Cooperation & Conflict in the Patriarchal Labyrinth

This essay offers a new way of visualizing structures of collective power based on gender, emphasizing the role of social institutions in shaping women's ability to bargain over the distribution of the gains from cooperation with men. It makes the case for an interdisciplinary conceptualization of bargaining power that emphasizes the role of imperfect information and inefficient outcomes, and explains important parallels between structures of collective power based on gender, age, and sexuality, and those based on other dimensions of socially assigned group membership such as race, ethnicity, citizenship, and class. Recognition of the importance of reproductive work helps advance the project of developing intersectional political economy.

C onsider the term "power structure." Social scientists often refer to hierarchical structures, sets of institutions, or economic arrangements that frame the environment in which people live and work. Both power and hierarchy imply a structure of pyramidal shape, with big groups at the bottom and small groups at the top. Yet the implied pyramids tend to be simple and two-dimensional drawings on a page. Their internal spaces as well as their relationships to one another remain largely unspecified, making it difficult to understand how people operate within them. In this essay, I develop a more complex visual metaphor: threedimensional pyramids of power with internal paths that allow some opportunity for those within them to climb upward but put greater obstacles in the way of others.
A patriarchal labyrinth describes risks and hurdles that vary by gender, age, and sexuality. It helps explain why, as psychologists Alice Eagly and Linda Carli have put it, "women's paths to power remain difficult to traverse." 1 Such institutional structures vary considerably over time and space and are vulnerable to renegotiation and redesign. They typically overlay and intersect other labyrinthine hierarchies based on different dimensions of group inequality. Together, they create complex strategic environments in which both women and men try to defend or advance their position and, sometimes, to modify the institutional environment in their favor. Such efforts, however, are hampered by the unpredictability of the labyrinth itself and the shifting alliances that affect the success of collective efforts to traverse it.
199 Nancy Folbre This visual image illustrates the ways in which patriarchal institutions constrain the choices that women make, reducing their ability to negotiate obstacles. The word "negotiate" is key here, because progress toward power and economic security is not like some athletic contest in which outcomes are determined in large part by individual prowess, like slogging through mud or climbing a rope. Rather, it is a team sport in which individual performance matters, but team or group membership shapes bargaining power: the ability to influence the rules and outcomes of the game. Progress upward in the labyrinth requires cooperation with others, but the gains from cooperation are often unequally distributed. Women are assigned greater responsibilities for the creation and maintenance of human capabilities than men are, and these responsibilities tend to reduce their bargaining power. The path to gender equality requires the development of new institutions to ensure more equitable sharing of the costs of caring for dependents.
This theoretical argument contends that global gains in women's legal rights over the last century have weakened but not demolished patriarchal institutional structures. Combining institutional analysis with bargaining logic, I explain how any group that is able to claim a disproportionate share of the gains from cooperation can develop social institutions to fortify their position. The balance of gendered power can shift for a variety of reasons, including changes in the division of labor wrought by economic development and change. However, structures of patriarchal power are reinforced by their intersections and overlaps with other labyrinthine hierarchies based on race, citizenship, and class. A symmetry in human rights represents asymmetry in paths to economic equality. Recent feminist discourse has successfully expanded a basic human rights framework to insist on its application to the realms of gender, the family, and sexuality. 2 This is no mean feat and it merits great appreciation. Research published by multilateral institutions such as the Organisation for Economic Co-operation and Development (OECD) and the World Bank have helped expose laws and public policies that are inconsistent with liberal principles of equal opportunity. 3 In the more academic realm, sociologist Goran Therborn has exhaustively documented the global retreat of patriarchal law (acknowledging that some enclaves remain almost untouched). 4 Even from the vantage point of the United States and other affluent countries, however, this celebration seems premature. While women have long enjoyed legal rights largely commensurate with those of men, they remain economically and politically disadvantaged. Empirical research shows that women earn considerably less than men largely because they take more responsibility for the care of children and other dependents, incurring "motherhood penalties." 5 Women also remain concentrated in occupations and industries that provide care services, often incurring "care penalties" as a result. 6 200 Daedalus, the Journal of the American Academy of Arts & Sciences

Cooperation & Conflict in the Patriarchal Labyrinth
Some such research implies that women simply have different preferences than men do, choosing to sacrifice earnings in return for the emotional rewards of care. 7 Many feminist scholars, however, advocate a more structural view, noting that individual preferences are shaped by social norms and economic circumstances that are, in turn, shaped by political and economic power. 8 Some invoke patriarchy, or capitalism, or patriarchal capitalism; others point to complex forms of intersectional power. 9 I believe that these approaches can be effectively synthesized by more attention to specific institutional structures that both reflect and advance collective interests. 10 T he vast literature on social institutions offers many important but diffuse insights. Institutions can be categorized in a variety of ways, according to sites (such as state, market, and family), functions (such as production, distribution, and socialization), and social science domains (such as political, economic, and cultural, corresponding roughly to legal governance, resource allocations, and cultural norms). All of these categorizations can serve useful purposes.
