Lauren Edelman's presidential address is artfully provocative. Given my membership in both the Law & Society Association and the American Law & Economics Association, I endorse her twin suggestions that law and society (L&S) scholars give more attention to the economy and that L&S and law and economics (L&E) scholars consider building bridges between their divergent methods. But there is a tension in the way Edelman frames these recommendations. Edelman proposes that L&S spend more time studying the economy partly because she finds the existing study, dominated by L&E, so inadequate. One might wonder why L&S should seek any reconciliation or interaction with economic methodology if it does not offer something that L&S currently lacks. I respond here to Edelman's respectful critique by attempting to demonstrate that economics has sufficient value to L&S as to make worthwhile the efforts at collaboration. Indeed, I will provocatively assert that, because L&S is, methodologically speaking, a “big tent,” the difference between alternative branches of L&S methodology within that “tent” is as great as the difference between the more positivist branches of L&S scholarship and certain elements of L&E scholarship. In any event, I seek to identify a specific and unexpected convergence between the schools, where the theoretical topography is best suited for building collaborative bridges.
There are several points of possible convergences other than the one I will address. Worth noting, for example, are the trends in economics toward greater complexity of behavioral assumptions and greater empiricism. Many economists are replacing the more reductionist assumptions of homo economicus with models of boundedly rational and boundedly selfish individuals, based on the steady stream of psychological and economic experiments finding powerful evidence of cognitive biases and nonselfish motivations, and the importance of context to each (Reference CamererCamerer 2003; Reference Fehr and FischbacherFehr & Fischbacher 2002; Reference Korobkin and UlenKorobkin & Ulen 2000). Another trend—weaker than I would like—is a move toward the study of inequality. Perhaps the first to recognize the potential for using game theory to explore issues of inequality was the philosopher Reference Ullmann-MargalitUllmann-Margalit (1977:134–97). But there is increasing interest in using games to model social inequalities, especially those related to sex. See Reference RoseRose (1992), Reference HadfieldHadfield (1999), and Reference WaxWax (2000).
My primary focus, however, is on the contrast Edelman draws between L&E and L&S on the grounds of what I will term contingency. Edelman says that L&S uniquely emphasizes “the social embeddedness and politics of markets” (Reference EdelmanEdelman 2004:183), “the social, political, and legal construction of rational economic behavior” (2004:184, emphasis in original), and “social action as responsive to institutions, norms, and historical context” (2004:187). Addressing one aspect of the market, for example, Edelman asks “why we bargain at the car dealership or the flea market but not at the pharmacy or the grocery store” (2004:192). She answers that pricing practices “are governed by culturally ingrained practices as well as, and sometimes instead of, preference maximization” (2004:192). Although Edelman may mean other things by this contrast I believe she means also to emphasize contingency: that there was no necessity to arriving at the existing pricing practices, but that arbitrarily small changes in history and culture could produce very different patterns of contemporary behavior. We can understand pricing practices and other economic behavior only by understanding the history that has led us to them.
Here is where I wish to sketch out a possible ground of unexpected convergence. To a degree not fully appreciated, economic theory embraces the idea of contingency, and with it, the importance of history and culture. After describing how contingency arises in game theory, I return to a source of divergence: whether L&S can embrace the converse—that the noncontingent practicality or “function” of market practices sometimes explains their persistence better than contingency.
Game Theory Embraces Cultural Contingency
Economics appears to reject contingency because the one-shot prisoners' dilemma game and its variants are such popular theoretical devices, and they have only one possible “equilibrium.” A Nash equilibrium is an outcome at which no individual can benefit by unilaterally changing strategies; each is playing his or her best response to what the others are doing. Because an equilibrium is solely a function of the individual payoffs, which the economist derives from “stable preferences,” a single-equilibrium game is the antithesis of socially constructed or culturally determined values or rationality. And so L&E and L&S theorists quickly reach an impasse.
