Introduction

In their respective books, Richard McAdams (Reference McAdams2015) and Frederick Schauer (Reference Schauer2015) provide valuable insights that explain how law works. In The Force of Law, Schauer seeks to explore the conditions under which law “makes a difference” (2015, 49) in people's decisions and behavior, and argues that those conditions almost always involve the coercive power of law. By contrast, in Expressive Powers of Law, McAdams's chief ambition is to identify and explain two general functions of law that operate outside of deterrence-based legal coercion: the coordination and information functions of law. In this article, I offer further thoughts on how law might work outside of sanctions or direct coercion. My central argument is that to understand how law works, we must first appreciate that it governs a complex array of human activity, and that assertions about the general nature of law—for example, that law works mostly through coercion, or law works mostly through legitimacy—not only are difficult to prove, but also do not advance the ball very far. Instead, the question of how law works is more effectively examined in a context-specific way, focusing on the interaction of persons and situations. We then see that the way in which people interact with law is mediated by group life. Law does not generally influence individual behavior in a vacuum, devoid of social context; instead, group identity interacts with law to provide motivations to comply.

McAdams and Schauer introduce several intriguing notions about the ways in which law does or does not influence behavior. McAdams helpfully identifies sanctions, legitimacy, coordination, and information as separate mechanisms through which law can influence behavior. By contrast, Schauer argues that almost all mechanisms involving the influence of law outside of deterrence still involve coercion of some kind. But Schauer's conception of coercion is both exceedingly broad, and fails to acknowledge how law can interact with a variety of social processes to produce desirable behavior.

To conceive of law as an external force operating on discrete activities misses the basic fact that law permeates social life to such an extent that our norms, values, and understandings cannot be readily separated from the demands of law (Sarat and Kearns Reference Sarat, Kearns, Sarat and Kearns1993). In contrast to the instrumental view that law operates on autonomous individuals by providing a set of incentives, the social groups view holds that a person's attitude and behavior regarding any number of demands of law—whether it be driving, paying taxes, using drugs, or anything else—is a product of the interaction of law, social influence, and motivational goals that are shaped by that person's commitments to specific in-groups. To illustrate my argument, I discuss some examples of these interactions of law and social life, focusing on the legal regulation of risk. Before doing so, I briefly review standard conceptions of the coercive power of law, and explain why it is worthwhile to look beyond the standard account.

Legal Coercion

Legal systems enforce the obligations they create by means of a variety of sanctions. Coercive power is without a doubt an important part of how law functions, even though sanctions work imperfectly in practice. Notably, the coercive power of law manifests not just in punishment of criminal acts, but also in almost any activity that law governs (Schauer Reference Schauer2015). Thus, if a person wants to write a will, or form a corporation, or enter into a contract, law dictates that she must do these activities in one way rather than another. If she fails to act in the manner prescribed by law, she faces the sanction of nullity—no valid will, corporation, or contract exists (2015, 27–28). Moreover, to avoid civil liability or administrative sanctions, people are forced to refrain from doing things they might prefer to do. Thus, a manager who wants to fire an employee because she is pregnant, a food manufacturer that wishes to make a prohibited health claim, and a publisher that wants to engage in predatory pricing are coerced by law to refrain from these acts. It is readily apparent, then, that the coercive power of law is pervasive, extending beyond criminal punishment and into ordinary, everyday transactions and interactions (Ewick and Silbey Reference Ewick and Silbey1998; Valverde Reference Valverde2012).

In the domain of criminal law, the coercive aspect of law has given rise to analyses of the extent and circumstances under which punishment effectively deters (Nagin Reference Nagin1998). Consider anecdotally some instances of domains where the punishment that law threatens often does deter, as well as domains where threatened punishment makes little difference. For example, we indisputably rely on the threat of legal sanctions to prevent offenses like theft. Granted, moral values keep this behavior in check—most people think theft is wrong most of the time. And most people would desist from stealing where doing so would directly harm another, even though it is unlikely that the victim would engage in self-help. Thus, for example, most of us would not consider taking a frail elderly person's wallet even when that person would not notice and no one else was looking.