However, emphasis on collective conflict leads me to categorize institutional structures in terms of their distributional consequences: sets of institutions that shape access to gains from cooperation in unequal ways based on some dimension of socially assigned group identity, such as gender, sexual orientation, age, class, race, ethnicity, or citizenship. Rather than attributing gender inequality to a unique set of institutions and class inequality or race inequality to another unique set, I allow for the possibility that many types of institutions can separately or collectively reinforce many dimensions of group inequality.
Both obstacles and shortcuts in paths upward toward greater relative power and economic security can be labeled according to the way they filter and redirect members of specific groups, even though they can affect more than one group at a time. The important questions are who benefits, how, and how much? This approach departs from mainstream economic theory, which interprets social institutions as efficient solutions to coordination problems that cannot be easily resolved by decentralized individual choices (that is, markets). 11 Distributional conflict fades into the background. 12 Efficiency also implies transparency, or at least sufficient light in the corridors to be able to see the best path upward.
Some political scientists offer a more nuanced explanation of institutions, pointing to the impact of both distributional conflict and political power. 13 Yet here, too, capacities for rational choice and cost-benefit analysis are often taken for granted. Popular game-theoretic models stipulate payoff matrices and/or fallback positions in advance, as though all the players clearly see the consequences of their actions. In a labyrinth, by contrast, players may be quite rational but nonetheless quite vulnerable to taking wrong turns; the possible paths are confusing and official signposts misleading. The resulting uncertainty reinforces a tenden-201 Nancy Folbre cy to minimize risk and conform to precedent even when this does not lead to the best possible outcome.
The social institutions that constitute hierarchical structures reflect multilayered processes of bargaining over the level and distribution of gains from cooperation. Naming such structures by their distributional outcomes-who they most benefit-links the burgeoning literature on patriarchal institutions to structural concepts such as patriarchy, capitalism, racism, and nationalism. Yet it challenges the notion that such structures are independent or autonomous, defining them instead as structures of collective power constituted by social institutions: a jagged mountain range of hierarchical labyrinths. E fforts to follow a convoluted path or climb an economic ladder are obviously affected by legal and political rights. However, they are also affected by social obligations. For instance, those who bear the burdens of financial support and direct care of dependents often find it difficult to compete with those who are unencumbered. Tensions between gender equality and recognition of gender difference-a manifestation of tensions between rights and obligations-are deeply embedded in the history of feminist activism in the United States, reflecting the complementary but distinct priorities of liberal and socialist feminisms.
Legal restrictions on women's rights leave a particularly clear historical trail, offering an evidentiary advantage over research on other types of institutions. Yet the distribution of social obligations is equally important. Even when women freely choose to devote more time and energy to the care of others than men do, they do not choose the economic consequences. Both patriarchal and capitalist institutional structures enable people in general and men in particular to free ride on caregivers.
Capitalist employers do not reward activities that do not generate a profit, whatever the public or social benefits such activities create. Within capitalist structures of collective power, those who devote less time and effort to unprofitable activities than others have commensurately greater access to earnings and wealth. This asymmetry means that partnerships with men offer women important economic benefits, even when these partnerships render them vulnerable to the threatened withdrawal of support. Economic dependence reduces women's ability to bargain over the distribution of care responsibilities in the home, the workplace, and the polity. It also reduces their ability to defy or modify patriarchal norms.