There is a way to sidestep this impasse, at least in part. Most games have multiple equilibria. That is, even granting fixed preferences and all other standard assumptions of game theory, there are situations where the payoffs do not fully determine the behavior. Consider three examples. Before a society determines what side of the road its cars should travel on, potential drivers face a pure coordination game. Each prefers to drive on the right if most other drivers do, but to drive on the left if most others do. Language usually presents a coordination game because people usually prefer to use symbol A to refer to object X if others generally do, but to use symbol B to refer to object X if others generally do. Even disputes can involve coordination. Two disputants may each prefer that the other be the one to “give in,” but each may prefer to “give in” rather than endure the conflict that results if neither gives in. Thus, when the costs of conflict over the two equilibria (either one giving in to the other) are sufficiently high, disputants want to coordinate by avoiding conflict (Reference McAdamsMcAdams 2000). In all these examples, each individual's best course of action depends on what he or she expects the others to do. But if more than one equilibrium is possible, then here's the rub: the first time the game is played, no one knows what to expect the others to do.
If payoffs don't determine initial expectations, then something else might (in some cases, must). Decades ago, Reference SchellingSchelling (1960) noted that non-payoff features may influence behavior by making a particular equilibrium salient or “focal” to the players. People tend to play the strategy associated with the one equilibrium that they expect others to recognize as “standing out” from the others. For example, people who get lost from one another often coordinate by meeting at the place each will expect to be conspicuous to the other. In four decades since Schelling, game theory has made many strides, but it has not made much progress in understanding focal points, that is, in understanding how expectations emerge independent from payoffs. This should come as no surprise because, as Schelling pointed out, what is focal “may depend more on imagination than on logic, more on poetry or humor than on mathematics” (1960:57). Schelling also noted the inherently empirical quality of what is focal, depending as it does on “precedent” and on “who the parties are and what they know about each other” (1960:97).
Thus, multiple equilibria open up a gap in game theory.Footnote 1 Saying that individuals will tend to converge on the equilibrium that is “focal” is just a catchall term for all the nonmathematical features of the situation that economics has no particular theory for understanding. In my view, not only is this gap a place where L&S theory could be accommodated; explaining expectations based on “precedent” (history) and “who the parties are and what they know about each other” (part of culture) is exactly what much L&S scholarship already does. When L&S theorists discuss schema, frames, “legal consciousness,”“cultural software,” the cultural “tool kit,” and other patterns of thought (e.g., Reference BalkinBalkin 1998; Reference Ewick and SilbeyEwick & Silbey 1998; Reference SwidlerSwidler 1986), they are describing “non-payoff factors” that determine what individuals will expect in situations where the payoffs do not fully determine how individuals will behave. Thus, it is not the game theorists but other social scientists who have been doing most of the empirical work on “focal points” that Schelling urged.
Coordination situations are interesting not only because L&S potentially has more theory than economics for predicting what will emerge, but also because once a particular equilibrium does emerge, economics recognizes its contingency. In evolutionary game theory, small and arbitrary differences in early iterations of a recurrent game—e.g., the first person to refer to object X used symbol A, or the first time a class of dispute arose the person with one ascriptive trait “gave in” to a person of another trait—can have large consequences by putting the system on the path to a particular equilibrium. The particular coordination equilibrium that emerges when more than one is possible is therefore “path dependent.” (Reference ArthurArthur 1994; Reference DavidDavid 2000). Consequently, when the same coordination situation arises in different places among different populations, there is no reason to expect the same solution to emerge. Just as game theory does not expect every culture to drive on the same side of the road or to use the same language, game theory embraces the idea of cultural variation across a wide variety of problems.
Most important, game theorists understand that cultural variation will build on itself. When a new coordination situation arises within a society, nonpayoff features that are likely to determine what equilibrium emerges are the patterns of thought (schema, frames, etc.) borrowed (via analogical or metaphorical reasoning) from other coordination situations in that society. Thus, culture determines what is “focal,” which determines expectations, which determines behavior in a coordination game. Culture is the sum of past mechanisms of coordination and also the source for future coordination mechanisms.
