I am honored that Law & Social Inquiry sponsored a symposium examining my book, The Expressive Powers of Law: Theories and Limits (McAdams Reference McAdams2015), alongside Frederick Schauer's excellent book (Schauer Reference Schauer2015), and by the fact that such a distinguished group of scholars participated. I appreciate the opportunity to reply to the comments and criticisms, participating in what Don Herzog (Reference Herzog2017, 6) calls the “blood sport” of academic exchange, though that possibly exaggerates its popular appeal. In addition to a few brief replies and a substantial response to Herzog, I make one major point in reply to Robert Ellickson and Gillian Hadfield: that the book pervasively engages the existence of informal order by exploring its complex relationship with law's expressive power.

Regarding the comments by Daryl Levinson (Reference Levinson2017) and Janice Nadler (Reference Nadler2017), I shall have no significant reply because I agree with nearly everything they say. Levinson makes a nicely subtle point about the indeterminacy in game-theoretic analysis of rules—the unconstrained or unexplained choice of the level of generality in which to explain rules. For those who emphasize the coordinating function of law, as I do, should we locate the element of coordination in particular legal rules, in the entire legal system, or in some intermediate level? I certainly do not attempt to answer the difficult theoretical question Levinson poses. The social psychologist Janice Nadler illustrates the importance of context to understanding why people obey the law. I agree. As I put it in the book's introduction, I advocate “theoretical pluralism about compliance, the proposition that law brings to bear multiple powers at the same time” (2015, 7). Yet I confess that as a theorist who inclines more to “lumping” than “splitting,” I may occasionally need reminding of this point.

Robin Kar (Reference Kar2017) provocatively advocates the evolutionary analysis of the mental structures that make it possible for humans to coordinate, to share a sense of what is salient in a given situation. Here, too, my reply is brief (with one additional point in the next section). I agree with Kar that there is a deep puzzle in knowing exactly how people manage to coordinate on a focal point. When there are an infinite or extraordinarily large number of “features” of a given situation, how do people know which feature other humans will notice, much less the ones they will find focal? As I state in my book, I believe that rationality alone is insufficient to answer the question; what is mutually focal among strangers depends on some additional shared psychology. Kar's idea of evolved “obligata” offers a plausible explanation of this psychology.

Yet this evolutionary game theory strikes me as orthogonal to my project to demonstrate that law works, in part, as a focal point. It is similar to saying there is a deep puzzle in knowing exactly how people manage to form a common language. Given the empirical fact that language exists, we can make progress in explaining how the formalization of linguistic rules (as by a dictionary or grammar book) affects the use of language in a society without a theory solving the deep puzzle of language's origin. For my purposes, it is sufficient to say that expression can, as an empirical matter, influence behavior by creating a focal point. From that premise, I argue that legal expression can have the same power. All of which can be true without taking a position on Kar's theory of obligata.

Now I turn to the commentary of Robert Ellickson and Gillian Hadfield, which raise interestingly related complaints.

I. The Scope of Law's Coordinating Power and its Relationship to Informal Order

Ellickson (Reference Ellickson2016, 50) says that the analysis of The Expressive Power of Law tends toward “legal centralism,” being too focused on the efforts of governments to create focal points and provide information. He acknowledges that I “cite[] countless examples of private actors … who provide focal points,” but says I nonetheless end up with a discussion that is “overly state-centered” (53). Hadfield (Reference Hadfield2016, 20) suggests that the topics I selected do not include “the principle challenge” or “the real puzzle” of interest, which she posits to be an explanation of the origin of government and legal sanctions, that is, how order first emerges from disorder. As a result, she says that my account has “limited theoretical scope” and is “not ambitious enough,” applying only to a “restrictive set of cases” (23). Hadfield argues that the “more fundamental relationship” between law and focal points is not how the law's focal power incrementally adds to legal compliance in ordered societies with mature legal systems, but how focal points explain the origin of ordered societies, that is, how a state acquires the power to direct sanctions (23).

I offer two replies. First, I defend my reasons for prioritizing the influence of formal law. Second, I explain how, despite this priority, my book discusses informal order more extensively than these criticisms suggest.

