “identity” and “other” map onto sets of actually existing concrete individuals, viz., friend and enemy. I have already discussed Derrida's critique of the conflation of possibility and eventuality (and the reification of the causes of antagonism) involved in the Schmittian adversary claim contained in step (3). Here I focus on step (4), which is wholly incompatible with the Derridean argument about difference.

To see this, we need to attend to the specifics of Derrida's argument. Derrida's (1967, chap. 2) well-known thesis is that the “inside” of a supposedly closed conventional linguistic structure—what in Saussure's structuralist linguistics is called a langue—presupposes an “outside” that both is necessary for the closure of the linguistic structure and makes closure impossible. For Saussure (1960), a sign consists in the relation between a concept (signifié) and an acoustic image (signifiant). Saussure argued that the relation between a signifié and a signifiant with which it is associated in a linguistic structure is arbitrary and is wholly determined (as a matter of convention) by the sign's position within the totality of the linguistic structure. Meaning, therefore, does not derive from the relation of an atomistic sign to a nonlinguistic entity (such as a mental state or external object); rather, a sign derives its meaning solely from its relation to (and difference from) all the other signs within the totality of the linguistic structure itself.³⁵

For structuralist linguistics, then, the wholeness of a langue, and its wholly closed character, is necessary to explain both how the sign relation can be arbitrary and how the communication of linguistic meaning is nonetheless possible. Derrida (1967) has argued that this closure of the linguistic system, on which structuralist meaning depends, conceptually and so constitutively requires drawing a boundary around the langue, i.e., the delimitation of what is interior from what is exterior to it (49–53, 65–66). The problem is, according to Derrida, that such closure, and the determination of an outside— both of which are necessary for meaning—would also make meaning impossible. Within such a closed system, since each sign must be defined in relation to the totality of other signs within the same linguistic structure, its meaning is subject to a vicious circularity: if the meaning of sign a is given by its difference from b (etc.), and the meaning of b (etc.) is in turn given by its difference from a, then the meaning of a is ultimately dependent on the meaning of a itself, which is precisely what is in question. While the location of each sign within the structure of relations is fixed, its meaning would remain empty of content because ultimately self-referential.³⁶ Hence the need to refer to something outside of the linguistic structure. Thus the “outside” of a linguistic structure is constitutive of a langue in two (contradictory) ways: first, this outside is required to constitute the boundaries that give closure to the totality of the linguistic structure; second, this outside must penetrate that structure in order for meaning to be possible in the face of a vicious, self-referential circle. This requires the signification of something outside of the putatively closed system of signification. That what is supposedly “outside” is actually constitutive of the “inside” in this way fatally calls into question, according to the Derrida, the distinction between inside and outside, on which structuralist linguistics depends.³⁷

The upshot is that, even if Derrida's argument about language maps onto an argument about collective identity formation—even if collective identity constitutively presupposes difference—the Derridean category of the constitutive outside is not coterminous with an empirically specifiable set of concrete individuals. Even granting the Derridean argument, and even if “difference is the condition of the possibility of constituting unity,” as Mouffe (2000, 33) urges, it does not follow that collective political identities presuppose—or that unity can only be forged on the basis of—an “other” consisting of actually existing individuals. Derrida's argument does not even support the weaker version of the particularist thesis, much less the stronger Schmittian one that adds the adversary claim. To derive Mouffe's conclusion to the contrary is to confuse the Derridean language of difference with the Schmittian language of alterity. From the Derridean insight it does not follow that every “us” requires an actually existing “them” in Mouffe's sense.

So it is no surprise that Derrida himself attacks Schmitt's friend/enemy argument on precisely this issue. The slide in Schmitt's argument, from the possibility of the enemy-relation to its putative necessity, is important for Derrida not only because its exposure undermines Schmitt's defence of the adversary claim (step (3) above), but also because to see the possibility of the event of war as an event itself requires a prior identification of who the enemy is, i.e., it presupposes that the enemy consists in an identifiable, recognizable, concrete set of individuals, a kämpfende Gesamtheit von Menschen. As Derrida (1994) puts it, it is absolutely crucial for Schmitt that there be no doubts plaguing the knowledge of “who is the friend and who is the enemy” (136). And this takes us to the divided heart of the matter as Derrida sees it: the difference that constitutes identity constitutes identity and so penetrates it so that, ultimately, not only is difference internal to the identity itself, but the distinction between what is internal to the identity and what is external to it is always breaking down.³⁸

And if the “constitutive outside” is constitutive of what is inside—if even conceptually there is no pure “inside”—then the (impure) conceptual distinction between friend and enemy (or between “us” and “them”) cannot map onto existing, empirically identifiable bounded groups of individuals except through ideological obfuscation. This is because any act of identifying the “friend” will inevitably find within its members characteristics that supposedly belonged solely to the category of the “enemy.”³⁹ “The contrary of the political inhabits and politicizes the political” (160). On Derrida's deconstructive reading, then, the enemy is “other” only in the expansive sense that allows for the possibility of an other who inhabits the inside. Indeed, ultimately the enemy is “other” only in the radically expansive sense that breaks down the distinction between self and other: the enemy is within one's own self, which is, as Derrida points out, the conclusion to which Schmitt himself comes as a prisoner after World War II (187). This is why, Derrida claims, Schmitt begins by strictly defining the political in terms of antagonism against an external enemy, but ends up internalizing the friend–enemy relation, claiming that the lack of an external enemy would simply internalize war by directing it toward domestic enemies (140–42).⁴⁰

One need not endorse either Derrida's reading of Schmitt or his substantive philosophical argument about difference to see that his argument cannot be used to justify either the exclusion or the adversary claim. There are three distinct reasons why difference does not map onto an empirically identifiable concrete “other” consisting of actually existing persons excluded (whether formally or effectively) from membership.⁴¹

(1) First, even if identity presupposes difference, difference can exist within the putative inside. (2) Second, difference within the putative inside actually calls into question the sharp distinction between inside and outside. (3) Third, difference need not refer to actually existing persons at all: it can refer to nonhumans, it can refer to characteristics rather than individuals, and it need not refer to actually existing things at all. And the difference argument does not justify the adversary claim because (4) there are no metaphysical grounds for thinking that a relation of difference is necessarily a relation of antagonism (i.e., for thinking that the possibility of war is always effective and structuring).

The result is that the difference argument cannot service Mouffe's neo-Schmittian anticosmopolitanism, on either the strict or the expansive reading: either she means to say that (A) democracy requires an external other qua enemy, in which case she has provided no support for the particularist thesis on which that claim is predicated, or she means to say that (B) democracy requires an enemy whether external or domestic, in which case her point is not specific to global democracy—to say nothing of the fact that the Schmittian defense of this adversary claim conflates possibility and effectivity. The problem of power—in particular, the exclusionary violence that the constitution of any unity leaves as a remainder—is not a problem specific to cosmopolitan democracy. If one can conceive of democratic possibilities at the subglobal level—as Mouffe does, which is the point of her critique of Schmitt— then to bar the possibility of a global identity undergirding a democratically legitimated cosmopolitan order requires asserting the particularist thesis found in the strict reading (A). And the problem here is that— difference argument or not—the particularist thesis does not stand to reason: cosmopolitan solidarity or identity does not require an actually existing external other in contrast to which it must constitute itself.