But in other contexts, more people are tempted to steal, and threatened punishment successfully helps people resist temptation, perhaps even most of the time. Many people are tempted to steal when it would not directly harm another individual, such as taking and eating candy from the bulk bin of a large corporate food market, and it is mostly the threat of legal sanctions that stops these people from stealing under these circumstances. There are also domains where the sanctions that law threatens do not deter very effectively, especially where the probability of detection is close to zero. These might include, for example, use of unlawful drugs in private residences, exceeding the speed limit by a modest amount, and evading taxes by failing to report income not subject to withholding. Specifying when deterrence works, and how well, is a standard domain of law and economics, and while there is a great deal of empirical work yet to be done, this ground is well trodden (Becker Reference Becker1968; Posner Reference Posner1985; Shavell Reference Shavell1985).

Of course, the coercive use of law is not without problems. There is serious concern that distortions in the US political process have led to contemporary criminal punishment regimes that are often unduly harsh as well as racially unjust (Stuntz Reference Stuntz2011; Alexander Reference Alexander2012). At the same time, rape law is largely underenforced, in part because of law enforcement failure to pursue investigations, leaving victims of sexual assault unprotected and officers unaccountable (Tuerkheimer Reference Tuerkheimer2016). Existing empirical evidence suggests that current crime control doctrines in the United States do not accurately reflect the community's sense of justice (Robinson, Goodwin, and Reisig Reference Robinson, Goodwin and Reisig2010) and also do not effectively deter in many contexts (Carlsmith, Darley, and Robinson Reference Carlsmith, Darley and Robinson2002; Tyler Reference Tyler2009). For example, merely 5 percent of the variance in illegal drug use can be explained by variations in expected likelihood and severity of punishment (MacCoun Reference MacCoun1993). Perceived probability of detection below a certain threshold can sometimes become psychologically meaningless, causing any potential deterrent effect to disappear (Baron and Ritov Reference Baron and Ritov2009). Knowledge of the relevant legal rule is often weak, even among those who have reasons to know the rules (Robinson and Darley Reference Robinson and Darley2004). Instead, people assume the law is as they think it ought to be (Darley, Robinson, and Carlsmith Reference Darley, Robinson and Carlsmith2001).

It is not clear that most individuals make the relevant cost-benefit calculation that deterrence theory presumes, and even when they do make such a calculation, situational pressures sometimes leave individuals unable to calculate rationally. Personal characteristics can also play a role—that is, an individual's high discount rate might result in him erroneously concluding that the risk of engaging in the criminal act is worth taking. In domains such as copyright, where social norms strongly lean in favor of permitting the activity law seeks to prohibit, rigorous enforcement might signal widespread noncompliance, thereby triggering even more noncompliance (Kahan Reference Kahan2001; Depoorter and Vanneste Reference Depoorter and Vanneste2005). Finally, it is worth noting that enforcing sanctions is costly, and costs include not only spending vast amounts of public money, but also devastating social costs that are difficult to measure (Braman Reference Braman2004; Clear Reference Clear2007; Burch Reference Burch2013).

Beyond Coercion

Because the coercive use of law is expensive, sometimes unjust, and often ineffective, it is worth asking whether and how law can influence behavior apart from (or in conjunction with)Footnote ¹ direct coercion. One possibility is that the legitimacy of law and legal authorities produces a feeling of obligation to obey. Thus, if people accept the right of the law to dictate appropriate behavior, then people might voluntarily comply without regard to threatened sanctions. There is a great deal of evidence that the legitimacy of various legal authorities depends to a large extent on the fairness of the procedures that those authorities employ (Tyler Reference Tyler2006a). It is less clear that law can produce compliance solely by virtue of its legitimacy, although there is some evidence to this effect (Gibson, Caldeira, and Spence Reference Gibson, Caldeira and Kenyatta Spence2005; Tyler Reference Tyler2006b; Papachristos, Meares, and Fagan Reference Papachristos, Meares and Fagan2012; Tyler and Jackson Reference Tyler and Jackson2014).