When I first began working on the expressive effects of law, years ago, I was struck by two observations: (1) law and economics scholars inappropriately ignored the behavioral effects of law that did not arise from sanctions, treating the effect of law on behavior as the exact equivalent of the effect of legal sanctions on behavior; and yet (2) law professors outside of economics frequently made undertheorized causal claims about law's expressive effects. I was motivated to address these two shortcomings in the legal literature. I wanted to convince economists that their own tools—methodological individualism,Footnote ¹ rational choice, and game theory—predicted that law would sometimes influence behavior apart from its sanctions. I aimed to make expressive effects part of the standard economics toolkit, not present in every situation, but to be considered whenever the theory predicts their presence. Additionally, I wanted to articulate a causal theory that would identify limits to the expressive effects of law, so I could convince other law professors, not inclined toward economic analysis, that some of their exuberant claims were implausible or, in the alternative, to induce them to specify their own causal accounts.

I continue to believe that these legal claims are worthy of the attention I gave them. Being claims about contemporary law in mature legal systems, my goals have led me to focus primarily on governmental rules and the ability of those rules to influence behavior. The topics Ellickson and Hadfield emphasize are also important, but my first priority is not the literatures about social norms or the initial creation of formal order.

In any event, that is my first reply. My second reply is more important: The Expressive Powers of Law pervasively discusses and engages informal order, identifying a number of interactions between social norms and customs on the one hand and the law's expressive powers on the other, including the role of expression in the origin of legal sanctions. Before I elaborate these points, I wish to address a terminological disagreement. Hadfield and Kar each summarize my work by suggesting that it is centrally about a game of “pure coordination.” Hadfield (Reference Hadfield2016, 23) says that the expressive function of law that I identify “only operates in those particular settings in which individuals are playing a pure coordination game.” Kar (Reference Kar2017, 44) says I “introduce” my expressive claim by using “a stock example of a game of ‘pure coordination,’” which is the choice between driving on the left and on the right side of the road.

These descriptions are off the mark. Put aside the fact (as Hadfield acknowledges) that half the book is not about coordination, but elaborates a second expressive theory: that law communicates information (and thereby can change beliefs and thereby change behavior). As to the coordination theory, I explained that it is not limited to games of “pure” coordination, as that term is usually employed. I strived to distance my analysis from the entirely too familiar example of the choice between driving on the left or the right because that is conventionally understood as a pure coordination game (even though Kar and I agree it is probably not) (2015, 23). My introduction uses different traffic examples, such as the power of the yield sign and the center line (1–6), and throughout the book, I invoke the problem of two cars meeting at an intersection where the traffic light is broken or nonexistent, asking why the drivers would be influenced by a bystander who directs one to stop and the other to go. These situations are not conventionally understood as pure coordination games because the preferences of the parties conflict; they disagree about which equilibrium outcome is best. When roads merge or cross, each driver wants the other one to yield. Pure coordination games are rare, but games such as these mixing conflict and coordination are extremely common, so those who think the focal point power of law is limited to pure coordination games will greatly underestimate its power.

I devote many pages to developing the claim that such mixed situations are pervasive. A major purpose I had in writing a book that included the focal point theory after writing an article on the same subject was to give a more substantial reply to the concern that the theory, as Hadfield puts it, applies only to “a restrictive set of cases.” Early in the book, in a section titled “The Pervasiveness of Coordination” (2015, 27–44), I make a series of abstract game theory arguments that the cases to which the theory applies are not so “restrictive,” but common enough to merit serious attention, at least as common and as worthy of attention as the prisoners' dilemma, which has generated a vast social science literature.

Throughout the chapters on the focal point theory, I then supply a series of concrete examples, including constitutional law, international law, property disputes, the enforcement of custom, the prohibition of public smoking, and the regulation of traffic. To take only the last example, traffic accidents kill more than a million people annually worldwide, so it is a mundane but important domain for harnessing the law's focal influence. These arguments and examples obviously did not persuade Hadfield, but she does not really explain why, other than her view that her primary topic—the origin of order—is more important. In any event, I certainly agree that the focal point power I describe does not apply to an unrestricted set of cases, but I regard the limits of the theory as a feature, not a bug, as it helps to identify and reject some of the exuberant expressive claims law professors make. That is why my subtitle is: “Theories and Limits.”

