1 Introduction

This paper offers a realistFootnote 1 account of political obligation (PO). More precisely, it offers an account of PO that belongs to the Williamsian liberal strain of contemporary realist theory (as opposed to a Geussian radical realist strain) and will draw on and expand some ideas familiar from Bernard Williams’s oeuvre (namely, the distinctions between thick and thin ethical concepts, political realism and moralism, as well as a minimal normative threshold for distinctively political form of rule – see Williams 1985, 2002, 2005).

Accordingly, the paper will claim that the fact of membership in a polity provides people with sufficient reason for complying with those political authority claims whose source is that particular polity. The paper will explain that membership in a polity is constituted not by a communitarian identification (Horton 1992), nationhood (Tamir 1993), joined commitment and plural subjectivity (Gilbert 2006), or non-instrumental valuation of membership (Scheffler 2018), but by the social fact that a group of people is stably exposed to political authority claims associated with a particular polity which they find making sense as passing the minimal normative threshold (the correlative requirements of a Basic Legitimation Demand and a Basic Obliging Demand) for a distinctively political form of rule (as opposed to the rule of terror or sheer coercion).

A better-known name for this political-ethicalFootnote 2 phenomenon is PO but the paper will argue that the label ‘PO’ is in fact a misnomer because it does not refer to a generic obligation. If PO were indeed a generic obligation as theories of PO usually assume, it would provide an independent (moral or, in some accounts, non-moralized) reason for compliance that would need justification in terms of general (moral) principles, local practices or ethically significant facts. And this is exactly what most accounts of PO look for. But, based on the Williamsian distinction between thin and thick ethical concepts (Williams 1985: 140), this paper will argue that PO in fact is not a thin ethical concept (which generic obligations typically are) but merely a conceptual component of the thick ethical concept of membership in a polity. Thick ethical concepts typically work the following way: they fuse empirical content (world-guidedness) with reasons for action (action-guidance). In this case, the thick ethical concept of membership in a polity fuses a world-guided component (the fact of membership) with a reason-giving component (PO).

The paper will draw on or result in some nontrivial claims about PO and its connection to contemporary realist political theory. First, it will claim that realism and an affirmative answer to the question of whether people have PO are compatible with each other. Second, it will claim that some forms of realism (especially what the paper will call Williamsian liberal realism) are compatible with not only a strictly non-moralized account of PO but also with the attribution of some ethical significance to PO (even though it does not deny the possibility that other, Geussian radical forms would be more compatible with non-moralized accounts). Third, it will deny the conceptual independence of PO but still it will not merely repeat the so-called Conceptual Argument. Fourth, the paper will take the so-called correlativity thesis simply as a precondition of any satisfactory (mostly but not necessarily always affirmative) account of PO. Fifth, the paper will find existing accounts of PO unsatisfactory mainly on two grounds: either because of their so-called moralism or because of not finding any convincing alternative to a moralistic account. Sixth, the paper will argue that a Williamsian realist account of PO is possible and has some advantages but it also needs a somewhat enlarged view of what Williamsian realism can say about the relationship between politics and ethics, and the most fundamental demands of a distinctively political rule. All of these nontrivial claims will be addressed in the paper in due course.

Accordingly, the paper will be structured as follows: first, the criteria of a satisfactory theory of PO will be outlined; second, the difficulties of proposing a realist account of PO will be explained; third, the main elements of a Williamsian realist account of PO will be presented.

2 The Received View

This section will outline the criteria of a satisfactory account of PO in order to lay the groundwork for the argument of the later sections that, despite some obvious difficulties, a realist account of PO can meet these criteria. To begin with, the paradigmatic account of PO is still the actual consent theory (see Simmons 1979; Pateman 1985; Klosko 2005). As Klosko (2005: 10) put it, “it seems to me, the relevant criteria can be identified by looking at a theory of political obligation based on consent. […] The current skepticism about political obligations in large part stems from awareness that most citizens have not expressly consented to government and cannot be shown to have performed other actions that constitute tacit consent. […] But if some other theory could ground obligations analogous to those based on consent, it too would solve the problem of political obligation.” According to this paradigmatic account of PO the members of a polity are given sufficient, independent (moral) reason for complying with political authority claims associated with that polity as long as their previous consent makes these claims legitimate.