The economic literature exploring this implication of coordination games is modest but growing. Theorists beginning with Reference Kreps, Alt and ShepsleKreps (1990) use coordination to explain the divergent “corporate culture” of firms. Individuals working together in an organization face recurrent problems of coordination and, over time, learn techniques for resolving them that differ from the techniques created in other organizations to address the same problems. See Reference Hermalin, Cooper, Cartwright and EarlyHermalin 2001. In one recent experiment, researchers induced subjects in a laboratory “firm” to create their own private language as a way of solving coordination problems more quickly (Reference Weber and CamererWeber & Camerer 2003). Others use coordination to address social culture. Reference GreifGreif (1994) explains how different “cultural beliefs”—individualist and collectivist—led two premodern societies, the Genoese and Maghribis, to develop different trading practices and institutions for solving an iterated principal-agent problem with multiple equilibria. The movement to different paths explains larger subsequent divergences, including the differential development of legal institutions. Reference ChweChwe (2001) explains the form social rituals take by their ability to generate the “common knowledge” needed to solve a coordination game. Reference ChongChong (2000) uses coordination games to explain political conflict over cultural values. Some of my own collaborative work uses coordination to explain how law works expressively (i.e., independent of legal sanctions) by making focal particular outcomes of coordination situations (Reference McAdams and NadlerMcAdams & Nadler 2003).
There is potential here for bridge-building. Even if this game theory fails to capture what Edelman means by the “social embeddedness” of market practices, an interesting research agenda would be to examine the precise differences between that concept and the economic ideas just described.
The Limits of Contingency in Explaining the Economy
Having identified an unexpected point of convergence, I want to acknowledge a remaining divergence and pose a challenge to L&S scholars. While economics is somewhat open to claims of contingency, is L&S at all open to the claim of noncontingency? Specifically, can L&S embrace the possibility that certain economic practices persist because they serve the interests of the individuals affected by them? I am setting aside the possibility of inequality mentioned above—that a practice persists because it further the interests of the powerful who impose it on the weak—and asking if it is not also possible that an economic practice persists because it is generally useful.
Of course, to avoid naive functionalism, one must explain how the benefits the practice produces actually motivate individuals to adopt or maintain the practice. Even if a practice is useful, it may not exist because it is useful. A feedback mechanism is essential (Reference HardinHardin 1995:82–86). L&E scholars, being methodological individualists, try to tie the existence of a social practice to the incentives it creates for individuals whose actions bring the practice about or cause it to persist. Take, for example, the choice Edelman mentions between bargaining and accepting a posted price. An economist would offer something like this: Because bargaining consumes time, its costs may move busy individuals toward posted pricing for items of low price and standardized quality.
The skeptic of this sort of functional explanation might claim, however, that individuals in grocery stores do not consciously consider the reasons they accept posted prices in that context and not in another. This objection demands a mechanism for explaining how the usefulness of the practice causes its existence, given that most people seem oblivious to its usefulness most of the time. Here is one answer, inspired by Reference Pettit and MakiPettit (2001). Function sometimes explains the origin of a practice less than the resiliency of the practice once it arises. People may not reflect on what works, but they often think about what fails. If posted pricing for groceries doesn't serve the interests of buyers and sellers, there is no reason to expect the practice to persist. It might disappear in a slow, random drift to a different practice, or succumb to an external shock in the market. Suppose, however, that posted pricing does generally serve individual interests in this context for reasons I identified, and yet a particular grocery store tries to change the practice by haggling over price. Now people may reflect for the first and perhaps only time on why it is better for them not to bargain in this context. They might realize that any money saved in bargaining over individual grocery items of standardized quality is not worth the time it takes. If so, then the experiment fails; the drift does not occur. With just a little conscious reflection, the function of the practice explains its resilience. A similar point could be made from the side of the grocer—that the business practices that best serve sellers are most likely to persist, even if most grocery managers rarely reflect on why they price the way they do (Reference SmithSmith 1991:890–93).
Economics finds that many features of the economy are explained in part by their function, so understood. Bridge-building is possible even if L&E (over)emphasizes function and L&S (over)emphasizes contingency, so long as both schools remain open to the possibility of both types of explanation. I read Edelman as allowing this position. When she explains pricing behavior as the product of “culturally ingrained practices as well as, and sometimes instead of, preference maximization” (2004:192, emphasis added), she suggests an openness to function because sometimes it is about preference maximization. If so, then L&E and L&S scholars might collaborate on better understanding the precise combination of contingency and function for any given economic practice (including the common possibility that certain practices are functional given the prior existence of certain contingent practices). L&S scholars may offer game theorists a better understanding of the cultural influences that make a particular outcome “focal” and likely to occur, in situations of contingency. At the same time, L&E may offer a better understanding of certain noncontingent aspects of economic practice. I have no illusions about the differences that remain, but the possibility of new forms of collaboration is real.