Now back to the main complaint that I do not engage sufficiently the role of informal order and the origin of sanctions. To restate the differences with Ellickson and Hadfield, I focus on questions like why people comply with a particular law against public smoking or a one-way street sign, an unenforced law on how to honor the flag, or a property rule defining beach access rights (to name four of many examples), where Ellickson would focus more on the expressive influence on informal order and Hadfield would focus on how a government or sanctioning system first arises. Hadfield says it is the role of coordination in the function of sanctions that explains why “coordination is a central, not merely incidental, element of legal order” (2017, 23). She does acknowledge that I “give brief treatment to the coordination of third-party enforcement,” citing three pages of the book (where I reference her work with Barry Weingast) (23; Hadfield and Weingast Reference Hadfield and Weingast2012).

Three pages would be brief in a book of 261 pages, but my engagement with informal order is far more extensive and complex than that. Indeed, the existence of informal order plays an entirely necessary role in my focal point theory of the formal law's expressive power. Here is why: one challenge I faced in demonstrating that there was a focal point and information effect to law is in the difficulty of isolating those mechanisms from the conventional mechanisms of sanctions and legitimacy. When the law arguably possesses legitimacy and threatens sanctions, one can attribute all of the law's influence on behavior (whatever it is) to one of these mechanisms. To isolate the expressive mechanisms I proposed, I was at pains to identify examples where private nonlegal expression influenced behavior so that I could argue that whatever mechanism explained such influence was also at work in legal expression. Informal order plays a central role in my analysis.

Thus, even if informal order is not my primary topic, it is a recurrent one. One theme of the book is that law gains expressive power from its ability to harness and improve mechanisms enforcing preexisting private order. Whatever the source of informal order—however enforcement is managed—there is always ambiguity in the substantive customary rules, and law can influence behavior by clarifying those rules and then relying on existing informal enforcement. One example I discuss is the dictionary (2015, 107–09, 112). Language is a form of private ordering, one that usually first arises in a highly decentralized manner. People generally manage to coordinate on the components of language, but there are temporal and geographic sources of ambiguity, that is, conflicts over regional dialects and abandoned uses or neologisms. In this setting, a dictionary can reconcile the conflicts and ambiguities by selecting one set of usages as correct. Even though a private person publishes the dictionary for private gain, it can influence the language, which is to say, the dictionary can have an expressive power over linguistic behavior. Speakers of the language may disagree with the dictionary in certain cases, but nonetheless tend to give in to its focal resolution of their disagreements.

One of my themes is that law has this same expressive power when it incorporates customary law. Codification of custom involves choosing among regional and temporal variations of the customary rule. The formal law may influence behavior because people give in to its focal resolution of their disagreements over the rule's precise content. Formal law draws some of its power from the informal enforcement of customary rules, to which it makes marginal adjustments. Thus, rather than ignore the origin of formal order, I describe how formal order is parasitic on informal order.

This is a common claim in the book. As Ellickson acknowledges, my private ordering examples include discussions of the focal point effects of Roberts' Rules of Order (2015, 113), the rules of chess codified by the international chess federation FIDE (113), and the rules of card games published as According to Hoyle (112). I then identify how the basic structure of these examples—the operation of a focal point—is present in the emergence of: (1) property rules based on clarifying customs of possession (87–92, 110–12, 114–15); (2) a federal statue codifying and clarifying customary rules for respecting the US flag, (111–13); (3) international laws of war as first articulated and clarified in the Lieber Code (115–16); and (4) the initial emergence and eventual influence of the International Court of Justice, which draws power from clarifying international customary law (91–92, 119–22, 231). These discussions all bear on the origin of order. (I also explained how the process might work in the opposite direction, as where racially restrictive covenants, initially enforceable by legal sanctions, retained a focal point influence to maintain housing segregation even after the covenants were no longer enforceable, a point explored in great detail by Richard Brooks and Carol Rose [Reference Brooks and Rose2013, 105–06]).