There are at least four important background assumptions of this paradigmatic account that are still shaping the debates on PO. The paradigmatic account assumes: first, that the source of the normative force is some general moral principle, a local practice, an ethically significant fact or some other norm-generating mechanism (the justification thesis); second, that the source of the normative force of PO is conceptually (linguistically, grammatically etc.) independent from legitimacy (the conceptual independence thesis); third, that there is still a close (e.g. Simmons (1979) calls it a logical as opposed to grammatical) connection between the legitimacy of political authority as a claim-right and PO as a duty to obey (the correlativity thesis); fourth, that the source of the normative force of PO is independent from other moral reasons for action. In other words, people comply with a political authority claim simply because of the special status of the authority claim and not because they have some other good moral reason to do so (the content-independence thesis). Although the actual consent theory became almost universally rejected (with some notable and partial exceptions like Beran 1987; Gilbert 2006; Simmons 1993), most contemporary accounts of PO still rely on some version of these background assumptions or at least subscribe to most of them. In other words, fair play theories (Dagger 1997; Rawls 1964), associative accounts (Dworkin 1986; Hardimon 1994; Horton 2006, 2007; Horton and Windeknecht 2015), natural duty theories (Rawls 1971; Wellmann 2005) etc. still largely follow the logic of the paradigmatic account of PO.

What happens when one questions one of these background assumptions? It largely depends on the nature of the criticism. There are, for instance, disagreements on the justification thesis. Most accounts seek a clear moral ground for PO in some general moral principle (fair play, natural duties, gratitude, consent, common good etc.) in order to explain why people generally have PO. Some philosophical anarchists, most notably Simmons (1979) also argue that there are clear criteria of how to justify PO but in practice these criteria are not and most probably cannot be met, therefore most people in most polities for most of the time have no PO (for a criticism: Szűcs 2022). On a somewhat different note, Gilbert’s (2006: 156 − 16) non-moralized account claims that people generally do have PO but it is an obligation of joint commitment and as such it is not a moral requirement. Associative theorists occupy a middle-ground between the moralized and non-moralized accounts when arguing that the source of the normative force of PO is some local practice or an ethically significant fact. As we will later see, it is the most common criticism against them that they lack a clear moral grounding.

Others question the conceptual independence thesis. The so-called Conceptual Argument (CA) (MacDonald 1951; Pitkin 1966) argues that it makes no sense to look for the sources of the normative force of PO because PO is already conceptually (linguistically, grammatically etc.) connected to and dependent on the legitimacy of political authority. This is a very radical claim. If CA is right then any attempt to provide an independent normative account of PO is a misguided and unnecessary enterprise which is doomed to fail from the beginning. Unsurprisingly, however, CA has been widely dismissed as begging the question by reducing normative disagreements to conceptual issues (e.g. Greenawalt 1987; Mokrosinska 2012; Pateman 1973; Simmons 1979). Most recently Applbaum (2018) has argued that the CA fails because it confuses two levels of debates: the disagreements on concepts and conceptions. If this objection to CA is right then any argument that asserts the conceptual dependence of PO on other concepts is also doomed to fail like CA unless it can explain what is the upshot of a normative account of PO.

Others focus their critique on the correlativity thesis. Recall that the correlativity thesis (e.g. Mokrosinska 2012: 2–4; 33–38; Simmons 1979: 195–197) assumes a logical connection between PO and political authority - usually described as a duty of compliance and a claim-right for issuing binding decisions - while denying an even closer, conceptual (linguistic, grammatical etc.) connection between them. Applbaum (2018), in his critique of the correlativity thesis, argues that political authority is not a claim-right but power and, accordingly, its correlative is not a duty to obey but a liability to the exercise of power. Edmundson (1998) argues that the actual correlative of the claim-right of political authority is not a duty to obey but a duty of non-interference. Later he (2006) also argues that political authority is not claim-right and its correlative is not a duty to obey but law-abidance as a virtue. If they are right these criticisms require if not the elimination of any normative account of PO but a fundamentally different conceptualization (with considerably different implications) of the same political-ethical phenomenon. Remarkably, however, they still maintain a central element of the original correlativity thesis, namely, a symmetry between the legitimacy of political authority and some moral requirements on the part of the subjects of political authority.

Finally, there are multiple criticisms of the content-independence thesis. One version of it plays a central role in the philosophical anarchist rejection of PO: both Simmons (1979, 1993, 2005) and Huemer (2013) argue that we do not need PO because the special status of a political authority claim is not enough to provide us with sufficient reason for compliance unless we have some other good moral reasons to do so. Similarly, Raz (1979) argues that compliance with political authority claims cannot and should not have a content-independent moral justification: if the content of a political authority claim is justifiable then it by itself provides us with sufficient reason for compliance. In contrast, some (e.g. Klosko 2011) argue that reasons for compliance can come from multiple independent sources and their normative force is limited in scope but this is not fatal to PO theory.

