The Nature of Law and Potential Coercion

This paper argues for a novel&#160;understanding of the relationship between law and coercion. It&#160;firstly&#160;refutes Kenneth&#160;Himma&#8217;s&#160;claim that the authorisation of coercive enforcement mechanisms is a conceptually necessary feature of law.&#160;It&#160;then claims that the best way to understand the law is as coercion&#8208;apt. The &#8220;coercion&#8208;aptness&#8221; of law is clarified, in part, by appealing to an essential distinction between law and morality: Whereas it can be reasonable for the law to appeal to coercive means in order to motivate compliance, it seems decidedly unreasonable for morality to do so.


Introduction
Law coerces. This assertion seems banal and yet a debate concerning the conceptual relationship between law and coercion has been present in jurisprudential literature for decades. Of course, that law coerces is not the focus of discussion but rather whether coercion is a conceptually necessary feature of law. If X is a conceptually necessary feature of a concept, then X is an existence condition of the concept in question. As a crude example, a cause is a conceptually necessary feature of an effect because an effect cannot exist without a cause. Thus, claiming that coercion is a conceptually necessary feature of law is the same as saying that the existence of law is determined, at least in part, by the presence of coercion.
There are different ways in which coercion can be said to be a conceptually necessary feature of law. One of the more classic assertions is: It is a conceptual truth that every legal norm, R, is coercive such that R includes a coercive sanction for its violation. 1 Another classic explanation is: It is a conceptual truth that every legal system, L, is coercive such that it includes coercive sanctions for the violation of some of its rules. 2 Owing primarily to H. L. A. Hart's arguments found within The Concept of Law (Hart 2012), such classic assertions are generally taken by legal philosophers to be erroneous. Hart argues that the relationship between the concept of law and coercion is grounded exclusively in natural facts of our world-facts like resource scarcity and the fallible, selfish nature of humans. Coercion, according to Hart, is not a conceptually necessary feature of law (or of legal systems), but a natural necessity of our legal practices.
Over the past decade multiple works have arisen questioning Hart's rejection of coercion as a conceptually necessary feature of law. Some theorists, like Ekow Yankah (2007), Andrew Stumpff Morrison (2016), and Joseph D'Agostino (2017) defend variations of the classic claims mentioned above. Morrison attempts to raise the sovereign from the dead and reinstitute the command theory of law, whilst D'Agostino's (2017, 1) main claim is that the law, via legislators, inherently intends to threaten violence against those who violate its norms. In his work, Yankah (2007Yankah ( , 1197 argues that "the ability to use coercion under special conditions defines legal norms." As noteworthy as these articles are, Kenneth Einar Himma's work, spread over a series of articles and chapters, is more comprehensive, and his claim is more moderate. Contra Morrison, D'Agostino, and Yankah, Himma (2016, 594) explicitly denies the classical claims that coercion is a conceptually necessary feature of a law qua legal norm or law qua legal system. Himma's claim, rather, is that the authorisation of coercive enforcement mechanisms (henceforth CEMs) is a conceptually necessary feature of the concept of law qua legal system (Himma 2013(Himma , 2016(Himma , 2018(Himma , and 2020. 3 According to Himma, the central function of the legal system is to keep the peace. Furthermore, "law can efficaciously keep the peace only by backing some legal norms with authorized coercive enforcement mechanisms" (Himma 2018, 154). A system of norms that does not authorise the use of CEMs, per Himma, cannot keep the peace, cannot function in the way that law is supposed to function and, therefore, cannot be properly identified as legal.
The present work will argue that Himma's arguments in support of a conceptually necessary relationship between law and coercion (so understood as the authorisation of CEMs) are unsuccessful. The thought is: If Himma's more modest claim is not correct, then it is not clear how else a conceptually necessary link between law and coercion can be supported. My suggestion is that, whilst the authorisation of CEMs is not a conceptually necessary feature of law, it is a conceptual truth that law is coercion-apt. 4 It may not be immediately evident what the difference between these two positions is, 2 This claim is typically attributed to Hans Kelsen, who argues that coercion is a distinguishing feature of legal systems. According to Kelsen, we should think of legal orders as directing legal officials how to respond when subjects of the law act in a certain way (Kelsen 1945, 63). So, the legal duty to not steal ought to be interpreted as stipulating which sanctions are to be applied by officials if a theft occurs (ibid., 61). For Kelsen, then, part of the essence of a legal order is that it commands the use of sanctions by officials. 3 Many thanks to Ken Himma for sharing the penultimate version of his chapter "Can There Be Law in a Society of Angels?" (now Himma 2020). 4 This phrasing is directly influenced by Leslie Green's (2008Green's ( , 1050 observation: "Law is the kind of thing that is apt for inspection and appraisal in light of justice; we might say, then, that it is justice-apt."

A Brief Note on Coercion
Coercion is often used in a vague sense by legal theorists. Typically it serves as a general descriptor for the law's ability to motivate compliance with legal requirements and prohibitions. However, coercion, not unlike law, is itself a complex philosophical phenomenon with a deep bench of literature devoted to it.