I discuss social movements—private collective efforts to change social norms. Political scientists have discussed the role of coordination in the civil rights movement of the 1950s and 1960s, which tended to be organized by charismatic leaders (2015, 132). My primary example, by contrast, was the more mundane and decentralized effort to flip social norms regulating smoking from toleration to prohibition (100–05). This section explains how social movements use law as a way of coordinating private enforcement. A local population of ardent nonsmokers might find that their efforts to suppress smoking are too diffused over time and space to flip the norm. They need a way of “picking their battles” and concentrating their private sanctioning behavior. When there are no social movement leaders to fulfill this role, law can coordinate such efforts by designating a particular location as nonsmoking. Nonsmokers then withdraw sanctioning efforts elsewhere and focus them in the designated no-smoking zone (e.g., the bus station, the sports arena, the no-smoking section of a restaurant), which has become a focal point meeting place for nonsmokers. I argue that the law's expressive dimension could have a more powerful effect if the law changes incrementally precisely because small changes are better than large ones at leveraging the power of informal sanctions.

Perhaps of greatest relevance to Hadfield's criticism, I identify early in the book (2015, 57–60) a logical puzzle for any system of sanctions: Who will sanction the sanctioners? If one posits that Z complies with law out of fear that a judge will sanction him otherwise, the question is why Z will fear the judge if Z is physically stronger and/or wealthier than the judge. The answer is that the judge will order Y (say, the sheriff) to arrest Z or garnish his wages. But surely Y expects to incur costs by sanctioning Z, including Z's resistance, so the question becomes: Why will Y sanction Z? Why does Y comply with the judge's order? One can posit that Y—the sheriff—fears that she will be sanctioned by some party X (say, the bailiff), who acts at the command of the judge. But then the question is why X complies with the order of the judge. And so on, ad infinitum. At each level, the imposition of sanctions is costly, so theory needs to explain why a sanctioner bothers to impose sanctions.

My answer is to posit the importance of a coordinated set of expectations. In equilibrium, a population of individuals expects that, within the population, most others will obey the order given by a particular person—the leader, for example, a judge—and those expectations mean that there is fear of defying the leader's orders. That most others will enforce the judge's orders creates a selfish incentive for each to comply. I did not try to generalize about what I regard as a highly contingent and path-dependent process by which the office of judge is first created. I simply observed that the expectations can arise around judges just as they arise in various nonstate settings, such as organized crime bosses and pirate captains.

With this sketch of an answer to the logical puzzle, I drew an analogy between the legal sanctions in a mature legal system and a stable currency in a mature economy. It is also expectations that give a fiat currency its power—the expectation that others will accept currency as valuable in future transactions. I then say that in some economies “[t]he expectations are so stable that we tend to reify the value of currency, as if it had worth independent of the expectations underlying it. Money is valuable, we think, because it is money. The same mental error occurs in law. In mature legal systems, one may lapse into thinking that the sanctions backing law are an exogenous and independent force” (2015, 58).

Having identified the error, and the need to avoid it, we come to the three pages Hadfield references, where I sketch how sanctions can arise endogenously (2015, 117–19). Here, I cite Hadfield's work with Weingast, but I offer my own explanation. The fundamental problem for any decentralized third-party system of sanctioning, I say, is ambiguity, which creates the possibility that sanctioners will work against each other. There are a variety of ways to model an incentive-compatible system of third-party sanctions, but in any given case, third parties will sometimes hold a different view of the relevant facts or the relevant custom. Given a property dispute between A and B, for example, some third parties may regard A as the wrongdoer, and punish A, while others regard B as the wrongdoer, and punish B. The result is chaos, or at least the failure of sanctions to create a coherent incentive for any pattern of behavior. In this context, a salient leader who can influence the third-party sanctioners can give them the power of coordination. By following the leader—who directs that A but not B be sanctioned (or the reverse)—the group can enforce a coherent set of norms. I conclude (119): “At this point, the focal speaker (oracle, shaman, elder, judge, etc.) effectively wields the power of sanctions. If the actor is recognized as a government official, then the government wields the power of sanctions and we call these legal sanctions.”