It is an important question whether and to what extent these four background assumptions inherited from the paradigmatic, consent-based account of PO are necessary for a satisfactory account of PO. The characteristics of disagreements on these four points seem to highlight two lessons. First, PO theory is still not over these background assumptions and, accordingly, one should reasonably expect that any satisfactory account of PO would retain some commensurability with these background assumptions. Second, even though PO theory is not entirely over these assumptions, there is ample room left for different opinions without denying the existence of PO.

Given all this as well as the basic characteristics of the account of PO already outlined in the introduction, the main challenge of the remaining part of the paper will be as follows: if it wants to retain some connection to and comparability with other accounts of PO then it has to show: first, that it is possible and useful to offer a normative account of PO in realist terms that explains the basic characteristics of the same political-ethical phenomenon as other accounts seek to explain even if this account will deliberately seek to depart from what it sees as unacceptable moralism in other accounts (most notably, that PO would be a generic obligation); second, that it is possible to argue for the conceptual dependence of PO and thereby to deny that PO is a generic obligation without accepting the conclusion of CA that an independent account of PO makes no sense; third, that it is possible to deny that PO is a generic obligation but maintain that there is some kind of connection between the legitimacy of political authority and PO; fourth, that the source of the normative force of PO has some content-independence even if PO is not a generic obligation.

3 The Realist Challenge

This section will address the special difficulties of offering a realist account of PO.

Let us call the substantive desideratum the requirement that any realist account of PO should be compatible with the general political outlook of political realism. There are many lists on the market enumerating the criteria of what constitutes the distinctively realist political outlook in political theory (e.g. Galston 2010; McQueen 2018; Sleat 2013). From the perspective of this paper, the single most important issue mentioned in these lists is the dividing line between realism and what realists see moralism (‘applied morality’, ‘the ethics-first view’).

Given the special importance of the realism/moralism divide for the argument of this paper, let us focus on what this divide means in realism. All realists seem to agree that moralism is a view of politics that seeks to reduce the normative challenges raised by politics to questions of morality and thinks that these questions can be solved in terms of a coherent ethical theory grounded in general moral principles. However, and this is a source of many misunderstandings about realism (e.g. Erman and Möller 2015), the realist critique might take two completely different forms (see Rossi and Sleat 2014): Geussian radical realism (as well as some other version of realism) claims that moralism is a fundamentally wrong way of looking at the normative challenges raised by politics because these normative challenges can be best understood in terms other norm-generating mechanisms (epistemic, instrumental) than morality (e.g. Burelli and Destri 2022; Kreuz and Rossi 2022). In contrast, Williamsian liberal realism claims that moralism (in which the concept of obligation plays a highly problematic but central role: ‘Morality is distinguished by the special notion of obligation it uses, and by the significance it gives to it.’ (Williams 1985: 174)) is a fundamentally wrong way of looking at the normative challenges raised by politics because ‘morality’ is a peculiar and misleading understanding of norm-generating mechanisms, ethics included (e.g. Hall and Sleat 2017; Sleat 2022). In other words, Geussian radical realism is based on a separation of political and moral norms, Williamsian liberal realism is based on a conception of a complex political-ethical realm whose normative questions are primarily endogenous to politics and, at the same time, may or may not carry some ethical significance.

Seen in this light, the fundamental problem with PO theory is that most accounts of PO apply the logic of moralism in very transparent and easily identifiable form. Most of these accounts interpret the justification thesis in a way that requires to ground PO in a single general moral principle and to provide a moral justification for it in terms of a coherent ethical theory based on that principle. Accordingly, these accounts are clearly unacceptable for a realist as normative explanations for having PO. Not just because they do not fit into some boxes but because, on a realist view, they fail to provide an adequate account of politics.

Two exceptions seem to be to this general observation. One is the idea that PO should be understood as a non-moralized concept. Margaret Gilbert’s theory (2006) based on joint commitment and plural subjectivity is the most elaborate version of this view. Her account of PO is all the more interesting because she categorically denies that PO would be an ethical phenomenon. From a moralist viewpoint, this view is clearly unacceptable because a non-moralized PO makes little if any sense. The other exception is the family of those associative theories that seek to ground political obligation in some local practice or an ethically significant fact (like membership). From a moralist viewpoint, such accounts are deeply unsatisfying because they do not seem to provide a clear moral ground for PO (for such criticism see Klosko 2005; Simmons 2000). John Horton’s associative theory is the most interesting example of this second view of PO. All the more so, because Horton is a representative of contemporary realist political theory, too.