While patriarchal institutional structures have disempowered women, they have also imposed significant forms of social obligation on men. In many traditional patriarchal societies, strong community sanctions and informal rules-if not always specific laws-required that a man provide support for a woman he impregnat-  and take economic responsibility for his offspring. Likewise, support for aging parents, including widows, was often effectively enforced. As women have gained new rights, however, men have often offloaded some of their responsibilities. The breakdown of the "shotgun marriage" rule in the United States, for instance, contributed to an increase in the percentage of families maintained by women alone. 14 The expansion of women's access to paid employment has often been accompanied by poor specification of women's rights to remuneration for family care from their children's fathers, their adult children, and more broadly, from the state. Increased family instability offers women more scope for individual choice, but less reliable networks of support. The distribution of income and leisure within married couples often benefits men, but co-residence typically guarantees some commonality in living standards. Nonmarriage, separation, divorce, and longdistance migration in search of individual employment tend to reduce family income pooling, particularly costly to mothers of young children.
Access to market income sometimes comes at the expense of intrafamily transfers, which remain poorly measured and largely ignored by most standard measures of family welfare. The tally of patriarchal laws that have been-or need to be-stricken from the books should be accompanied by a list of new laws needed to encourage more equal distribution of the costs of caring for dependents and investing in the development of human capabilities. T he labyrinth metaphor leaves ample room for the role of individual and group agency: All enter the structure at different starting points, some more advantageous than others. Still, effort, skill, and teamwork affect their probability of avoiding the hazards and reaching the prizes within. Everyone may engage in efforts to clear their own path, or to close paths to others. The set of choices that individual and collective agents face, however, is far more complex than that implied by economic models of utility maximization that assumes perfect information, exogenously given preferences, and sovereign self-interest.
Individual agents are, from the outset, socially assigned to groups. As children they cannot choose their gender, age, sexual orientation, class background, race and ethnicity, citizenship, or a host of other collective identities that influence the preferences they form and the capabilities they develop as well as the opportunities available. As they mature, agents gain scope for self-awareness and rational choice, and their most important choices concern the strength of allegiances to the various groups to which they are assigned, which may be a source of intrinsic satisfaction as well as a determinant of their future economic success. The consequences of such choices, however, will always remain uncertain, because they are affected by the simultaneous choices of others.
In other words, the scope for rational decisions by individuals is limited not only by the constraints imposed by institutional structures, but also by the dif-203 Nancy Folbre ficulty of coordinating enforceable decisions with other independent agents or forming stable alliances with them. This latter difficulty helps explain why institutional labyrinths are resistant to change: even the most disadvantaged participants derive some benefit from them if the only alternative is to exit to an even more dangerous, unpredictable, and unstructured environment. Nonetheless, the threat of an exit by the disadvantaged remains significant, exercising some influence on the actions of those who benefit from their cooperation. T he concept of cooperative conflict plays a central role in John Rawls's A Theory of Justice and has been applied persuasively to gender bargaining in households by economist Amartya Sen and others. 15 Distinct groups or individuals may have much to gain from cooperation with one another, but nonetheless struggle over the distribution of the gains from cooperation. Philosopher Charles Mills has correctly observed that the term "cooperative conflict," applied, for instance, to institutions such as slavery, is far too benign. 16 As Catharine MacKinnon has pointed out in her discerning critique of the concept of "consent" to sex, hierarchical institutional structures can lead to something that could be more aptly described as coerced cooperation. 17 What we call "freedom" depends largely on the existence of viable alternatives to subordination.
Within the economics discipline, attention to bargaining between men and women within households has now largely displaced microeconomic models that treat the household itself as a unitary decision-maker. Yet most bargaining models retain a focus on individuals operating in a social environment that is taken as a given, ignoring the incentives for joint efforts to challenge or to reinforce political, cultural, and economic institutions. A broader approach to bargaining that brings collective identity and action into the picture helps explain how it both shapes and is shaped by hierarchical structures. The complexities of bargaining in a dark labyrinthine environment also deserve explicit consideration. C ooperation can take many different forms, ranging from agreement to participation in a market exchange to willingness to follow orders. The distribution of the gains from cooperation is not determined by each person's "value added," but by their fallback position: what happens to them if cooperation breaks down. A credible threat of physical violence weakens fallbacks, as does lack of independent access to income or withdrawal of economic support for the care of dependents.