Consider the following scenario: A would-be robber approaches a man from behind, holds a gun to the man's head and says, "Give me your wallet and I won't pull the trigger." The man slowly pulls his wallet out of his jacket and holds it up until the gunman takes the wallet and runs off.
This scenario is generally taken to be a paradigm instance of coercion. Few would claim that the man decided to give his wallet away, qualifying the scenario by claiming that he was coerced into giving up his wallet. A fruitful account of the nature of coercion should be able to sufficiently explain why it is an instance of coercion in addition to being an instance of theft.
The task of isolating what it is about a certain scenario that makes it "coercion" is easier said than done. Conflicts in the literature are numerous, and some theorists may disagree on one point only to agree on others. For example, there is disagreement as to whether offers can be coercive in the same way as threats (see Nozick 1969;Zimmerman 1981). There is also disagreement over whether coercion denotes success, that is, is whether the concept "failed coercion" is conceptually impossible in the same way as "married bachelor" (see Bayles 1972;Oberdiek 1976). Another prominent debate is whether coercion is a moralised concept (see Frankfurt 1973;Nozick 1969;Oberdiek 1976;Wertheimer 1987). 5 We can furthermore ask if the use of force to physically manipulate another as a mere means (e.g., manipulating someone's hand to force their signature) counts as an instance of coercion. Again the literature is split; some maintain that the physical manipulation of another as a mere means is coercive (see Anderson 2008;Frankfurt 1973;Lamond 1996Lamond , 2000Lucas 1966;Yankah 2007). Those who disagree argue that any instance in which the agency of another is completely 5 A concept is "moralised" if a moral assessment is a feature of its nature. As Hans Oberdiek (1976, 80) writes, "coercion embodies a moral assessment: insofar as an act or institution is coercive, it is morally unjustified and therefore stands in need of a moral defense or excuse." Thus, if coercion is taken to be a moralised concept, then the question of "wrongness" is not separate from the question of whether an act is coercive. It is not: Is act y coercive and, furthermore, is act y wrong? Rather, it is: Act y is coercive and therefore it is prima facie wrong and stands in need of justification (see Lamond 1996Lamond , 2000Zimmerman 2002). neutralized (as in the physical manipulation of another as a mere means) is not an instance of coercion (see McCloskey 1980;Nozick 1969;Oberdiek 1976). Instances of coercion, according to this latter group, are best conceived as instances where the agency of the coerced agent has been substantially reduced or manipulated by a coercing agent, but not entirely neutralised. For legal philosophy, this schism is particularly interesting as it marks the difference between whether enforcement tactics like imprisonment, restraints, and barricades are properly identified as "coercive force." Yet another debate asks whether an instance of coercion is identified with reference to its cause (e.g., the actions of the coercing agent), or its effect (e.g., the pressure a coerced agent feels to comply with the coercer's demands) (see Anderson 2010). This debate is of particular importance for this paper as it has interesting consequences for how we identify examples of the law's coercive force. For instance, we can consider those occasions where a legal official acts with coercive intent, like when a judge threatens to further sanction an individual who refuses to pay an outstanding fine. Here, the judge is acting in her capacity as a legal official with the intent of forcing the defendant to do what the law requires: to pay their original fine.
We can also consider those instances where an individual felt pressured by the threat of a legal sanction and, correspondingly, altered their behaviour to keep out of trouble. Instances of coercion like this probably occur on a near-constant basis. Anytime an individual behaves "otherwise" because they think the threat of legal sanction looms too large, it seems appropriate to conclude they fell prey to the coercive force of law. However, the identification of such behaviours as "coerced" is problematic because of their subjective nature. Different people experience this sort of coercive force of law with varying severity, due not only to differences in personal psychology (including, but not limited to, phobias, personality disorders, and intellectual limitations), but also to biological and sociological considerations such as race, sex, gender-identity, and socioeconomic status. As a simple example: A poorer person is more likely to feel coerced by the threat of a parking ticket than an affluent person.
Thus, pointing to an empirical outcome (e.g., a person decides to not park illegally) does not guarantee the precise identification of the law's coercive force. It is just as possible that an individual decided against parking illegally simply because they did not want to park in an accessible space, in which case the threat of sanction had no coercive power. In other words, such an understanding of the law's coercive force (e.g., that it can possibly be identified anytime someone decides to act legally) makes for a shaky analytic foundation. If, however, the focus is kept on those instances of coercion that originate with a specific legal directive (you must do this, or face a sanction), as with the judge issuing a directive, we have an unambiguous notion of the law's coercive force to consider.
The treatment offered in this section of the concept "coercion" and the ways in which it pertains to law is far less thorough than the topic deserves. For present purposes, however, it will suffice, as it has at least provided a somewhat clearer sense of the query being investigated: Is the integration of coercive mechanisms a conceptually necessary feature of law? This is the query that Himma addresses, but his response is not simply that the integration of coercive mechanisms is a conceptually necessary feature of law but, rather, that the authorisation of their use is.