There are two questions here. First, whether either Gilbert’s or Horton’s approach does provide a satisfactory account of PO from a realist viewpoint. Second, if not what we should look for instead.

To begin with, Gilbert’s account might be appealing for Geussian radical realists because a non-moralized concept of PO would fit nicely into an understanding of politics that looks for non-moralized norm-generating mechanisms in politics. It does not necessarily mean that Gilbert’s account should be acceptable for them: instead, this is a question that would need further investigation and the outcome of such an investigation would depend mostly on whether the notions of joint commitment and plural subjectivity are useful for radical realist theorizing. Since this paper is committed to another, Williamsian liberal version of realism, the question for us looks a bit different: is Gilbert’s (2006: 156–161) stipulation that the obligations of joint commitment (PO included) are not moral obligations appealing enough from a Williamsian realist viewpoint? Alas, neither the stipulation itself nor her argument for it (namely, that obligations of joint commitment are not context-sensitive) seem particularly persuasive from a Williamsian realist viewpoint. Why? Because even if we accept that PO is context-insensitive in some sense, it does not seem to follow that it could not carry some political-ethical significance for those who are under PO.

Horton’s account (1992) that finds the source of the normative force of PO in the norm-generating mechanism of identification with a polity is obviously much less appealing to radical realists than Gilbert’s exactly because it does not rest on a strict separation of moral and political norms. In contrast, Horton’s account seems to be more compatible with a Williamsian version of realism than hers because Horton’s allows the ethical significance of the fact of membership. There is a reason, however, why it still does not seem a satisfactory account of PO. The objection has been already stated by many critics of Horton and I do not see in his later responses (Horton 2006, 2007) a really convincing answer to this objection. The objection is quite simple: Horton’s central notion - identification - does not seem to be either a necessary or a sufficient condition for PO (Dagger 2000; Simmons 2000: 65–92; Vernon 2007; Wellman 1997; for a more sympathetic but critical reading see: Fives 2022): it is not a necessary condition because one’s membership can be a fact without any positive identification on one’s part (one can recognize the special standing of the issuer of a political authority claim over oneself and, at the same time, distance oneself from the community) and it is not a sufficient condition because one’s membership cannot be a fact based solely on one’s identification. Consequently, identification cannot be the source of the normative force of the fact of membership either. If this objection holds (and I think it does) then a Williamsian realist account of political has to look for another, more plausible source for the normative force of PO.

From a Williamsian realist viewpoint, therefore, neither Gilbert’s nor Horton’s account seem satisfactory enough. Two lessons can we learn from this for the remaining part of the paper. First, we need an account that accommodates the general political outlook of realism, especially in its Williamsian form (including the suspicion against generic obligations). Second, we need a better normative explanation for PO than identification. It will be the task of the next section to show how it is possible.

4 A Williamsian Realist Proposal

This section will unfold the major elements of a broadly Williamsian account of PO, explain to what extent it is compatible with the background assumptions of PO theory and why this account is also an extension of and arguably an improvement on Williamsian realism.Footnote 3

As we already saw, there are multiple obstacles to offering a Williamsian realist account of PO. First of all, it cannot offer a moralistic justification for having PO and the available alternatives (most notably, Gilbert’s and Horton’s) are not satisfying from this perspective either because they offer a non-moralized alternative conceptualization or because they focus on a no less problematic alternative source of normative force in terms of identification. Not unrelated to this difficulty, even if a Williamsian realist might think that the political-ethical phenomenon usually described as ‘PO’ is worth to explore, its traditional characterization as a generic obligation must seem highly suspicious. How is it possible to overcome these difficulties?

The answer lies in two important Williamsian ideas: first, the distinction between thick and thin ethical concepts and, second, the minimal normative threshold for a genuinely political form of rule. The first will help us find a way of conceptualizing the same political-ethical phenomenon as PO theory in general without resulting in either a moralistic or a non-moralized concept of PO. Alas, it will lead to another difficulty: namely, that it will have to question the conceptual independence thesis, so, in order to succeed, it will also have to explain the difference between the Conceptual Argument (CA) and a Williamsian conceptualization of PO. The second idea, the minimal normative threshold will provide a more fundamental and more robust normative explanation for having PO than those accounts that seek to ground PO in local ethical practices or ethically significant facts like nationhood, identification with a community, valuation etc.