Beyond this is a separate discussion of constitutional law, in which I directly address the formation of order (2015, 71–76). A number of scholars have observed that the selection of a constitution involves a coordination game, not the “pure” kind, of course, but the mixed variety. That is because, for the overwhelming majority of individuals, a large number of constitutional orders is preferable to anarchy. So there is a common interest in coordinating on one constitution, yet individuals sharply differ on which constitution is best, producing conflict. Here, I rely on some early work of Hadfield's coauthor Barry Weingast, who joined Russell Hardin in describing constitutions as focal point solutions to coordination games (in particular, a Stag Hunt game) (Hardin Reference Hardin1989; Weingast Reference Weingast1997). I review some more recent political science theory on how the focal point of law creates incentives for government bureaucrats to comply with law. I link this point back to the leadership discussion, saying that the constitution helps to both create expectations that people will follow the orders of the political leader who comes to power under the constitution, and also to limit the leader's power by creating expectations that the people will not follow blatantly unconstitutional orders. I also discuss the particular example of the inauguration of a president, a ritual that effectively obliterates the old expectations about whose orders will be followed and creates a new set around the new president (98).

Finally, where I focus in Chapter 7 on the expressive power of arbiters, I use this analysis to explain how adjudication works without state sanctions. Among my examples is one that Hadfield mentions in her work: the effect of judicial decisions in medieval Iceland (2015, 202–03, 230), despite the absence of a state apparatus to enforce the decisions. Tom Ginsburg and I first discussed Iceland in 2004 in our article on compliance with the International Court of Justice, which we say worked like the courts in medieval Iceland (Ginsburg and McAdams Reference Ginsburg and McAdams2004, 1241–42). In the book, I claim that expressive mechanisms resolve the puzzle of how adjudication works before there exists a mature state with a developed legal order, which in turn reveals some of the continued function of focal points in mature legal systems.

In sum, a substantial part of the book discusses informal order by way of explaining the origin of formal order. I think of the theme as secondary to my primary project of explaining why people comply with particular laws in a mature legal system, which may explain why the secondary theme crops up across many sections of several chapters rather than being the sustained focus of one section or chapter. But it is a major theme. If you are interested in the origins of formal order, including the work Hadfield has done with Weingast, I will immodestly suggest that you should be interested in The Expressive Powers of Law. If you are interested in the interaction of formal and informal order, I'm aware of no legal book since Eric Posner's Law and Social Norms (2000) that addressed the subject more broadly than I did, with the exception of Ellickson's The Household: Informal Order Around the Hearth (Reference Ellickson2008).

II. The Relationship of Law's Expressive Power to Alternative Theories

Don Herzog (Reference Herzog2017, 6) relates an example of legal compliance that he says is not explained by anything within the theory “menageries” that Schauer and I create, but is instead “another critter.” Herzog says that he and his colleagues fully comply with the Michigan law forbidding racial affirmative action in public education despite their strong disagreement with the law and despite the fact that they do not have reason to fear sanctions (11).

I confess and seek absolution for being a lumper and not a splitter (well, more of a lumper and less of a splitter than Herzog; as I once wrote in another context [2001, 681], at a certain point I, too, feel the need to “get off the reductionism train”). So it is no surprise that I believe that the compliance he describes belongs to a previously identified species of compliance theory—a legitimacy theory. Note how Herzog explains the compliance: “[S]ometimes we comply with the law because we think we owe it to a democratic majority” and “affirmative action is the kind of policy dispute properly left to democratic decision making” (2017, 12). He notes that some laws would be so “flagrantly unjust” that their democratic creation would not justify submission to them, but that the Michigan law is not so unjust (12). All of this sounds to me like a theory by which one accepts and feels obligated by the rules of a legitimate authority. Legitimacy theories differ by the source of legitimacy. Herzog's theory stresses a democratic source, subject to the constraint of substantive justice. That sounds related to Tom Tyler's procedural legitimacy—though Tyler locates the source of legitimacy in the fair treatment by the institutions of courts and police (see, e.g., Sunshine and Tyler Reference Sunshine and Tyler2003). And Herzog's justice side constraint sounds related to Paul Robinson and John Darley's substantive idea of legitimacy—though they address only criminal law (Robinson and Darley Reference Robinson and Darley2007). A state referendum might produce especially strong sense of the democratic legitimacy of the law, even for those who disagree with it, and such legitimacy might causally contribute to legal compliance.