(i) The Hartian claim that the sense of "law" requiring explication picks out municipal legal systems in the modern state; (ii) widely accepted Razian claims about how our legal practices construct the content of our legal concepts; (iii) claims showing the centrality of [CEMs] in every paradigmatic instance of law we have ever known; and (iv) logical difficulties arising in connection with explaining legal normativity in a system without such mechanisms. (Ibid.,594) Broadly construed, Himma's argument progresses as follows: Per Himma, "Hart identifies 'municipal law in a modern state' as picking out the sense of 'law' that forms the subject of conceptual jurisprudence" (ibid., 598). From this observation, Himma goes on to claim that "Hart's selection of 'municipal law in a modern state' as the subject of conceptual jurisprudence indicates the conceptual relationship between the notion of a legal system and the notion of a state" (ibid., 599). Himma then asserts that the concept of law qua legal system ought to be connected in some way to the conceptual function of the modern state, which, per Himma, is to provide efficient social control, to "keep the peace," so that a Hobbesian state of nature is kept at bay (Himma 2016(Himma , 2018. He writes: Little argument is needed to see that these conflicts of interest [e.g., material scarcity, greed, selfishness] among beings like us in such a world are likely to become violent at times. It is this natural propensity for violence in a world like ours that creates a need for keeping the peace through social control. (Himma 2016, 601) Himma is correct that "little argument is needed" for one to accept the claim that a means of social control is necessary for maintaining a relatively peaceful society. What, however, explains the link between successful social control and the need for coercion? For Himma (2018, 154), it seems to be effectiveness: "[L]aw can efficaciously keep the peace only by backing some legal norms with authorised [CEMs]." Granting such a link [that the need for social control entails the need for coercion] puts Himma's claim in synch with Hart's justification for what he termed a claim of natural necessity.
As mentioned in the introduction, Hart argues that coercion is not a conceptually necessary feature of law qua legal norm, nor of law qua legal system. Despite this, Hart maintains that there is a sense in which coercion is necessary to law. His claim is that given the natural facts of the world and the people who inhabit it, coercion is "naturally necessary" to the function and practice of our law. By natural necessity, Hart means to pick out the way in which we need coercion not only to compel the compliance of would-be lawbreakers, but to also provide a guarantee that those individuals who voluntarily live according to the rules of law will not be harmed by (or "sacrificed to," as Hart writes) those would-be lawbreakers (Hart 2012, 198). The relationship between law and coercion, then, is "contingent on human beings and the world they live in retaining the salient characteristics which they have" (ibid., 199-200). 6 It seems, then, that Himma is not saying anything too different from Hart. Yet, Himma maintains that coercion (so understood as the authorisation of CEMs) is not merely naturally necessary to law, but is a conceptually necessary feature of law qua legal system. The leap from natural to conceptual necessity is granted by Himma's premises (ii) and (iii). Premise (iii) is that the authorisation of CEMs is central to the function of "every paradigmatic instance of law we have ever known" (Himma 2016, 594)-"central" in such a way that it would be mistaken to call a system of norms "legal" if they lacked such authorisation. Therefore, per premise (ii) (that legal practices construct the content of legal concepts), the authorisation of CEMs (as a central legal practice) is a conceptually necessary feature of law.
In order to unpack Himma's claim that the authorisation of CEMs is a conceptually necessary feature of law, I will defend the following conclusions: Firstly, that premises (i) and (ii) result from misapplications of Hartian and Razian claims. Second, that premise (iii) suggests the flawed methodology of using empirical evidence to ground a conceptual claim. Lastly, whilst premise (iv) may flag an important concern for legal philosophy, it is not in itself a good enough reason to accept Himma's claim.

Law of Angels
Premise (i), the claim that the central concept of interest for legal philosophy is municipal law in the modern state, is relatively uncontroversial. The worry, however, is that Himma overextends premise (i) by claiming that the concept of interest for jurisprudence ought to be restricted to municipal law in a nation state. In his own words, he sees "no plausible reason for adopting a concept of law that departs from what we are and what we do with the law"-here, Himma (2016, 603) is referencing Raz's society of angels thought experiment.