Relatedly, there is some evidence that if individuals perceive the criminal justice system as unjust, their compliance decreases because their motivation to defer is weak (Nadler Reference Nadler2005; Mullen and Nadler Reference Mullen and Nadler2008; Robinson, Goodwin, and Reisig Reference Robinson, Goodwin and Reisig2010). In one study, people who read newspaper articles about unjust laws later reported a greater willingness (compared to those who read about just laws) to violate the law themselves, such as stealing office supplies or engaging in underage drinking (Nadler Reference Nadler2005). In another study, people who perceived the outcome of an abortion trial as highly unjust were more likely later to steal a pen than those who did not (Mullen and Nadler Reference Mullen and Nadler2008). Findings like these strongly suggest that coercion is not the only mechanism by which law influences behavior. Rather, qualities of the law itself, including the extent to which it is perceived as furthering justice or reflecting community values, influence the extent to which people feel bound by law in general.

Sanctions, legitimacy, and related perceptions of law's credibility and propensity to do justice are all mechanisms by which law might influence behavior. At the same time, law functions in ways other than coercion and legitimacy. Richard McAdams argues that two of these functions are important and yet overlooked—coordination and information. These are both instances of expressive law: the claim that law influences attitudes and behavior by what it expresses.

The coordination function operates through expectations. When law highlights a behavioral choice in a coordination setting, it changes expectations about how others will behave. Thus, a law that announces “No Smoking” empowers nonsmokers easily to coordinate to confront any smoker who flouts the legal announcement. Knowing this, smokers who think about lighting up might then demur.

The information function operates through evaluation. If a new law requires the use of seatbelts, that legal requirement tells citizens a few things. First, it tells citizens what legislators collectively know about the risk of not wearing a seatbelt. Second, it tells each individual what other citizens generally think about seatbelts, thereby raising the possibility that others will think less of us if we do not wear a seatbelt. Finally, if a particular individual already favored requiring seatbelts, the new law makes that person realize that others favor it as well.

McAdams's (Reference McAdams2015) chief contention is that law has the power to influence behavior independent of threatened sanctions, and independent of its perceived legitimacy—by facilitating coordination, and by providing information. But Schauer is skeptical, and he contends that sanction-independent obedience to law is in fact hardly widespread. On Schauer's account, “people generally do what they want to do or what they think is right to do unless some external force makes them do otherwise” (2015, 4).

Despite appearing on the surface to be irreconcilable, these two accounts of how law works actually have substantial areas of overlap. For example, Schauer's contention that most instances of obeying the law occur “because of what will happen to us if we do not” (2015, 6) sounds on its face as if it is only referring to fear of punishment or other legal sanctions. But as McAdams (Reference McAdams2015) demonstrates, what will happen to us if we do not obey law encompasses other important reasons for following the law. Most prominently, when law provides a focal point in a coordination situation, that signal allows us to see precisely what might happen if we do not obey, and those bad consequences might not entail state sanctions at all. Law provides focal points in coordination situations that create expectations, providing reasons to obey and to fear the consequences of not obeying. In this way, law causes behavior, but not because of fear of legal sanctions, but because of the desire to maximize one's own outcome in light of information about what others are likely to do.

According to Schauer's framing, this is an example of “people generally do[ing] what they want to do or what they think is right to do unless some external force makes them do otherwise” (2015, 4). The external force here is the anticipation of the other person's behavior. But it is a big stretch to say that the creation of focal points by law constitutes legal coercion. Schauer's account therefore must include this focal point influence of noncoercive law, for which there is experimental evidence (McAdams and Nadler Reference McAdams and Nadler2005,Reference McAdams and Nadler2008), or it must characterize this influence as somehow coercive.

One key point of departure between McAdams (Reference McAdams2015) and Schauer (Reference Schauer2015) is how they conceptualize the information function of law. McAdams frames this function as a set of reasons why people obey the law independent of legal sanctions and independent of legitimacy. Legislation reveals attitudes approving or disapproving of a given behavior, which causes people to update their beliefs about what others think, which in turn might lead to behavior change in order to avoid the disapproval of others. Schauer, on the other hand, frames this function as yet another way that law coerces—“insofar as the effect is produced by way of triggering the coercive force of social norms, then coercion still occupies center stage in explaining how law affects behavior, although in this context we might think of law as simply outsourcing its coercive power to private actors” (2015, 148). Once again, this conception of coercion is unusually broad. It seems to rest on the assumption that individuals ordinarily derive their beliefs, motives, goals, and attitudes using conscious, deductive processes, independently of their social world. From this standpoint, when the individual is confronted with a social norm that makes her change her behavior to avoid losing the esteem of others, this looks coercive.