Perhaps we use the word “legitimacy” differently. Relatedly, at one point in the book, I attempt to explain the possible synergy between legal legitimacy and law's focal point power. I endeavor to identify a characteristic of law that makes legal expression “stand out” from much other expression. In this context, I say: “By definition, a legitimate authority is the quality to which people defer” (2015, 124). Herzog strongly objects (2017, 8). He is certainly right that I did not want to resolve in this sentence, by definition, the empirical matter of how much legal obedience is due to legitimacy. By saying people “defer,” I did not mean to imply that they obey. Not only would it be extremely odd for me to deny that people can disobey law they perceive to be legitimate, but also Herzog is quoting a page where I set out a hypothetical scenario in which a person who feels obligated to obey a law nonetheless disobeys it. Instead, by “defer,” I meant merely that they “accept.” If people regard the law as legitimate, then by definition, they accept the right of the institution to make law of the sort made, whether or not they substantively agree with the law and/or obey it. My only point here was to assume arguendo some legitimacy theory of law so I could explore how it might interact with my focal point theory.

Elsewhere, Herzog says I “smudge[] a distinction we need.” He continues: “One thing to provide a focal point and solve a coordination dilemma, another thing to provide information—and still another to symbolically say, this conduct is wrong. Those I ordinarily think of as expressive theorists are interested in that last, and it is interestingly different from the first two” (2017, 8). I can only say that I am transparent about what I am doing and not doing in the book. In my first chapter, “Expressive Claims About Law,” I identify four literatures examining four different expressive dimensions of law. Two of these literatures are normative; two are positive. The fact that I devote my book to one of the positive questions—What are the effects of legal expression on human behavior?—does not blur any distinction arising from the other positive and normative questions. And of course law frequently declares “this conduct is wrong.” I engage the point most directly when I discuss Joel Feinberg's theory of the expressive dimension of criminal punishment (2015, 176–77, 241–44; Feinberg Reference Feinberg1970) and Kenworthey Bilz's explanation of crime victims' usual willingness to forgo private vengeance in favor of public punishment (192–93; Bilz Reference Bilz2007). But all of the book engages legal symbolism by seeking to understand why the law's symbolic condemnation affects behavior.

Finally, Herzog objects to my critique of social meaning theory (2015, 166–67), as advanced by Dan Kahan (Reference Kahan1997) and Lawrence Lessig (Reference Lessig1995,Reference Lessig1996). One of Lessig's examples is seatbelt use: if seatbelt use is low in a society, then one who uses seatbelts signals unusual fear of the riskiness of the driver, which raises the costs of wearing a belt; if seatbelt use is high, no such signal is sent. Given that setup, the law cannot change the social meaning unless it first changes the behavior, moving usage rates from low to high. If so, then the social meaning change for seatbelt laws is not the cause of compliance, but the result of compliance, which must be otherwise explained.

I stick to that position, but I now see how my claims here may seem to be too strong, that is, may seem to rule out any legal theory of social meaning change, rather than just the ones that I was addressing (the ones I knew of). As I more generally advocate pluralism in understanding legal compliance (2015, 7), I did not mean to advance such a strong claim, that no expressive mechanism exists other than the ones I describe.

Nonetheless, I am skeptical of Herzog's social meaning counterexample. He imagines that “[a]n antiracist city council votes to post a sign on the court: Only Racists Play Here,” and the result is that “whites might stop playing there … because they don't want to be seen as racists” (2017, 9). I assume that nonwhites might stop playing there for the same reason, but perhaps the problem hypothesized is that racist whites had previously intimidated other races from using the court.

Predictably, perhaps, I do not understand why this is not just a story of informational signaling of the sort the book explores. Because the hypothetical specifies that the elected city council voted to post the sign, that vote constitutes a revealed aggregation of the beliefs of the council members, which apparently include the belief that racists play on this particular court, and because those elected officials tend to be experts about the attitudes of their constituents, is further evidence of the attitudes of the community (2015, 141–43), which apparently is hostile toward racists playing on the court. One reason I gave in the book for why attitudes tend to affect behavior is that people value approval and seek to avoid disapproval as an end (139–40). So the vote and the sign reveal that a white person was, even before the sign was erected, risking disapproval as a racist for being one of the whites to play there. And, as some people may have previously been ignorant of the correlation, the addition of the sign signaling the correlation only strengthens the inference, thereby strengthening the incentive to avoid playing there. I do not assert that there is nothing else going on here, but I am not yet convinced there is a pragmatic gain to complicating the story of the causal mechanism at work.