One of the early pioneers of household bargaining models, Marjorie Mc-Elroy, has observed that social institutions, or what she calls "extra-environmental household parameters," could affect the fallback position of household members. 18 For instance, divorce law, access to public assistance, and cultural norms influence the relative consequences of marital dissolution for husbands and  19 Empirical research substantiates this important linkage between social institutions and microeconomic outcomes. For instance, economists Shelly Lundberg and Robert Pollak have shown that a policy innovation in the United Kingdom that directed child allowances to mothers rather than fathers altered the allocation of household spending. 20 Applying a qualitative, rather than quantitative approach, economist Bina Agarwal has persuasively explained the impact of cultural norms on household bargaining outcomes in developed countries. 21 Legal institutions have particularly clear implications for household bargaining. Until well into the twentieth century, for instance, U.S. law obligated a husband to meet the subsistence needs of his wife and children, but not to share family income equally with them. 22 Today, family law in the United States enforces a responsibility for the mutual support of spouses, but does not require equal sharing (though in some states, community property laws dictate equal sharing of wealth acquired during marriage in the event of divorce). In many countries, men historically enjoyed the right to physically abuse their wives, as long as they caused no lasting injury. Even after these rules were reformed, protection against domestic violence within marriage remained largely ineffective until quite late in the twentieth century.
The threat of physical abuse of some women by some men exerts a disciplinary influence on all women, making them more appreciative of (and dependent on) nonabusive men. Likewise, the extreme physical abuse of many slaves in the American South, however costly to their owners, created an implicit threat that represented a kind of social externality for slaveowners who were less abusive, allowing them to don a mantle of benevolence. In both cases, members of powerful groups have an incentive to avert their eyes from abuses of power that they would never directly perpetrate.
Bargaining can, and often does, take more subtle forms. Fallback positions are not limited to exit options. Individuals who believe they are being treated unfairly are particularly likely to shirk or engage in sabotage, reducing the gains from cooperation in ways particularly costly to those who benefit the most from them. In the context of household bargaining, this strategy can be labeled "burnt toast." The extensive literature on efficiency wages in capitalist firms explains why employers may benefit from paying a higher-than-market clearing wage: the resulting cost of job loss creates an incentive for workers whose performance is difficult or costly to monitor to increase their effort. 23 Similarly, a man may offer a woman a higher-than-necessary share of the gains from cooperation, because this increases her effort in or fealty to the partnership. Altruistic sentiments and personal affections also soften bargaining. What economists refer to as "gift exchange" between some employers and their workers is far more likely to take place with-

(1) Winter 2020
205 Nancy Folbre in households than within capitalist firms. 24 Yet it cannot always be relied upon. Women's vulnerability to inequality within households governed by patriarchal rules results in part from the fragile quality of emotional attachments. T he effects of social institutions on household bargaining have been more thoroughly explored than the causal arrows that run the other direction, because these are more difficult to pin down. Yet the links between the micro-and macrolevels are obvious: if patriarchal institutions affect the relative bargaining position of men and women within households, then men and women have incentives to engage in collective efforts to modify those institutions in their favor.
This approach subsumes both the neoclassically influenced concept of rentseeking and the Marxian theory of class conflict under a larger rubric. 25 Group success often leads to the consolidation of institutional power, which can, in turn, increase a group's share of gains from cooperation. A person or group in an initially strong fallback position can capture a large share of the gains, then invest those gains in efforts to improve their fallback position or weaken the fallback position of others. Once a group occupies the top of a hierarchical institutional structure, they can use their leverage to develop political, economic, and cultural institutions that preserve their advantage.
This dynamic helps explain a dialectic between power and efficiency that contributes to the persistence of social institutions that are inefficient as well as unfair. Sometimes, innovations that could increase the size of the "social pie" are blocked because they might reduce the share of powerful groups. A big slice of a small pie can be more desirable than a much smaller piece of a slightly larger pie.
Capitalist development sometimes delivers opportunities for women to earn wages outside the home that can potentially yield greater benefits to their households than their nonmarket work. However, husbands may be reluctant to let them take advantage of such opportunities because the potential reduction in their share of the total benefits may outweigh increases in the overall size of those benefits. 26 Similar reasoning explains why slaveholders in the United States prohibited the education of slaves, even though such education could have improved the productivity of the slave-based economy. In both of these examples, the structure of collective power impedes institutional changes that could potentially increase efficiency. E mphasis on individual and collective bargaining does not imply rational choices informed by systematic cost-benefit analysis. There are no round oak tables in hierarchical labyrinths at which agents sit down to make offers and counteroffers. Still, most forms of collaboration are influenced to some degree by implicit bargaining. As advertisements for a prominent business train-  27 Threats and promises, fakes and feints, persuasion and coercion, coalition and compromise often take implicit forms. Even market exchange is often contested. 28 Process may matter as much as payoffs: some people are trained to bargain harder than others. Payoffs are seldom perfectly clear and may often be misperceived or unknown (what military strategists refer to the "fog of war" also applies to the cloud of sex).