But there is a problem with this account—we do not consciously adopt a set of beliefs and goals in a vacuum. The individual is not a self-contained entity containing fixed attributes detached from social context. For one thing, people think, feel, and act as group members. Cognition, emotion, and motivation are shaped by attention to relevant others in the social context (Markus and Kitayama Reference Markus and Kitayama1991). Groups with which we identify help us make sense of new situations. We experience greater empathy and attention to the emotional experiences of those in our group compared to those outside of it (Brown, Bradley, and Lang Reference Brown, Bradley and Lang2006). Expression and experience of emotion are influenced by anticipated reactions of others. To complicate matters even more, our knowledge and attitudes are influenced by processes that often operate outside of conscious awareness. Social norms are but a single instance of various kinds of social influence we experience throughout social life. If we posit that social norms are inherently coercive, we must consider that all of social influence is similarly coercive, which implies that social life itself is inherently coercive. This again would seem to denude the term “coercive” of meaning.

Law in the Everyday Context

To bolster his claim that law works primarily through coercive means, Schauer argues that there is no evidence that “people are willing voluntarily to subjugate their own best judgment of what morality and policy require to what they consider law's mistaken moral or policy judgment” (2015, 98). This is probably true. If a physician strongly believes in a terminally ill patient's right to decide to end his life, but a law prohibits the physician from assisting in the patient's suicide, then the physician's decision not to assist the patient is most likely attributable to the physician's submitting to the law under threat of sanctions rather than the physician voluntarily subjugating her own best judgment of what morality and policy requires. Similarly, if a parent strongly believes that mandatory vaccinations will endanger her child's health, the parent will comply with the vaccination law because of the threat of legal sanctions rather than voluntarily subjugating her best judgment. In these examples it is the coercive power of law that shapes behavior.

Cases where individuals feel strongly about an issue tend to be salient, and examples come to mind easily. At the same time, instances of law conflicting with strongly held beliefs are a narrow and relatively infrequent category within the broader array of occurrences in which law might have influence on beliefs and behavior outside of the deterrence framework. Law governs many aspects of social life, and not just those about which people have moral opinions or strong policy preferences. It is in the everyday, humdrum practices that we observe something like Hart's puzzled person (in contrast to Holmes's bad man), who is simply trying to figure out what law requires under the circumstances, not because she wishes to avoid punishment or other legal sanctions, but because she just wants to do what law requires.

To illustrate this point, consider that much of law is bureaucracy, and here we find puzzled people without an opinion on what morality or policy requires, and at the same time with an accompanying motivation to comply for reasons other than (or in addition to) threatened sanctions. Such puzzlement often leads people to seek out legal advice where their own expertise is lacking. Examples might include: Do I need I business license to be an independent contractor? Do I need to complete an e-verify form when a visa extension is approved? How do I prepare an LLC operating agreement? What are the required bylaws for my corporation? These are examples mostly from the single domain of small business, but other domains like this surely exist. The puzzled person is motivated chiefly by the desire to do things the right way. This terrain is underexplored because unlike the physician and the parent in the examples above, this behavior never generates headlines, rarely generates debate, and fades into the background of everyday practices.

Many of the coordination situations discussed by McAdams are in some ways similar to the bureaucratic cases in the sense that people often have no opinion on what morality or policy requires. On the one hand, in contrast to the small business owner, a driver approaching a stop sign is not a puzzled person because the stop sign is readily visible and everyone already knows what a stop sign means. But on the other hand, like the small business owner, the driver in this situation wants to do what the law requires, simply because that is the best thing to do in this particular context, quite apart from the threat of legal sanctions. Thus, most reasonable drivers who approach a stop sign want only to know how to coordinate with other drivers. Drivers might inch forward rather than stop completely in order to gain an advantage.Footnote ² But in the end the main concern of the driver is avoiding a collision. Here, law prompts compliance, but usually not because of the sanctions law threatens. Instead, compliance comes as a result of a desire to coordinate to avoid a costly outcome.