4.1 The Thick Versus thin Ethical Concept Distinction

As regards the first part: how else can be the political-ethical phenomenon usually called ‘PO’ understood if not as a generic obligation? To answer this question, we should rely on the Williamsian distinction between thin and thick conceptsFootnote 4. A generic obligation is seen by Williams as a textbook example of thin concepts. Thin concepts contain minimal information about the matters of fact of the world around us and therefore they fit nicely into a moralist separation of ‘is’ and ‘ought’, ‘facts’ and ‘values’. Their normative force springs not from how things are in the world (from the ‘is’ or ‘facts’) but from their fit with other thin ethical concepts (from the ’ought’ or ‘values’). Hence the pressure for justifying PO in terms of general moral principles in moralist political theory. In contrast, thick ethical concepts bridge the gap between ‘is’ and ‘ought’, ‘facts’ and ‘values’ (Williams 1985: 129). They are both world-guided and action-guiding in a way thin ethical concepts are not (Williams 1985: 140–141). In other words, when a thick ethical concept is properly applied to someone then if a certain thing is a fact with respect to that person, then that person is given a reason to act in a particular way. As it must be clear from this, the reason-giving component of a thick ethical concept is closely dependent on a factual component and usually it is impossible to clearly separate them.

Based on this, the obvious alternative to seeing PO as a generic obligation is to offer a conceptualization of the same political-ethical phenomenon in terms of thick ethical concepts. But which concept can do that? Can be, in the first place, PO itself a thick ethical concept? Can we confidently say about having sufficient reason for compliance with political authority claims that it is both world-governed and action-guiding? This does not seem a very promising idea because something is conspicuously missing from the picture. Indeed, PO as a thick ethical concept could account for having a reason for compliance but by itself it does not seem to be able to properly point to the factual element that would provide the reason for compliance. Interestingly, the need for some factual element appears even in more traditional accounts of PO too. There it is usually described as the ‘particularity problem’ (Simmons 1979) and it is seen to raise an almost insurmountable challenge to PO theory because most proposed moral grounds cannot explain a particularized relationship between someone and a certain polity.

Hence the idea that probably the relevant thick ethical concept that could ground PO in a Williamsian realist manner is membership in a polity because it contains exactly that factual information that is needed to solve the particularity problem. How is it possible? The obvious (albeit undeniably radical) answer is that PO describes a political-ethical phenomenon that is not conceptually independent from membership in a polity but, quite the contrary, is the reason-giving conceptual component of membership while the mere fact of membership is the world-guided conceptual component of the same concept. Put simply, people have PO due to the fact that they are members in a particular polity.

So far so good. We have an alternative way of conceptualizing PO that is compatible with Williamsian realism and addresses Williamsian realism’s main substantive concerns about moralism. But it has an obvious difficulty: namely, it denies the conceptual independence of PO and thereby might seem vulnerable to the same objection as CA. Importantly, however, there is a fundamental difference between CA and this account. CA has sought to eliminate independent theorizing about PO as not making sense at all because of the conceptual dependence of PO on legitimate political authority. In contrast, this account seeks to draw attention to the largely neglected theoretical importance for realism of the political-ethical phenomenon usually described as PO. PO may not be a generic obligation but, as we will later see, it plays an important part in creating and maintaining a distinctively political form of rule that has not received proper attention in realist theory. Consequently, on this paper’s account, the rejection of the conceptual independence should not necessarily be fatal to PO theory.

Further, minor, difficulties with this account arise from two additional sources. The first difficulty is that this Williamsian realist account cannot produce a justification for PO that could be seen as satisfactory from a moralistic perspective. But this is a limitation that this account must happily live with and nothing else needs to be said about it. The second difficulty is that if this account admits that the political-ethical phenomenon in question is not a generic obligation then it might seem far from obvious why it is better to insist that we should describe it as a PO than following Applbaum’s (2018) suggestion that legitimacy’s correlate is liability or Edmundson’s (2006) proposal for law-abidance as a virtue. This is a more serious difficulty that is not that easy to answer but not impossible. The main reason for why the Williamsian realist account of this paper seeks to conceptualize the political-ethical phenomenon in question as PO not as liability or law-abidance is that Applbaum’s and Edmundson’s attacks on PO for its apparently too demanding nature do not seem particularly convincing. What makes the paradigmatic assumptions of PO too demanding in their eyes is the allegedly irrational possibility that people might be requited for not complying with political authority claims in cases when non-compliance does not seem to be warranted. The paradigmatic case for this is jaywalking across an empty road or not stopping at a stop sign in a desert at midnight. Both Applbaum and Edmundson are hesitant to accept that there could be a well-founded (for them: moral) obligation for not jaywalking or stopping in cases like these. But, as many others were keen to point out (Applbaum himself quotes Raz but Gilbert (2006) makes a similar point), compliance is not necessarily about doing exactly what is told us to do but acknowledging the special standing of the issuers of political authority claims and the special status of these claims. What one ought to do all things considered is a completely different question (see also Simmons 1979) and, therefore, the case against PO theory by Applbaum and Edmundson, despite their merits, simply does not seem convincing enough.