In the society-of-angels thought experiment, Raz (1990, 159) asks us to imagine a society comprised of angels who "[lack] all desire to disobey their [legal] rulings." A society built and maintained by such creatures, Raz claims, would need a legal system to provide for authoritative norms, societal coordination, and legal institutions such as courts and a legislative body. Courts, according to Raz, are required for reasons beyond allocating coercive sanctions, such as the interpretation of legal disputes. Even angels, he suggests, could disagree over the terms of a contract, or be unsure who should pay damages if a tree damages a neighbour's property. A society of angels would therefore also require laws providing for rights and duties, "[s]ince accidental damage might occur [...] from [angels] acting wrongly because they misapprehended the facts or misinterpret[ed] the law" (ibid.). 7 The upshot of the thought 6 Coercion is not the only feature Hart thought was "naturally necessary" to the concept of law-he leaves that an open question, only wanting to point out that there is this "other" sense in which common features of law are related to the concept of law; that legal theory does not only have at its disposal conceiving of features of law as either conceptually necessary or not. 7 It is important to note that Raz denies such civil remedies count as coercive force because they are designed to impart duties on responsible parties, and not to deter would-be lawbreakers. Neil MacCormick (1982) supports such an understanding of the nature of civil remedies: That civil remedies coerce, he claims, is not an essential characteristic of theirs-what is essential is that they repair harm. experiment is that it is unreasonable to claim the angels lack a legal system simply because their system does not need to institute, nor authorise, coercive mechanisms. As Raz (perhaps infamously) writes: A legal system without coercive mechanisms "is humanly impossible but logically possible" (ibid., 158). 8 Undercutting the effectiveness of the society-of-angels thought experiment is critical for the success of Himma's argument. Recall Himma's Premise (ii): "widely accepted Razian claims about how our legal practices construct the content of our legal concepts" (Himma 2016, 594). Himma supports premise (ii) by attributing the claim that "it is the core features of our legal practices that construct our concept of law" to Raz (ibid.,603). 9 However, as was just briefly explained, Raz maintains that coercive mechanisms are not conceptually necessary in order to identify a normative system as legal. So, if Himma can successfully show that the angels in Raz's thought experiment do not actually have a legal system, then he can advocate more strongly for the position that coercive mechanisms are central to not only our legal practices, but to legal practices in general. And if coercive mechanisms are central to legal practices in general, then Himma's premise (ii) will successfully support his overall conceptual claim regarding the conceptual necessity of the authorisation of CEMs.
According to Himma, morally perfect angels would only really need their normative system to solve coordination problems and make provisions to regulate disputes. Such a system would not be legal in character because it would not need to provide for any level of social control. "Social control," Himma (2016, 600) writes, "is only necessary where there is a significant probability of conflicts that threaten the peace among self-interested persons living together in a society of material scarcity." There are, however, more types of social control than merely prohibiting or requiring conduct and, if we follow Raz, the law in the society of angels is a tool of social control-similar to ours-that grounds and maintains the organization of that society. That a legal system of angels is successful at keeping the peace without the need to authorise CEMs is down to the population being angels predisposed to doing not only what is morally right, but also what the law requires. 8 Despite the theoretical weight the society-of-angels thought experiment is generally given, it is not Raz's argument that coercion is not a conceptually necessary feature of law. Rather, the thought experiment is simply an illustration of his claim that coercive sanctions are auxiliary reasons for action. The sanction, per Raz, is not a reason against a reason to act (the sanction is not an exclusionary reason for action). Rather, whereas the legal norm provides a reason against a reason to act, the sanction merely complements the legal norm. The general thought is that the coercive sanction does not enhance our understanding of the normativity of the law, because the person who is coerced into following the law because they want to avoid a sanction is simply avoiding a sanction. This is why sanctions, per Raz, are "reason[s] of the wrong kind" (Raz 1990, 161). Wrong not because it is morally wrong to act on the strength of auxiliary reasons, but wrong because in practical reasoning the exclusionary reason is enough of a reason to refrain from acting in a particular way. The sanction, in other words, should not be necessary. This is the foundation upon which the society-of-angels thought experiment rests. 9 Himma does not offer explicit evidence for the Razian claim as he puts it, but perhaps the passage of Raz's that he is referring to is the following: "The concept of law is part of our culture and of our cultural traditions. [...] But the culture and tradition of which the concept is a part provide it with neither the sharply defined contours nor a clearly identifiable focus. Various, sometimes conflicting, ideas are displayed in them. It falls to legal theory to pick on those which are central and significant to the way the concept plays its role in people's understanding of society, to elaborate and explain them" (Raz 2001, 237). If Raz is correct, Himma (2016, 608) claims we may then conclude that "it is not a conceptually necessary feature of a legal system that it include a system of criminal law." Himma takes this conclusion to be absurd-the law cannot claim to regulate conduct if it does not in some way prohibit or require specific conduct. However, the claim that morally perfect beings would not require a criminal code depends on a particular understanding of morality as both universal and static. Part of the reason a society of angels would require law, in Raz's view, is that morality is at least partly underdetermined. Because there are questions as to what definitively counts as a moral wrong, choices need to be made regarding what should count as a legal wrong and there is no reason to think this would not extend to criminal law. Raz argues, e.g., that it may not be certain what constitutes murder as opposed to manslaughter; we know that there are differences here, but if the lines between them are indeterminate, then we need someone to draw them for us so that we all know exactly where we stand in relation to each other (Raz 1979, 245).
Perhaps an angel would never commit murder, but the point here is that not all questions of morality are as clear-cut as whether or not killing is wrong. To extend the point, of equal relevance are those factual disputes that need legal settlement. Whether a particular act, X, is an instance of some criminal category might not be obvious, and even angels would need someone to settle such uncertainties for them. Being morally perfect does not entail that the angels are perfect when it comes to empirical matters. It therefore seems intelligible to conclude, as Raz does, that even morally perfect beings would need a legal system to fill in those gaps that moral perfection cannot solve in society.