Legal Regulation of Risk

The examples of the small business owner and the driver illustrate the idea that individuals can be influenced by law—but not because of fear of sanctions—in situations that are basically morally neutral. Let us contrast stop signs and business licenses with the legal regulation of risk, which generates examples that are exceptionally interesting because they often involve a complicated interaction of social and cognitive phenomena. The mechanisms of influence on people's attitudes and behavior in the domain of legal regulation of risk include informational social influence (what others do), normative social influence (what others approve of), personal moral values, attitude characteristics (importance, knowledge, elaboration, certainty, extremity, and accessibility of the attitude to the person), and motivational goals (e.g., cultural identity expression). Examples in this domain include the legal regulation of tobacco, guns, GMOs, vaccines, food, seatbelts, helmets, child safety seats, local building permits, fireworks, drugs, and drunk driving.

The ability of law to regulate risky activities successfully sometimes depends on the perceived moral characteristics of those activities. In the United States, policy debates about hazardous activities often have strong libertarian overtones, so that the activity in question is framed as a personal choice and therefore not the proper subject of government regulation. Examples include regulation of unpasteurized milk, motorcycle helmet use, vaccinations, guns, and smoking.

The smoking example provides an interesting case study. Over the latter half of the twentieth century, cigarette smoking moved from being viewed as a matter of personal preference to being an object of danger and disgust (Rozin Reference Rozin1999). Because the United States has strong cultural roots in individualism, an activity like smoking is considered by many as off limits as the subject of regulation unless it can be demonstrated that it causes harm to people other than the smoker herself. When such information about harm to others is disseminated, regulating the activity can help decrease the incidence of harmful behavior through a variety of mechanisms. In the case of smoking, legal regulations that worked through the mechanism of incapacitation were quite powerful: when smoking involved leaving the office building, restaurant, bar, or stadium, smokers often decided that the pleasure derived from the cigarette is outweighed by the cost. Other coercive legal mechanisms, such as taxation, were also influential in decreasing smoking.

At the same time, the evidence of harm to others that emerged in the latter half of the last century transformed smoking into an activity with moral implications. The large variety of legal regulations and legal activity that accompanied this evidence (place restrictions, taxation, labeling regulations, advertising regulations, sales-to-minors restrictions, OSHA restrictions, Federal Aviation Act restrictions, and tort litigation) might not only have coerced, but also contributed to the development of social norms that discouraged smoking. In the end, it is plausible that legal activities (including regulations and litigation), health information about harm to self and others, and social and moral norms about who smokes and who disapproves of smoking all mutually influenced one another over the course of the latter half of the twentieth century, with the result that smoking rates in the United States are now at an all-time low (Ng et al. Reference Ng, Freeman, Fleming, Robinson, Dwyer‐Lindgren, Thomson, Wollum, Sanman, Wulf, Lopez, Murray and Gakidou2014).Footnote ³ The coercive powers of law undoubtedly contributed to this decline. But to stop there when explaining what caused the decline in smoking in the United States would be missing much of what is interesting about how law, information, and norms interact over time.

Risky activities become harder to influence through expressive regulation when those activities reflect commitments based on cultural values. Consider gun control laws, which regulate behavior considered to be honorable by certain cultural groups. Hunting, for example, is considered by many Americans to be a sport, a means for feeding a family, and the foundation for strengthening the bond within a family. As such, hunting is perceived by many families as “a way of life that we believe in” (Brown Reference Brown2015, 41). Laws regulating guns, therefore, are perceived by some Americans to attack the fundamental identity and status of their cultural group, and to be an implicit statement that members of their group do not matter. The risk perceptions formed by members of these groups reflect and reinforce their beliefs, so that guns are viewed as normal and venerable rather than dangerous and undesirable (Kahan and Braman Reference Kahan and Braman2003,Reference Kahan, Dan and Braman2006).