4.2 The Minimal Normative Threshold

Even after answering the above difficulties, there remains one further question that needs serious consideration: it is why and how membership in a polity can have its normative force. The challenge here is to offer a more robust alternative to Horton’s ‘identification with political community’ as the source of the normative force of membership in a polity (or to any other accounts following a similar logic). This challenge is three-pronged: first, it must be shown that there is a norm-generating mechanism that lends normative force to membership in a polity; second, it must be shown how this norm-generating mechanism is uniquely connected to membership in a polity (as well as to other polity-like political-ethical phenomena) and in what sense it offers a better normative explanation for PO than other mechanisms; third, some explanation must be offered for the difference between membership in a polity and other political-ethical phenomena that display a family resemblance with membership in a polity (i.e. membership in polity-like structures like mafias or bands of robbers or evil states) because the disturbing similarities between them is considered to be an Achilles heel for membership accounts in general.

To start with, Williams (2005: 1–3) claims that the norms of politics cannot be simply reduced to the norms of morality either in a utilitarian or a deontological form because, as he argues, there is an autonomous source of political-ethical norms which is the demand to create stable political order, let people live in peace, and cooperate with each other. This is called by him the ‘first political question’ and he argues that among the various responses to the first political question, we can distinguish those that offer a genuinely political answer from those that simply rely on domination or sheer coercion. According to Williams, the basis for this distinction is that a genuinely political answer must offer some explanation to the subjects of political rule and that explanation must ‘make sense’ for them in a given historical context. Here lies the source of the normative force of the requirements of genuinely political rule which Williams calls the Basic Legitimation Demand (Williams 2005: 1–17) but is in fact a minimal normative threshold for a distinctively political form of rule. The normative difference between genuinely political answers to the first question of politics and examples of sheer coercion or successful domination is that the formers pass the minimal normative threshold, the latter do not.Footnote 5

Is this minimal normative threshold moral or non-moral? The question is wrong. As Williams himself somewhat vaguely puts it, the first question of politics may be moral but definitely not pre-political.Footnote 6 To make this seemingly obscure Williamsian claim a bit more accessible, we need to take two things into account. First, the autonomy of the norms of politics is not based on a strict separation of moral and non-moral (Hall and Sleat 2017; Sleat 2022). Second, a strict separation of moral and non-moral is hardly tenable even on a general level (Williams 1985: 7). Consequently, it is best to conceive of the minimal normative threshold for genuinely political rule as constituting a broader political-ethical domain within which normative demands include but are not reducible to moral considerations.

Hence the first prong of the Williamsian realist answer to the question about the source of the normative force of membership in a polity. Undeniably, membership in a polity is a specific kind of political relationship (just one among many different kinds) but, at the bottom, the source of its normative force springs from the fact that the political authority claims that constitute it pass the minimal normative threshold, like in any other kinds of political relationship. In other words, a Williamsian realist account does not need a moralistic (or else, external moral) justification for PO because it can offer a different kind of normative explanation: people have a sufficient reason for compliance with political authority claims because these claims pass the minimal normative threshold. In turn, when these claims do not pass the threshold, people do not have sufficient reason for compliance. In such cases, people still might be coerced into obedience or they might be trapped into a dominating relationship but these are qualitatively different from being part of the kind of political relationship on which membership in a polity is built. In other words, this account of PO is normative in the theoretically important sense that it makes it possible to tell the difference between coerced obedience to red-handed tyrants, foreign occupiers and other forms of rule by domination and sheer coercion on the one hand and compliance with political authority claims on the other. No doubt, the minimal normative threshold puts the bar very lowFootnote 7 with respect to what counts as a sufficient reason for compliance. But is this really a problem? First, there is no reason to think that we cannot have some other - more rigorous, more demanding - political-ethical considerations in addition to what PO gave us reason to do. Second, there is no reason to think that we cannot have some more rigorous, more demanding political-ethical considerations that, in case of conflict between them and the content of the PO, will weigh more than our PO. Neither of these possibilities would make the requirement of passing the minimal normative threshold irrelevant.