We can press this point a bit further by considering the Razian condition of comprehensiveness, which Raz takes to be a conceptually necessary feature of law qua legal system. 10 By comprehensive, Raz (1979, 151) means to capture that quality of a legal system that claims the authority to regulate "any form of behaviour of a certain community." Whilst a legal system that lacked a criminal code would lack a great deal of regulations that we take to be central to our legal practices, the lack of a criminal code does not necessarily mean a normative system would lack comprehensiveness. Raz explicitly states that comprehensiveness does not require a legal system to regulate all forms of behaviour, only that the legal system claims the authority to do so (ibid.). The lack of a criminal code, then, does not represent a lack of comprehensiveness so long as that legal system has conferred the power to enact such norms on a legal official. As Raz explicitly states that the society of angels would have a legislature, it is not a stretch to presume that such legal official(s) exists(s).
Worries about the content of an angelic legal system aside, Himma also questions how relevant the thought experiment actually is to legal philosophy. He writes, "it seems reasonable to characterise law in a society of angels as, at best, a borderline case of law that is significantly removed from the core cases of legal systems that construct the content of our legal concept" (Himma 2016, 604). Such a claim relies on a dichotomy between the concept of our law and a general concept of law that would hold in all possible worlds. He claims, per premise (iii), that the concept of law "of interest" ought to speak to what is central and ubiquitous to all instances of law that we know-that we have ever known (ibid., 602). If it cannot, then how can it be of any philosophical use to us? However, what is missing is an argument for how claims regarding the "ubiquity" and "centrality" of certain legal practices, claims that can only be verified through empirical study, can promote an observation (all legal systems authorise the use of CEMs) to the level of conceptual truth (the authorisation of CEMs is a conceptually necessary feature of law). According to Himma, this is merely a conflict of intuitions: "at the end of the day, the dispute between those of us who believe that the relevant concept of law does not extend beyond beings sufficiently like us and those who do not will rest on different intuitions" (ibid., 604).
However, what the previous discussion on the society-of-angels thought experiment has shown is that despite the fact that angels are very different from "beings like us" (because they have different motivations and require different things from their legal systems), they still have the need for a legal system and their need for a legal system is actually similar to our need for a legal system (to organise and structure society so that we can coexist peacefully). Therefore, the conclusion that coercion (or its authorised use) is not a conceptually necessary feature of law is not a failure to recognise the importance, or even the centrality, of coercion to our legal systems. 11 It is simply a conceptual claim about law.

What about Normativity?
Himma's fourth premise, that there are insurmountable logical difficulties that arise when attempting to explain the normativity of a legal system that lacks the authorization of CEMs, is the most difficult to unpack. The claim that "law is normative" is how legal philosophers capture the sense in which law is action-guiding. Himma (2016, 620) thinks this claim requires philosophical explication, as "it is not always clear how to properly understand this claim [that law is normative]." He goes on to assert that properly explaining the normativity of law requires answering the following questions: Firstly, how it is that the law provides reasons for action (if, indeed, it does do such a thing)? Second, what kind of reason for action does the law provide? 12 In The Force of Law, Schauer (2015, 72-3) asks a similar question: [W]e do not know how much of law's effect on moral and policy attitudes is a function of law's sanction-independent content and how much is a function of the emphasis supplied by the sanction. Could law have the opinion-forming or opinion-influencing it has, however much that may be, without the way in which the sanction arguably underlines the importance of the legal norm itself? 11 Frederick Schauer (2010, 9) claims something similar when he writes: "Without just that empirical foundation, a theory of law, or even an account of the concept of law, seems distortingly detached from reality. It is central to Hart's claim against Austin that the focus on coercion is distorting precisely because it ignores the non-coercive, sanction-independent, empowering, and non-hierarchical nature of much of law as it actually exists, but if it is grounds for criticism of a descriptive theory of law that it ignores dimensions of legality and legal systems that are widely present in the world, then it seems to be also grounds for criticism of Hart's account that it ignores the coercive dimensions of law, dimensions that are no less real than the noncoercive dimensions that Hart properly chides Austin for neglecting." 12 Unfortunately, it is beyond the scope of this piece to discuss whether there is such a need to explain the normativity of law to the degree Himma takes to be necessary.
We do not know what mechanism(s) drive the law's normative force and, according to Himma, it is impossible to tell that story without invoking the authorization of CEMs. If the point of law is to structure and organize society, then that society needs to be able to trust that the law is able to motivate compliance with its directives. Many will be unlikely to generally follow the laws if they do not have some guarantee that everyone else will generally follow them. In our world, Himma notes, that feat is impossible without the authorisation of CEMs. Therefore, Himma's claim is that we can only sufficiently explain the normativity of law by accepting that the authorisation of CEMs is a conceptually necessary feature of law. However, notice that this again presents a conflict between conceptual analysis and empirical observations. Himma (2016, 601) claims the conceptual necessity of CEMs is linked to the fact that "[t]he most effective method for deterring [...] conflicts, though it will clearly not always be successful, is to threaten a violent legal response." It is an empirical claim that municipal law in a modern state cannot efficiently motivate compliance without the authorisation of CEMs. But is it entirely accurate? Just a simple example to consider: Access to education has been demonstrated time and again to dramatically reduce crime rates. 13 While it would be very interesting to know, for example, the extent to which education serves to reduce crime rates compared with the law's authorisation of coercive mechanisms, we do not have access to such data.