Working together, these cultural values and risk perceptions strengthen the individuals' ties to their in-group, and distinguish their group from other groups. Whereas the regulation of smoking helped contribute to enormous changes in attitudes and behavior, the regulation of guns, by contrast, is unlikely to have a direct influence in shaping behavior among people who possess these cultural beliefs and ties. Indeed, looming regulation can cause backlash effects among those who feel that such regulation would threaten their values. For example, in the past few years, gun sales have spiked immediately following mass shootings. Gun dealers anecdotally report that these sales are driven by anticipation of new restrictions (Feldmann Reference Feldmann2012), and the data suggest that it is indeed the fear of new restrictions rather than the fear of impending violence that drives spikes in sales (Aisch and Keller Reference Aisch and Keller2015).

Risky activities become harder to influence through expressive regulation when those activities are seen as falling outside of the domain of legitimate government regulation. McAdams (Reference McAdams2015) argues that learning about new legal rules can cause individual citizens to update their beliefs because law provides information. Law can function as an attitude signal of what most other people approve or disapprove. In addition, when law prohibits certain risky behavior, such as using cell phones while driving, it is providing information not only about what other people approve of, but also actuarial information about the costs and benefits of the behavior. McAdams argues that because people are motivated to reduce risks that pose hazards to their health and well-being, they update their beliefs to reflect the costs and benefits of the prohibited behavior. This is true as far as it goes. However, this model is often too simple: the motivation to reduce health risks sometimes competes with several other motivations, some of which can crowd out self-protective attitudes and behaviors.

Consider, for example, the legal regulation of sugar-sweetened beverages (SSBs) in the United States. The idea of imposing a sales or excise tax on soda was introduced several years ago by public health scholars who wanted to reduce diabetes and obesity. State and local governments have proposed taxes on SSBs dozens of times, often unsuccessfully.Footnote ⁴ Health officials had hoped that by taxing SSBs, sugar consumption would decline because higher prices would lead to a decline in SSB purchases. Indeed, an empirical study of a nationwide tax on SSBs in Mexico demonstrated that the tax did in fact lead to a substantial decline in SSB purchases (Colchero et al. Reference Colchero, Popkin, Rivera and Wen Ng2016).

Those promoting efforts to enact taxes on SSBs have sought to gain support by trying to convince the public that reduction in sugar consumption is beneficial for health. In virtually all these proposed tax schemes, the anticipated revenue generated from the tax was earmarked to fund health care initiatives to treat the damage caused in part by the SSBs being taxed. This echoed closely the efforts of several US state governments, which have used tobacco tax revenues to fund antitobacco public education campaigns and health services. The proposed laws sought to work through deterrence, but also through attitude change along the lines that McAdams outlines: law would signal that legislators have concluded that costs of SSBs outweigh their benefits, which would cause people to update their beliefs about drinking these beverages, which would cause them to decrease their intake, all because people are motivated to increase their own levels of health and well-being.

But the fate of these proposed tax laws illustrates that people hold other motivations that compete with the motivation to increase their own health. The explicit effort to nudge people into making a healthier choice was met with very strong popular resistance in nearly every instance of proposed taxes on SSBs. Interestingly, the popular backlash seems to have been based substantially on resistance to nanny-statism: the idea that consumers have the right to choose their beverages without paternalistic meddling by the state. In efforts across the country to reduce SSB consumption through law, a substantial portion of the public was disturbed by the idea that their city or state should be in the business of influencing their attitudes and behavior about decision so personal as what and how much to drink.Footnote ⁵ The motivation to increase one's own health was certainly present, but it was overwhelmed by the motivation to maintain one's own status as an autonomous decision maker in the domain of beverage consumption.

Thus, in order for law to have an opportunity to have informational influence on attitudes and behavior, people must be open to being influenced in this way. In the domain of the legal regulation of risk, people are sometimes strongly resistant to such governmental influence, and this resistance can grow or diminish over time, depending on the subject matter that law addresses. For example, in the face of clear evidence that cigarettes are harmful to not only the smoker, but also those around her, cigarette taxes earmarked for health and tobacco reduction are now uncontroversial. However, the health case against SSBs is relatively new and subject to resistance, so that the ability of law to influence attitudes and behaviors expressively can be diminished or even eliminated by dynamics of influence that block legal change precisely because of the law's anticipated expressive effects. Thus, sometimes efforts to enact expressive law can have ironic effects because people recognize and resent the expressive nature of law, causing them to oppose the enactment of the proposed law.