Lauren Edelman's presidential address is artfully provocative. Given my membership in both the Law & Society Association and the American Law & Economics Association, I endorse her twin suggestions that law and society (L&S) scholars give more attention to the economy and that L&S and law and economics (L&E) scholars consider building bridges between their divergent methods. But there is a tension in the way Edelman frames these recommendations. Edelman proposes that L&S spend more time studying the economy partly because she finds the existing study, dominated by L&E, so inadequate. One might wonder why L&S should seek any reconciliation or interaction with economic methodology if it does not offer something that L&S currently lacks. I respond here to Edelman's respectful critique by attempting to demonstrate that economics has sufficient value to L&S as to make worthwhile the efforts at collaboration. Indeed, I will provocatively assert that, because L&S is, methodologically speaking, a “big tent,” the difference between alternative branches of L&S methodology within that “tent” is as great as the difference between the more positivist branches of L&S scholarship and certain elements of L&E scholarship. In any event, I seek to identify a specific and unexpected convergence between the schools, where the theoretical topography is best suited for building collaborative bridges.
There are several points of possible convergences other than the one I will address. Worth noting, for example, are the trends in economics toward greater complexity of behavioral assumptions and greater empiricism. Many economists are replacing the more reductionist assumptions of homo economicus with models of boundedly rational and boundedly selfish individuals, based on the steady stream of psychological and economic experiments finding powerful evidence of cognitive biases and nonselfish motivations, and the importance of context to each (Reference CamererCamerer 2003; Reference Fehr and FischbacherFehr & Fischbacher 2002; Reference Korobkin and UlenKorobkin & Ulen 2000). Another trend—weaker than I would like—is a move toward the study of inequality. Perhaps the first to recognize the potential for using game theory to explore issues of inequality was the philosopher Reference Ullmann-MargalitUllmann-Margalit (1977:134–97). But there is increasing interest in using games to model social inequalities, especially those related to sex. See Reference RoseRose (1992), Reference HadfieldHadfield (1999), and Reference WaxWax (2000).
My primary focus, however, is on the contrast Edelman draws between L&E and L&S on the grounds of what I will term contingency. Edelman says that L&S uniquely emphasizes “the social embeddedness and politics of markets” (Reference EdelmanEdelman 2004:183), “the social, political, and legal construction of rational economic behavior” (2004:184, emphasis in original), and “social action as responsive to institutions, norms, and historical context” (2004:187). Addressing one aspect of the market, for example, Edelman asks “why we bargain at the car dealership or the flea market but not at the pharmacy or the grocery store” (2004:192). She answers that pricing practices “are governed by culturally ingrained practices as well as, and sometimes instead of, preference maximization” (2004:192). Although Edelman may mean other things by this contrast I believe she means also to emphasize contingency: that there was no necessity to arriving at the existing pricing practices, but that arbitrarily small changes in history and culture could produce very different patterns of contemporary behavior. We can understand pricing practices and other economic behavior only by understanding the history that has led us to them.
Here is where I wish to sketch out a possible ground of unexpected convergence. To a degree not fully appreciated, economic theory embraces the idea of contingency, and with it, the importance of history and culture. After describing how contingency arises in game theory, I return to a source of divergence: whether L&S can embrace the converse—that the noncontingent practicality or “function” of market practices sometimes explains their persistence better than contingency.
Game Theory Embraces Cultural Contingency
Economics appears to reject contingency because the one-shot prisoners' dilemma game and its variants are such popular theoretical devices, and they have only one possible “equilibrium.” A Nash equilibrium is an outcome at which no individual can benefit by unilaterally changing strategies; each is playing his or her best response to what the others are doing. Because an equilibrium is solely a function of the individual payoffs, which the economist derives from “stable preferences,” a single-equilibrium game is the antithesis of socially constructed or culturally determined values or rationality. And so L&E and L&S theorists quickly reach an impasse.