Explicit bargaining is not only costly and time-consuming, it can also create resentment and ill will. Economists typically assume that married partners will not try to block any move that improves their partner's outcome, as long as it comes at no cost to them. 29 In the real world, however, spite often trumps reason, invites retaliation, and leaves both bargainers worse off than they were before. Some men kill their partners, then commit suicide.
Social norms can lubricate cooperation by offering implicit rules, and sometimes explicit solutions, to costly forms of disagreement. How best to divide a pie? "You slice, and I'll choose." Want to avoid a fight? "Let's flip a coin." If a husband and wife cannot agree on how to divide tasks they may simply fall back to social expectations. 30 Yet social expectations tend to favor those in already favored positions. 31 Norms of appropriate femininity can reinforce gender inequality, just as patriotic norms can justify aggression against countries, racial pride can fuel White supremacy, and elitist values can legitimate class disparities.
Altruistic norms are often internalized in ways that affect individual preferences and perceptions. If one bargainer cares more about the other (or about third parties who may be affected), bargaining outcomes will be skewed in favor of the less altruistic or more assertive member, as suggested by the phrase "nice guys (and nice gals) finish last." In a generalization of the "hidden injuries" of class, the subordination of women weakens their individual and group agency. 32 Heteronormative values as well as homophobic attitudes can undermine the confidence of those labeled deviant. Oppression itself is often internalized. 33 C ooperative processes in which individuals cannot necessarily identify or claim their own value added characterize many aspects of economic life, but play a particularly important role in processes of reproduction (defined here as the creation and maintenance of human capabilities) and social reproduction (defined here as the creation and maintenance of social groups). Since women devote relatively more time and energy to these tasks than men do, this makes institutional bargaining particularly relevant to the analysis of gender inequality.
The specific demands of reproduction have direct implications for the bargaining power of those who specialize in it. Responsibility for the care of dependents often renders women dependent on men to meet their own subsistence needs. The emotional attachment to others that is intrinsic to caregiving also weakens 149 (1) Winter 2020 207 Nancy Folbre fallback positions. 34 As a result, declining specialization in reproductive tasks has contributed to women's empowerment. Two very long-run historical trendsfertility decline and technological changes favoring intellectual capabilities over physical strength-have increased women's collective bargaining power. Yet institutional resistance to gender equality remains significant, in part because such equality requires a redistribution of the costs of caring for others. R eproduction creates and maintains human capabilities of enormous economic value. Care for children is sometimes referred to as investment, and care for the seriously disabled and elderly as consumption, because it is unlikely to generate future economic gains. In standard capital accounting, however, investment covers the costs of depreciation. Both financial support and direct care for the sick and elderly, like that devoted to children, should be considered a form of social investment. Their important insurance function should also be recognized: productive workers are more willing to support dependents when they anticipate reciprocal assistance when they become dependent themselves.
Care providers, however, are seldom able to capture fully a share of the social value they create. Whether unpaid or paid, their contributions are difficult to standardize. Care work typically requires collaboration with others, including family members and other care providers. It often requires cooperation from care recipients themselves: children must heed their parents, patients must follow their doctors' orders, and students must do their homework. The quality of care services often hinges on concern for the well-being of the care recipient. All of these characteristics limit the bargaining power of caregivers. Men, as well as women, can become "prisoners of love," but cultural norms make it easier for men to escape: fathers can ignore or abandon children in part because of their confidence that mothers will fill the breach; adult sons are more likely to provide direct care for their elderly parents if they lack a sister. Gender norms often make women dependent on other women, such as their own mothers or their daughters, for significant assistance with care; not surprisingly, this dependency can make them fearful of change (if women began to act as "carelessly" as men, what would happen to families?). M any trends associated with processes of capitalist development have encouraged a reallocation of women's time and effort away from unpaid care activities in families toward paid employment. The global fertility rate now hovers between two and three children per woman, and in many affluent countries, including the United States, has dropped below the population replacement level of about 2.1 children per woman. 35 Rates of childlessness have also increased in the United States, particularly among highly educated women. ing for young children, improving their access to labor market earnings and enhancing their ability to mobilize for institutional change.