Furthermore, that properly explaining the law's normativity is complex does not mean that understanding the reasons people follow the law is similarly complex. The observation from Section 2, that pointing to an empirical outcome (e.g., a person decides to not park illegally) does not always guarantee the precise identification of an instance of coercion, can, I think, be fruitfully applied to this point. Different people follow the law for different reasons: Some may think an action is right (or wrong) independent of its legality; some (like the angels) think it is simply right to follow the law in general; some may fear being socially shunned for breaking certain laws; and yes, some will follow a law simply because they do not want to be punished. All can be true at once without all necessarily reflecting conceptual truths about law. From the standpoint of legal philosophy, perhaps it simply ought to be accepted that most people follow most laws most of the time. 14 The question, then, becomes: Even if the threat of a sanction sometimes, or even reliably, serves to coerce a person into complying with the law, does that necessarily mean, as premise (iv) contends, that there are "insurmountable difficulties" associated with explaining normativity in a system that does not authorise CEMs? The best answer here seems to be that it is contingent, i.e., dependent, on the facts of the matter. Of those legal systems that have authorised CEMs, it is utterly appropriate to say that the authorisation of CEMs is necessary to that system's normative character. However, stating that the authorisation of CEMs partly explains that system's normative force does not mean that such authorisation constitutes the normativity of law in general-even if all the systems we can point to have integrated such authorisation.
A strength of Himma's argument is the fact that all current examples of municipal law authorise CEMs and, as such, it is correct to conclude that the authorisation of CEMs is a significant part of understanding the normativity of our municipal law in our modern states. Phrased as such, however, it is a redundant claim and, per the previous discussion, when the concept of law in general is being considered there do not seem to be any reasons to accept premises (i)-(iii) as supporting the conclusion that the authorisation of CEMs is a conceptually necessary feature of law. This is a significant impasse. For those who pursue the study of the law at its most general, beyond the confines of our natural world, it does not seem possible to defend the claim that coercion is a conceptually necessary feature of law. And yet, as Himma rightly claims, the relationship between the two is ubiquitous and we should be able to parse out why this is. The best we have, it seems, is to revert back to Hart's claim that coercion is a naturally necessary feature of law. He writes: [W]e do need to distinguish the place that sanctions must have within a municipal system, if it is to serve the minimum purposes of beings constituted as men are. We can say, given the setting of natural facts and aims, which make sanctions both possible and necessary in a municipal system, that this is a natural necessity. (Hart 2012, 199;Hart's italics) So, we can explain the relationship between law and coercion solely with reference to human nature and the needs of our societies, which furthermore means that investigations into the concept of law will have nothing to contribute to our understanding of the relationship between law and coercion. And yet, such claims, particularly as justified within Hartian jurisprudence, are inadequate. In an early work, Hans Oberdiek (1976, 74) writes: "Even if Hart is correct [about natural necessity], this only tells us something about humans and the human predicament; it tells us nothing about the concept of law." The suggestion that coercion is a naturally necessary feature of law, Oberdiek points out, does nothing to enhance our understanding of the concept of law.
To essentially brush aside a consistently observed feature of legal systems as merely a human fact and exclude it from philosophical inquiry is a central worry Schauer articulates in The Force of Law (2015). It seems intuitively wrong that Hart's claim of natural necessity is used by theorists to dismiss a seemingly universal character of law as outside the scope of legal philosophy, and it is these nagging worries that seem to be behind the recent resurgence in those works defending a conceptual relationship between law and coercion. The question, then, becomes: Is there a way to explain the ubiquitous relationship law has with coercion at the highest level of abstraction of the legal system, where not even the authorization of CEMs is a conceptually necessary feature of law? The goal is to provide an account of the relationship between law and coercion as originating from law, not from the human predicament.

Potential Coercion as a Conceptually Necessary Feature of Law
The core claim of the present work can be summed up thusly: Law is coercion-apt. Thinking of law as coercion-apt can give theorists a way to think about and discuss a facet of law's nature that legal philosophy has hitherto struggled to explicate. What, however, does it mean to suggest that law is coercion-apt? This section will fill in this picture by explaining what makes law specifically coercion-apt, and how this quality can help to further refine the distinction between law and other normative systems like positive moralities.
Another way to put the claim "law is coercion-apt" is to say that law is (in part) constituted by potential coercion, and the allusion here to potential energy is intentional. Potential energy, as present in a drawn bow, for example, does not necessarily result in kinetic energy in any conceptual sense, and yet potential energy both accounts for and explains kinetic energy when it occurs. Likewise, conceiving of law as being constituted (in part) by potential coercion can help account for the seemingly ubiquitous institutionalization of coercive mechanisms without logically entailing the claim that the fact of coercion (or even the authorisation of its use) is a conceptually necessary feature of law.