There is a way to sidestep this impasse, at least in part. Most games have multiple equilibria. That is, even granting fixed preferences and all other standard assumptions of game theory, there are situations where the payoffs do not fully determine the behavior. Consider three examples. Before a society determines what side of the road its cars should travel on, potential drivers face a pure coordination game. Each prefers to drive on the right if most other drivers do, but to drive on the left if most others do. Language usually presents a coordination game because people usually prefer to use symbol A to refer to object X if others generally do, but to use symbol B to refer to object X if others generally do. Even disputes can involve coordination. Two disputants may each prefer that the other be the one to “give in,” but each may prefer to “give in” rather than endure the conflict that results if neither gives in. Thus, when the costs of conflict over the two equilibria (either one giving in to the other) are sufficiently high, disputants want to coordinate by avoiding conflict (Reference McAdamsMcAdams 2000). In all these examples, each individual's best course of action depends on what he or she expects the others to do. But if more than one equilibrium is possible, then here's the rub: the first time the game is played, no one knows what to expect the others to do.
If payoffs don't determine initial expectations, then something else might (in some cases, must). Decades ago, Reference SchellingSchelling (1960) noted that non-payoff features may influence behavior by making a particular equilibrium salient or “focal” to the players. People tend to play the strategy associated with the one equilibrium that they expect others to recognize as “standing out” from the others. For example, people who get lost from one another often coordinate by meeting at the place each will expect to be conspicuous to the other. In four decades since Schelling, game theory has made many strides, but it has not made much progress in understanding focal points, that is, in understanding how expectations emerge independent from payoffs. This should come as no surprise because, as Schelling pointed out, what is focal “may depend more on imagination than on logic, more on poetry or humor than on mathematics” (1960:57). Schelling also noted the inherently empirical quality of what is focal, depending as it does on “precedent” and on “who the parties are and what they know about each other” (1960:97).
Thus, multiple equilibria open up a gap in game theory.Footnote 1 Saying that individuals will tend to converge on the equilibrium that is “focal” is just a catchall term for all the nonmathematical features of the situation that economics has no particular theory for understanding. In my view, not only is this gap a place where L&S theory could be accommodated; explaining expectations based on “precedent” (history) and “who the parties are and what they know about each other” (part of culture) is exactly what much L&S scholarship already does. When L&S theorists discuss schema, frames, “legal consciousness,”“cultural software,” the cultural “tool kit,” and other patterns of thought (e.g., Reference BalkinBalkin 1998; Reference Ewick and SilbeyEwick & Silbey 1998; Reference SwidlerSwidler 1986), they are describing “non-payoff factors” that determine what individuals will expect in situations where the payoffs do not fully determine how individuals will behave. Thus, it is not the game theorists but other social scientists who have been doing most of the empirical work on “focal points” that Schelling urged.
Coordination situations are interesting not only because L&S potentially has more theory than economics for predicting what will emerge, but also because once a particular equilibrium does emerge, economics recognizes its contingency. In evolutionary game theory, small and arbitrary differences in early iterations of a recurrent game—e.g., the first person to refer to object X used symbol A, or the first time a class of dispute arose the person with one ascriptive trait “gave in” to a person of another trait—can have large consequences by putting the system on the path to a particular equilibrium. The particular coordination equilibrium that emerges when more than one is possible is therefore “path dependent.” (Reference ArthurArthur 1994; Reference DavidDavid 2000). Consequently, when the same coordination situation arises in different places among different populations, there is no reason to expect the same solution to emerge. Just as game theory does not expect every culture to drive on the same side of the road or to use the same language, game theory embraces the idea of cultural variation across a wide variety of problems.
Most important, game theorists understand that cultural variation will build on itself. When a new coordination situation arises within a society, nonpayoff features that are likely to determine what equilibrium emerges are the patterns of thought (schema, frames, etc.) borrowed (via analogical or metaphorical reasoning) from other coordination situations in that society. Thus, culture determines what is “focal,” which determines expectations, which determines behavior in a coordination game. Culture is the sum of past mechanisms of coordination and also the source for future coordination mechanisms.