Economic changes on the demand side have also contributed to changes in women's roles. In the United States as in other affluent countries, the growth of the service sector created new labor market opportunities rewarding brains more than brawn. Once considered relatively poor pickings compared with manufacturing jobs, service jobs are now considered less vulnerable to the ups and downs of the business cycle, and they are often difficult to outsource or move offshore. Some economists argue that women have interpersonal or "people skills" that will be richly rewarded by current labor market trends. 36 This optimistic outlook ignores some significant countervailing trends. The decline in children per woman in the United States has been accompanied by an increase in the economic costs that children impose on mothers, driven by new preoccupations with "child quality" and cultural pressure for "intensive mothering" at the top end of the income distribution, and increases in the percentage of children supported by mothers alone at the low end. Further, women's access to high-paying professional and managerial jobs has been restricted by high temporal demands that such jobs impose: long work weeks, flexibility for late-night and weekend work, and availability for travel away from home. At the low end of the labor market, mothers face the opposite problem: part-time shift work and unpredictable hours often leave them with inadequate childcare and inadequate earnings. The growing demands of elder care exacerbate these pressures.
Furthermore, employed women are highly concentrated in the care sector of the U.S. market economy, in health, education, and social service jobs that pay significantly less than jobs in private business services and public administration, controlling for educational credentials. 37 While social skills relevant to management and marketing pay off in the private sector, skills directed at caring or helping people who lack much ability to pay are undervalued in the labor market. Because women in the United States are also disproportionately represented in public sector jobs, budgetary austerity has affected them even more than men.
Today, collective bargaining over the distribution of the costs of reproduction takes place largely within the arena of the welfare state. Women's role in care provision generally makes them more appreciative of social spending than men and helps account for the growing gap in gender voting preferences in the United States. Women would be a more unified and powerful voting bloc were it not for the way other collective interests shape incentives to offload the costs of social reproduction. Here, other labyrinthine institutional structures come into play. E veryone belongs to more than one socially assigned group, and the paths available to them are shaped by many intersecting or overlapping collective identities. The institutional organization of reproduction impinges on Nancy Folbre all groups who hope to persist over time and has implications that reach far beyond interests based on gender, age, and sexuality. Modern welfare states have socialized many of the costs of health, education, and insurance, and typically restrict access to such benefits through immigration controls. As a result, citizenship in a country such as the United States is a marker of significant economic privilege.
Perceived threats to the collective interests of groups based on citizenship and race and ethnicity are often related to social reproduction rather than more narrowly defined economic interests. Congressman Steven King (R-Iowa) has complained that "we can't restore our civilization with other people's babies" and "if we let our birthrate get below the replacement rate, we're a dying civilization." 38 Such complaints provide a rationale for denying women access to reproductive rights in the name of national and racial and ethnic interests: asking them to forgo "selfish" decisions for the sake of Christian civilization.
Class differences in access to care services are also consequential. Highly educated women living in metropolitan areas benefit economically from a large supply of low-wage immigrant women who provide childcare and elder care at a relatively low cost; less-educated women can seldom afford care services that are not publicly provided. Many public universities in the Midwest, starved of state support, have increased admissions from overseas students who can pay full tuition. For instance, at the University of Illinois at Urbana-Champaign, the number of Chinese students has increased by a factor of five since 2008. 39 Such enrollments help subsidize the cost of educating state residents, but also reduce the spaces available to them. T he concept of patriarchal institutional structures offers a guide for tracing the effect of gender on the intersecting and overlapping paths created by multiple group interests. Rather than focusing on one overarching dimension of collective conflict, it emphasizes multiple, complex, and contextdependent interactions. It encourages more attention to institutional specificsand to the organization of reproduction and social reproduction-than conventional economic frameworks. While it leaves room for individual agency, it also emphasizes the impact of the altruism and solidarity that can make groups effective economic actors.
The stylized model of cooperative conflict between women and men draws from the rational-choice tradition of game theory but emphasizes the complexity of a strategic environment that precludes any clear perception of payoffs. The insights that emerge from such models are best considered a guide for historical and comparative analysis, a spool of many different threads that can be used to explore possible escapes from the labyrinths of the past.