The potential energy of a drawn bow is the consequence of the bow being pulled, generally by an archer. Accordingly, a bow at rest has no potential energy. Unlike a bow, however, the law does not require an "archer" in order for it to contain potential coercion; the law, unlike a bow, is never "at rest." One way to cash this out is to appeal to the assertion that legal systems are systems of norms and the essence of a norm is that it aims to guide conduct. Legal systems, if we follow Raz, claim the comprehensive authority to serve as the supreme guide of conduct for all those who are within its jurisdiction, be they citizens, residents, or visitors. Because law claims this comprehensive authority, the law also assumes the dual responsibility of ensuring compliance with its norms and protecting those who would voluntarily comply against those who would not. It therefore claims the standing authority to institute coercive mechanisms if necessary and-because of this-we can say it exists in a constant state of being "pulled," of being coercion-apt.
Contrast law to something like conversation, which ought not to be thought of as coercion-apt. Conversation, if we consider its typical use, is an information-sharing opportunity; its end is to make your thoughts explicit so that your interlocutor understands you. Conversation, then, does not have any underlying normative quality because the goal is simply mutual understanding. Even if a conversation were to slip into an argument, we should resist the urge to claim that such a verbal exchange is coercion-apt. The point of argument is to persuade, and for that you must (ideally) fully explicate your position to your interlocutor in order to convince them that your position is the better one. A successful argument would convince your interlocutor to agree with you of their own volition and, logically speaking, there cannot be any type of force or threat involved. So, if you were to say to your interlocutor, "If you agree with me, I will do X for you" or, "Agree with me, or else," then the argument has ended.
That an argument can slip into a coercive exchange should the arguer become desperate enough to "win" that they resort to threats is not sufficient reason to claim that argument coercion-apt. This is because once such coercive threats are exchanged the argument has morphed into something else, something of a very different nature to "argument." Contrast this to law: If officials trigger its coercion-aptness by instituting coercive mechanisms, it does not make sense to say that the law has morphed into something of a different nature. Rather, the officials have tapped into the law's potential to be coercive. Thus, to claim that something is coercion-apt is to claim that its nature will not change if it then becomes coercive. 15 What of positive moralities? Just as there is instrumental value in threatening someone to agree with you in an argument, there is also instrumental value in resorting to coercive mechanisms to motivate compliance with generally accepted moral norms. And yet, despite this instrumental value, there is something distinctly inappropriate in resorting to coercive mechanisms within the context of positive moralities. If, for example, someone said, "We should sustain our moral practices by coercing people into complying," a suitable reply would be, "But morality is not the kind of thing for which this is an appropriate option! Once you have resorted to coercive mechanisms, you are no longer playing the morality game and have transformed morality into something else-something more like law." However, if someone said, "We need to sustain our legal system by coercing people into complying because, in the circumstances in which we find ourselves, this is the only way the system can be sustained," it would not make sense to reply that law isn't the kind of thing for which this could ever be justified.
What is true at the most general level, despite the many different ways in which "coercion" can cash out as noted in Section 2, is that coercion is a distinctive way in which an agent is compelled to act. The person who is coerced is, in some sense of the word, forced against their will to do something. Morality, however, hinges on voluntary compliance and appeals to reason. Once coercion is co-opted into its sphere, it becomes something very unlike morality. Here I have in mind what Hart writes in Law, Liberty, and Morality (Hart 1963): that using coercion to enforce positive morality does not actually serve to preserve its moral character, but rather contorts it so that it is no longer moral.
Positive moral norms endure when they are voluntarily adopted, when those who adopt them do so because they believe such principles to be the right thing. This is why Hart writes that the preservation of positive morality is done through argument, advice, and exhortation (ibid., 75). He does not deny that coercion has an instrumental value where the maintenance of positive moral norms is concerned. Rather, Hart denies that what is protected or promoted is actually morality. He writes: "Much morality is certainly taught without [fear of legal punishment], and where morality is taught with it, there is the standing danger that fear of punishment may remain the sole motive for conformity" (ibid., 58). Moreover, if at a certain point in time the only motive for conformity with a moral norm is fear of punishment, then there is good reason to believe that morality no longer plays a role in practical reasoning.
Unlike positive moralities, at least as Hart discusses them, law is not primarily concerned with whether or not its subjects think it is a legitimate authority, or that what is legal is also moral, or even that doing the legal thing is the same as doing the "right thing." Law, especially if we focus on duty-imposing norms, is very concerned with compliance. The law, however, is more than duty-imposing norms; how, then, 15 It seems, then, that aptness is nothing more than that which is possible. A car can be blue or not blue and its nature will not change; is it correct to say that a car is blue-apt? I do not mean coercion-aptness to be as anaemic as that and yet, there does seem to be something about the nature of a car that it is apt to be decorated, to be customised. It has the potential to be blue, or yellow, or red, or wrapped in a metallic or matte film, etc. Aptness may, at its most crude, be reduced to mere possibility, but I mean for it to signal something more like a disposition. is the claim that law is coercion-apt connected to power-conferring legal norms? It seems immediately apparent that it would be strange for the law to coerce a subject to exercise a power-conferring norm. Indeed, in contract law it is illegal to coerce someone into entering into contract with you. However, in exercising some power-conferring legal norms, one can expand their legal duties, and it seems correct to claim that duties generated by power-conferring legal norms are coercion-apt.