The economic literature exploring this implication of coordination games is modest but growing. Theorists beginning with Reference Kreps, Alt and ShepsleKreps (1990) use coordination to explain the divergent “corporate culture” of firms. Individuals working together in an organization face recurrent problems of coordination and, over time, learn techniques for resolving them that differ from the techniques created in other organizations to address the same problems. See Reference Hermalin, Cooper, Cartwright and EarlyHermalin 2001. In one recent experiment, researchers induced subjects in a laboratory “firm” to create their own private language as a way of solving coordination problems more quickly (Reference Weber and CamererWeber & Camerer 2003). Others use coordination to address social culture. Reference GreifGreif (1994) explains how different “cultural beliefs”—individualist and collectivist—led two premodern societies, the Genoese and Maghribis, to develop different trading practices and institutions for solving an iterated principal-agent problem with multiple equilibria. The movement to different paths explains larger subsequent divergences, including the differential development of legal institutions. Reference ChweChwe (2001) explains the form social rituals take by their ability to generate the “common knowledge” needed to solve a coordination game. Reference ChongChong (2000) uses coordination games to explain political conflict over cultural values. Some of my own collaborative work uses coordination to explain how law works expressively (i.e., independent of legal sanctions) by making focal particular outcomes of coordination situations (Reference McAdams and NadlerMcAdams & Nadler 2003).
There is potential here for bridge-building. Even if this game theory fails to capture what Edelman means by the “social embeddedness” of market practices, an interesting research agenda would be to examine the precise differences between that concept and the economic ideas just described.
The Limits of Contingency in Explaining the Economy
Having identified an unexpected point of convergence, I want to acknowledge a remaining divergence and pose a challenge to L&S scholars. While economics is somewhat open to claims of contingency, is L&S at all open to the claim of noncontingency? Specifically, can L&S embrace the possibility that certain economic practices persist because they serve the interests of the individuals affected by them? I am setting aside the possibility of inequality mentioned above—that a practice persists because it further the interests of the powerful who impose it on the weak—and asking if it is not also possible that an economic practice persists because it is generally useful.
Of course, to avoid naive functionalism, one must explain how the benefits the practice produces actually motivate individuals to adopt or maintain the practice. Even if a practice is useful, it may not exist because it is useful. A feedback mechanism is essential (Reference HardinHardin 1995:82–86). L&E scholars, being methodological individualists, try to tie the existence of a social practice to the incentives it creates for individuals whose actions bring the practice about or cause it to persist. Take, for example, the choice Edelman mentions between bargaining and accepting a posted price. An economist would offer something like this: Because bargaining consumes time, its costs may move busy individuals toward posted pricing for items of low price and standardized quality.
The skeptic of this sort of functional explanation might claim, however, that individuals in grocery stores do not consciously consider the reasons they accept posted prices in that context and not in another. This objection demands a mechanism for explaining how the usefulness of the practice causes its existence, given that most people seem oblivious to its usefulness most of the time. Here is one answer, inspired by Reference Pettit and MakiPettit (2001). Function sometimes explains the origin of a practice less than the resiliency of the practice once it arises. People may not reflect on what works, but they often think about what fails. If posted pricing for groceries doesn't serve the interests of buyers and sellers, there is no reason to expect the practice to persist. It might disappear in a slow, random drift to a different practice, or succumb to an external shock in the market. Suppose, however, that posted pricing does generally serve individual interests in this context for reasons I identified, and yet a particular grocery store tries to change the practice by haggling over price. Now people may reflect for the first and perhaps only time on why it is better for them not to bargain in this context. They might realize that any money saved in bargaining over individual grocery items of standardized quality is not worth the time it takes. If so, then the experiment fails; the drift does not occur. With just a little conscious reflection, the function of the practice explains its resilience. A similar point could be made from the side of the grocer—that the business practices that best serve sellers are most likely to persist, even if most grocery managers rarely reflect on why they price the way they do (Reference SmithSmith 1991:890–93).
Economics finds that many features of the economy are explained in part by their function, so understood. Bridge-building is possible even if L&E (over)emphasizes function and L&S (over)emphasizes contingency, so long as both schools remain open to the possibility of both types of explanation. I read Edelman as allowing this position. When she explains pricing behavior as the product of “culturally ingrained practices as well as, and sometimes instead of, preference maximization” (2004:192, emphasis added), she suggests an openness to function because sometimes it is about preference maximization. If so, then L&E and L&S scholars might collaborate on better understanding the precise combination of contingency and function for any given economic practice (including the common possibility that certain practices are functional given the prior existence of certain contingent practices). L&S scholars may offer game theorists a better understanding of the cultural influences that make a particular outcome “focal” and likely to occur, in situations of contingency. At the same time, L&E may offer a better understanding of certain noncontingent aspects of economic practice. I have no illusions about the differences that remain, but the possibility of new forms of collaboration is real.