The second prong of the Williamsian realist account is the explanation for the unique connection between the minimal normative threshold and membership in a polity (or, more broadly, a polity-like political-ethical phenomenon). In contrast to other membership accounts that offer identification with a political community or nationhood or joint commitment etc. as the norm-generating mechanism for membership in a polity, this account claims that membership in a polity is constituted by being stably exposed to political authority claims associated with a single polity and by finding these political authority claims as making sense in terms of the criteria of the minimal normative threshold. The basic idea is that it is neither a necessary nor a sufficient condition for this special relationship that someone identify with a political community or consider oneself to belong to the same nation as one’s compatriots etc. while being stably exposed to political authority claims and finding them to make sense in terms of the minimal normative threshold are both necessary and sufficient conditions for actual membership in a polity. Membership in a polity is typically not self-assumed (i.e. people are typically born into their membership and sometimes it is imposed on them). It does not require positive emotions towards one’s polity. One’s identity does not have to be primarily determined by one’s membership in a polity. People belonging to ethnic or religious minorities, holding unpopular political views etc. are still rightly considered to be members in polities until the political bonds between the polity and its members is dissolved by the collapse of political rule. Importantly, membership in a polity is a much broader notion than modern equal citizenship. It, of course, includes birthright-based as well as naturalized citizenship, but it also includes various forms of resident aliens, temporary visitors, guest workers, asylum seekers and even illegal aliens insofar as they are sustainedly treated by political rather than purely coercive means. Why? Because all these forms of political relationships stably bind people to a particular polity, providing them with sufficient reason for compliance. This understanding of membership might seem overly inclusive, but only if we take equal citizenship as our benchmark. Historically speaking, however, equal citizenship has hardly ever been the predominant form of membership in a polity and it is far from obvious that equal citizenship for all is the future of polities. A realist account of PO should be able to address this diverse reality and explain the normative implications of the empirical variety of forms of membership in a polity.

Some might wonder (as one of the reviewers of this paper did) how all this normative work can be done by the minimal normative threshold alone. Is Williams’s original Basic Legitimation Demand (BLD) not merely a top-down relationship between the exercisers of political power and their subjects? In other words, is this relationship not too thin to constitute a polity-like political-ethical phenomenon with identifiable membership? This is an important question that needs to be addressed properly. All the more so because there is a reason why this paper uses the term ‘the minimal normative threshold’ instead of using Williams’s original terminology. The original Williamsian formulation of the minimal normative threshold was unnecessarily narrow as it merely focused on how using coercive measures by those who have political power can be legitimized. If the minimal normative threshold were only about the conditions for legitimately using coercive measures then the political relationship constituted by it would be indeed a very thin top-down relationship (or, to quote Williams, it would be merely about being ‘in the power of the state’(Williams 2005: 4)). But, in fact, what Williams, on Hobbesian grounds, identifies as the first question of politics: “the securing of order, protection, safety, trust, and the conditions of cooperation” can be realized in multiple ways other than using coercive measures. Driving on the proper side of the road, not jaywalking, paying taxes, voting, obeying the laws, defending the country, supporting major public institutions etc. demand that people themselves act in one way rather than another; and the Hobbesian presumption that in cases like these people are either coerced into obedience or act only for fear seems to be an extremely implausible explanation. Just like the BLD cannot be met (as Williams wants us to believe) unless people are given some intelligible explanation for why they should face coercive measures and in what exactly these measures differ from sheer coercion or successful domination, people cannot be demanded to act on their own in these cases unless they are given intelligible explanation for why do should do so. If BLD is the name for the normative threshold for the distinctively political use of coercive measures, then maybe we should call ‘Basic Obliging Demand’ the normative threshold for the distinctively political cases when people act on their own. But the name does not really matter. What matters instead is the fact that the minimal normative threshold is not merely about a thin top-down relationship constituted by exposing the subjects of political authority claims to legitimate coercive measures but also necessarily includes a number of bottom-up relationships constituted by the fact that people are given sufficient reason for acting on their own in accordance with political authority claims. Taken together, these two kinds of political relationships are already thick enough to create a polity-like political-ethical phenomenon with identifiable membership.

Finally, the third prong of the Williamsian realist account is the distinction between membership in a polity and the variety of neighboring concepts that denote some other polity-like political-ethical phenomena. Williams himself does not make the job too easy because his original formulation of the minimal normative threshold is unnecessarily statist in the sense that it limits its scope to states - as it was pointed out and rightly criticized by Raekstad (2018). The main problem with statism is that it leaves unexplained both the well-known and disturbing similarities as well as the important dissimilarities between polities and many other polity-like political-ethical phenomena. Fortunately for us, however, the Williamsian framework can be used in an extended way to cover these issues too.