Can coercion-aptness also serve to distinguish law from social conventions? To be sure, societal pressures exist as a way to force compliance with various social conventions, and some of those societal pressures can feel coercive. Perhaps, however, the same inappropriateness that attaches to morality when coercive means are considered also attaches to social conventions? The point being advanced, then, is not that only law is coercion-apt, but that there is a distinction in appropriateness between the realms of morality and social conventions on the one side and law on the other where coercive mechanisms are concerned. And the distinction is such that one ought to be hesitant to bestow the characteristic "coercion-apt" on either morality or social conventions.
It is also important to consider whether the law is apt for any other types of motivation, or if coercion is a type of motivation distinctly appropriate to law in a way other means of motivating compliance are not. For example, some people may be motivated to follow the law only because they fear societal repercussions, like harm to their reputation. However, we cannot discuss the ways in which the law is apt to protect reputations in the same way it is coercion-apt because the law cannot intentionally generate social pressures in the same way that it can intentionally generate coercive mechanisms. Societal pressures to follow the law-or even to break the law, as in occurrences of civil disobedience-are not grounded in legal mechanisms, but in societal norms and opinions. Societal pressures may converge with legal norms, but such convergence is not necessary, nor even predictable.
Rewards can also motivate compliance with norms, so is law reward-apt in the same way that it is coercion-apt? Rewards can serve as powerful incentives, but as Bentham (1970, 134-6) points out, a reward-based system of law could not be depended on and would look quite absurd. For example, if a reward for not stealing were offered, would individuals be paid per day they did not commit an offense? For every time you do not steal, are you to be rewarded a credit to the store in question? Would rewards be universally applied, or only issued to those who have broken laws in the past in an effort to get them to "reform"? And, if in the latter case, wouldn't such a system lead to more lawbreaking because everyone would want to claim rewards? 16 It is true that the law can piggy-back off of reputational or moral motivations, and can motivate compliance by offering rewards. However, the point being pressed in this section is the following: What makes law distinctive from morality is that moral norms are not apt for invoking coercive force in order to motivate behaviours in the same way as legal norms. It is not an open question whether moral norms can be enforced coercively, because morality is the kind of thing that excludes resorting to The Nature of Law and Potential Coercion coercion. When it comes to law, however, such questions relating to the incorporation of coercive mechanisms are always live.

Conclusion
In this paper I set out to explain, firstly, that coercion is not a conceptually necessary feature of law, and second, that coercion-aptness is a conceptually necessary feature of law. Thus, the claim being promoted is that "law" and "coercion-aptness" are conceptually connected in the sense that we would not identify L as an instance of law were we not also able to identify L as coercion-apt. As explained in the previous section, conceiving of law as coercion-apt can help to distinguish law from morality: Moral norms are duty-imposing, but are not coercion-apt because the character of morality is such that it would contort its nature to resort to coercion to motivate compliance with its directives.
If we return to the society of angels, the claim "law is coercion-apt" can also develop Hart's claim that coercion is naturally necessary to law. Yankah writes that the thought experiment should cast serious doubts on whether the angels actually have a legal system: If one (fallen) angel inexplicably decided to disobey those edicts, the inability of any structure to even theoretically compel him to do so would, I suggest, cast serious doubt on the system's claim to be law. (Yankah 2007(Yankah , 1236 Casting doubt on the legality of the angels' system of norms, however, is not the only possibility. As Green (2016, 167) points out: "Even the angelic legal system contains all the powers necessary to get a coercive apparatus up and running in short order. It is unusual only in that it does not use its necessary powers in that way." The lack of a coercive mechanism in a society of angles is not evidence that the angels lack a legal system, but rather is evidence that either the legislature has not yet conferred the relevant powers on certain legal official(s), or the powers have been conferred, but not yet exercised. One could therefore respond to Yankah's worry by claiming that the fallen angel has simply given the relevant official(s) of such a legal system a very good reason to trigger the coercion-aptness of their legal system and create, authorize, and enact coercive mechanisms.
Moreover, were the angel-officials to institutionalise coercive mechanisms in an attempt to redeem the fallen angel, the possibility exists that they would be instrumentally justified in doing so because it would be a productive measure to take in securing the desired aim of successfully motivating the compliance of the fallen angel and protecting the law-abiding angels from other fallen angels. It seems correct to borrow from Hart's original statement of natural necessity and say that considering law's nature (and its claim to normative authority in particular), authorising coercive mechanisms is not only instrumentally justified, but is also an instrumentally optimal way to maintain the rule of law. This is not a moral claim, as instrumental value is independent of moral value. There are obvious examples of wicked legal systems, and wicked laws, for which there can be no moral justification for the use of institutional coercion. And yet, there is a sense in which it is instrumentally justified for even a wicked legal system to authorise coercive mechanisms to sustain the rule of its law.