Further evidence of these backlash effects of proposed expressive law can be found in the example of Philadelphia, which in 2016 became the first major US city to impose a tax on SSBs. In proposing the tax, the mayor made it politically palatable by denying that the proposed law would be expressive or have any expressive effects, thereby avoiding the potential backlash of proposed expressive law. He did so by carefully avoiding any mention of health-related expressive effects of the law. The stated purpose of the SSB tax was not to reduce consumption of sugar or improve health, but to fund universal pre-K, renovations to libraries, parks, and recreation centers, as well as municipal pension obligations. During public debate about the proposed tax it became apparent that the mayor “is not using the word obesity, or suggesting that people should drink less soda.… [W]hen asked about the health benefits of the tax, he says, ‘There's really serious health benefits in pre-K’” (Sanger-Katz Reference Sanger‐Katz2016).

With the “non-expressive law” fiction firmly in place, the city council agreed to approve the measure, but not before doubling down on the fiction: the initial proposal was for a 3-cents-per-ounce tax only on SSBs; the version that the city council ultimately passed was a 1.5-cent-per-ounce tax on all soft drinks—including artificially sweetened drinks (e.g., Diet Coke) that do not contribute to diabetes, obesity, and heart disease. The city government thereby completely disclaimed any expressive purpose of the law, and the proposed tax was accepted because it would bring tangible benefits that are unrelated to any government effort to change attitudes or behavior about SSB consumption.

This example illustrates that the very prospect of expressive law can lead to backlash when the topic of law or regulation is deeply embedded in cultural beliefs and practices. These embedded cultural values are not merely attitudes held by individuals in a vacuum. Rather, they exist in a larger context, and they develop and change along with the needs of the relevant group in which they are embedded.

Law and Group Dynamics

The earlier example of cultural values surrounding gun regulation illustrates a more general point about how the expressive influence of law interacts with group dynamics. Individuals are guided by the values of groups that they identify with (Dawes, van de Kragt, and Orbell Reference Dawes, van de Kragt and Orbell1988). People do not think about themselves and others as a collection of individuals; people instead tend to think in terms of social clusters, and assign themselves and others to those clusters. This gives rise to a motivation to belong (Baumeister and Leary Reference Baumeister and Leary1995), which in turn provides individuals with a framework for navigating the social world. To talk about how law influences individual behavior in a vacuum, devoid of social context, is to ignore the ways that group identity interacts with law to provide motivations to comply. These include motives to cooperate, to be loyal to the group, to adhere to group norms, and to avoid social exclusion (Brewer Reference Brewer2004).

Law can work expressively not so much by shaping independent individual attitudes as by shaping group values and norms, which in turn influence individual attitudes. In short, the way in which people interact with law is mediated by group life. For example, in Turkey, where seatbelts are required, the belief that wearing seatbelts saves lives does not predict seatbelt use but the belief that “people who are important to me approve my using a seatbelt” does predict seatbelt use (Şimşekoğlu and Lajunen Reference Şimşekoğlu and Lajunen2008). The motivation to belong and to identify with one's group leads individual group members to make efforts to understand what others in the group would approve of, and to act accordingly.

This example is similar to McAdams's information account of expressive law, which posits that legislation can signal what is socially approved. McAdams argues that sometimes law expresses what most people expect one to do or refrain from doing. If we assume that individuals are concerned with what most people approve, law serves as a signal for the judgment of most people, whose approval individuals seek. Yet this account oversimplifies things by focusing on the individual, on the one hand, and “most other people” on the other. Individuals usually identify with specific groups, rather than society in general or some general notion of social approval. Much of the time, individuals are concerned not so much about what most other people think, but rather by what members of their relevant in-group think.

For example, the legal regulation of activities in domains involving risk, such as driving, food, and drugs, is complicated and often controversial because risk perception differs across groups, leading groups to differ sharply about appropriate legal regulation. Interestingly, these differences and similarities do not always cleave according to predictable boundaries. Consider legal regulations mandating childhood vaccinations, which are disapproved of by discrete and disparate groups such as affluent, educated parents in Malibu, CA, on the one hand, and low-income midwestern rural parents with high-school diplomas on the other (Omer et al. Reference Omer, Salmon, Orenstein, de Hart and Halsey2009). Individuals belonging to those groups seem relatively uninfluenced by information signaled by mandatory vaccination laws because they perceive nonvaccination as socially approved within their own specific in-group, despite the law.