The ground for the distinction between membership in a polity and membership in other, polity-like political-ethical phenomena is far from obvious or easy to explain. Consider Augustine’s question in The City of God about the difference between a kingdom and a gang of bandits (Augustine 1998: 147). Augustine’s famous rhetorical question: ‘Justice removed, then, what are kingdoms but great bands of robbers?’ makes the explicit claim that what should distinguish polities from bands of robbers is the former’s observance of the demands of justice. In this we can see the textbook moralist solution to the problem in question (interestingly, coming from someone who is often associated with the realist tradition in a broader sense and who also suggests that in real life the boundaries between kingdoms and bands are unfortunately blurred).

The Augustinian question is echoed by those criticisms of associative accounts of PO that argue that associative accounts cannot explain the normative difference between membership in a polity and membership in some polity-like political-ethical phenomenon like, say, the mafia. These – moralistic - criticisms (e.g. Dagger 2000: 110–111; Simmons 1979: 16–23; Simmons 2000: 65–101; Vernon 2007: 872) practically repeat the Augustinian claim that without some additional justificatory move we cannot really explain why PO exists in a polity and not in a mafia even though we cannot really want to admit the existence of genuine PO in a mafia. The next step of these criticisms seems especially devastating: if we have such external justification then we do not really need local ethical practices or ethically significant facts to explain the normative force of membership in a polity.

There are multiple problems with this moralistic criticism and taking a closer look at these problems can help us better understand not just the limitations of moralistic accounts of PO but also what makes this paper’s account importantly different from typical associative accounts. The first problem is that the last step of the moralistic argument is in fact a non sequitur: even if associative accounts do mistakenly believe that they identified a sufficient condition for PO, still they can be right about a necessary condition for PO that other accounts have failed to take seriously enough. And they do seem to be right about that. The fault of Horton’s account, for example, is not that it would not be right to emphasize that the fact of membership in a polity plays a central role in grounding PO. Instead, its fault is that it fails to identify the proper norm-generating mechanism linking the fact of membership to PO.

Another problem with the moralistic criticisms is that, in fact, we may not even need an external justification at all to explain the normative difference between membership in a polity and membership in a mafia. It all depends on whether we can find an alternative norm-generating mechanism that can explain the normative difference between a mafia and a polity without falling back onto a moralistic account. Such an alternative explanation consists of two parts: first, it is important to see that the factual difference between a mafia and a polity does not need explanation. This factual difference is already presumed by those moralistic accounts that criticize associative accounts for being unable to properly accounting for it. Second, if this paper’s account holds true then there is more basic explanation for the normative difference between membership in a polity and membership in a mafia than what a moralistic account could offer. The explanation is that the authority claims that spring from membership in a mafia are qualitatively different from the authority claims that spring from membership in a polity because a mafia as a particular kind of political relationship is qualitatively different from a polity as a particular kind of political relationship. In other words, what makes sense with respect to a mafia does not necessarily make sense with respect to a polity and, accordingly, people are given sufficient reason to comply with different kinds of political authority claims in these two cases. This explanation is more basic than a moralistic explanation because it traces the normative differences between mafias and polities to the different kinds of minimal normative thresholds these two different kinds of political relationships aspire and are able to pass. Inasmuch as a mafia can deploy political means to bind its members together, it can be rightly described as a polity-like political-ethical phenomenon and it can be said to provide its subjects with sufficient reason for compliance. However, inasmuch as it does not aspire to, and even if it would, it could not rule its subjects as politically as a polity, a mafia can also be rightly said to be a qualitatively different political-ethical phenomenon than a polity and, therefore, it cannot be said to provide its subjects with PO.

5 Conclusion

This paper has offered a broadly Williamsian realist account of PO that traces the source of its normative force to the concept of membership in a polity. This account has sought to provide a realist alternative to the moralistic accounts of PO (i.e. has sought to show that PO is not a generic obligation and is not conceptually independent) in a way that retains enough similarity with the background assumptions of those accounts to make it obvious that still they conceptualize the same political-ethical phenomenon in a different manner. It also has sought to provide a plausible and more robust alternative to those membership accounts of PO that try to ground PO in someone’s communitarian identification (Horton), belonging to a nation (Tamir), joint commitment and plural subjectivity (Gilbert). In contrast to these accounts, this paper’s account has claimed that the source of the normative force of PO is someone’s membership in a polity because membership in a polity means that someone is being stably exposed to political authority claims associated with a polity that pass the minimal normative threshold of genuinely political form of rule and if a political authority claim passes this threshold then it provides someone with sufficient reason for compliance. The implications of this account are manifold but two seem rather obvious: first, the political-ethical phenomenon known as PO seems to make sense even beyond the realm of moralist political theory and, second, realist political theory should (and could) pay more attention to the problem of compliance than it usually does.