At the same time, the vast majority of Americans think that vaccines are generally safe, and the vast majority of parents ensure that their children's vaccinations are up to date (Omer et al. Reference Omer, Salmon, Orenstein, de Hart and Halsey2009). For the large proportion of parents who do comply with mandatory vaccination requirements, that compliance is likely explained by a combination of threatened sanctions, advice and information from medical professionals, and an understanding of what is socially approved by their relevant in-group. But where, as here, social approval varies sharply between social groups, the source of influence is the group most relevant for the individual. When the expressive function of law plays a role, its effects are felt most directly on group norms, which then in turn affect individual behavior.

In this way, law can make a difference to behavior apart from sanctions, but as part of a complex system involving the interaction of individuals and groups. It is for this reason that I believe Schauer (Reference Schauer2015) is mistaken in his general skepticism that law “makes a difference” to behavior apart from the threat of sanctions. Schauer contrasts people who comply with law for prudential reasons related to threat of sanctions with those who comply because of law's status as law—“law qua law.” “[I]f those who take the very fact of law as a reason for action or reason for decision are few and far between, then coercion resurfaces as the likely most significant source of law's widespread effectiveness and longstanding appeal in achieving various social goals” (2015, 52). But this sets up a false dichotomy that forms the underlying premise of the book: there are people who comply with law because of the threat of sanctions and people who comply with law just because it is law. Schauer argues that the latter group is either small or nonexistent because there is a lack of empirical evidence that members of the group exist in large numbers, and because when we see people acting consistently with law for reasons other than threat of sanctions, the behavior that is consistent with law is not in fact caused by law. It is instead caused by nonlegal phenomena like social norms.

But this is a gross oversimplification because it assumes that when people act consistently with law for reasons of social norms their conduct is not caused by law. Schauer does not sufficiently consider the ways in which social norms and law can mutually influence each other. For example, it might be that but for the seatbelt law, the Turkish seatbelt wearers would not perceive that important members of their in-group approve of their wearing seatbelts because social approval itself developed through a dynamic of the mutual influence of law and norms. Perhaps some people became convinced of the efficacy of seatbelts, and as a result legislators began considering a law requiring them, the debate about which caused more people to approve of seatbelt wearing, causing more to wear seatbelts along the way, which led to momentum for the legislature to enact the law. Thus, individuals noticed that people they care about would approve of their wearing a seatbelt, which caused them to do so, but not completely independently of law.

Empirical demonstrations of the interaction of law and social norms come from the domain of contract law. Assignment of contract is both legally uncontroversial and common. At the same time, assignment of contract is in tension with social norms of promising and reciprocity. When you promise to give something to me in exchange for my giving something else to you, I feel bound to you by this promise. But if you assign your rights under the contract to an assignee, my feelings of obligation do not necessarily extend to the assignee, as that is not the party to whom I made the promise and with whom I previously interacted. Because of this, I might be more inclined to break my promise to the assignee than to break my promise to you. And, indeed, empirically it turns out that efficient breach is more likely toward an assignee than toward the original party (Wilkinson-Ryan Reference Wilkinson‐Ryan2012). Here, the law imposes identical duties on the promisor regardless of whether the duty is owed to the original party or to the assignee. But the interaction of law and social norms regarding promise and reciprocity causes the promisor to behave differently toward an assignee.

Understanding how law works requires recognizing how law influences and interacts with individual attitudes and behavior and, more broadly, social norms and social movements, and political institutions and their leaders. McAdams and Schauer have each made strides along this path. At the same time, law is not a monolith—it is composed of rules and standards, sanctions and remedies, procedures, agents, and institutions, among other things. To talk about “how law works” is to attempt to describe the consequences of an exceedingly complex system. McAdams and Schauer are doing critically important foundational work to advance our understanding of the influence (and noninfluence) of law. The time is now ripe for other scholars to build on their foundational insights to construct more socially situated accounts of how law works across a variety of contexts.