Market Hegemony and Economic Theory

It is central to standard economic theory that people act on their interests. People are interested in a variety of things, so a range of values should influence market behavior. When engaged in commerce, however, people generally act for personal gain; the influence of other values usually just disappears in the marketplace. What is missing from the standard account is that people often act on proper subsets of their interests. Economics can, however, be extended to capture this insight.

I n marketplace transactions people generally act for personal gain; other values are reflected only in what they do with those gains. This is no accident. One of the stylized facts about markets is that they are hegemonic: they lead people to act on "bottom line" considerations, that is, to get money, products, services, and so on for themselves. 1 About this fact critics and advocates of markets agree; they disagree only about whether this focus on "bottom line" issues is a good thing. 2 A descriptively adequate economic theory should account for market hegemony. One mainstream view holds that markets are hegemonic because acting on nonmarket values is costly and competition drives out high-cost competitors. Firms attend only to profits because they can't afford to focus on prejudices or social causes. This account, however, isn't good economics. It is central to standard economic theory that markets will reflect consumer attitudes. But people care about a variety of things, including many that aren't self-focused. 3 When people care about, say, labor rights or avoiding racial "inferiors," they will be willing to pay to further those goals. Entrepreneurs, in response, should compete to satisfy the demand for products that advance these goals (e.g., labor-friendly shoes, Caucasian-produced bread). The usual competitive pressure story overlooks consumers' nonpecuniary values and firms' incentives to satisfy those values. Market pressures should lead even profit-maximizing firms to attend to all of the interests consumers have. Standard economic theory, understood fully, suggests that people won't act solely on personal gain when they trade. It misses something here, however. People do tend to focus on narrowly "economic" considerations when they make purchases, hire workers, and so on.
To the extent that people focus on the "bottom line," economic models capture behavior pretty well. This, however, shows only that they presuppose market hegemony rather than explain it. Economic theory provides no insight into a fundamental fact about markets. In what follows, I argue that economic theory can be extended to account for focus on the "bottom line." The key to understanding market hegemony is the contextual character of human evaluation. As circumstances vary, people attend to different subsets of their interests. It is this human capacity to "lock on" to some considerations and exclude others that standard economic theory fails to capture. I examine this capacity in the context of market behavior and suggest a method for expanding economic theory to account for it. 514 Philosophy of the Social Sciences 3. I take this as noncontroversial, at least at the level of individual behavior. The evidence that people aren't purely self-regarding is, of course, compatible with a great deal of self-regarding behavior. There is a dispute about whether humans can be modeled "without departing from the naïve assumption that players are seeking to maximize their average monetary gain" or "we must instead attribute their behavior to their holding 'other regarding preferences'" (Binmore 2006, 83). Even advocates of the former position "do not deny the existence of other-regarding preferences. On the contrary, [they] trace their origin . . . to Hamilton's rule, which expresses the extent to which animals should be expected to make sacrifices on behalf of their kin" (Binmore 2006, 83). Evolution selects for animals that actually help their kin (to the right extent) but not necessarily for animals that want to help only their kin (e.g., Ross 2006, 63-65). From an evolutionary perspective, we can say that people "mistakenly apply [moral] norms" or "mistakenly model one another as relatives," but these aren't necessarily mistakes from the perspective of individual-level psychology (Ross 2006, 54-55). Furthermore, selection pressure for deviating from Hamilton's rule in behavior is imperceptible to particular individuals (Ross 2006, 66).
Standard economic theory holds that people act on their interests, whatever those interests are. The idea that people focus on "bottom line" considerations in the market is, however, entrenched in economic analysis (see, e.g., Fehr and Gatcher 2000, 159;Sen 1999, 2-3). 4 Basic economic models, for example, treat profit maximization as the single objective of business firms. 5 A widely held view holds that businesses have the particular social role of providing goods and services. Within the confines of that role, business managers (appropriately) ignore their nonmarket values (see, e.g., Barry 2000;Primeaux and Steiber 1994). Milton Friedman (1982, 133;also see Friedman 1983) gives this position its classic formulation when he claims that there "there is one and only one social responsibility of business-to use its resources and engage in activities designed to increase its profits so long as it stays within the rules of the game." There is good evidence that firms embrace the "responsibility" they are assigned (see, e.g., Blinder 2000, 18-20). Increasing competition, for example, seems to decrease gender discrimination in a variety of industries, exactly what the market-hegemony claim predicts (Black and Brainerd 2004;Schwartz-Miller and Talley 2000). More importantly, there is little support for the view that nonpecuniary values explicitly influence business decisions in a significant way. Very few firms base their business strategies on social values. Limited attempts to integrate social concerns into market decisions are widely recognized as cynical gimmicks (see, e.g., Stecklow and White 2004;Benady 2004;Kitchin 2003;Gunther 2003;Driessen 2003;Cowe 2002).
The same pattern holds for consumers. Standard economic models assume that they are interested in their own consumption; the evidence from markets vindicates this assumption. It is clear that consumers take into account both price and (several dimensions of) product quality when they make decisions in the marketplace. Still, most people don't consider factors that are unrelated to their use of goods. Evidence from surveys suggests that Ellis / Market Hegemony and Economic Theory 515 4. Economics is concerned with the allocation of scarce resources. The story it tells about resource allocation, however, appeals to what is essentially a psychological account: people act to get what they want (given their beliefs). Economic models must specify wants to make predictions. They often characterize people as self-regarding, but this is an auxiliary assumption. 5. While most economists recognize that this is a convenient fiction, the problem is usually thought to involve principal-agent issues (see, e.g., Kreps 1990, 233, 577-614, 723-29, 743-69). Even sophisticated models usually assume that owners are interested solely in profit. although consumers "are positive . . . toward socially responsible companies," the majority "do not regularly use CSR [corporate social responsibility] as a purchasing criterion" (Mohr, Webb, and Harris 2001, 67; also see Cloninger et al. 2000, 25-26). Marketing promotions that allow consumers to designate money for charities are much less successful than those that offer "free" gifts (Benady 2003). All of this explains why there is no hard evidence that attention to social values helps firms make a profit (Gunther 2003). Even those who advocate corporate social responsibility admit that they are working against the dominant view (Stern 2002). The very existence of groups that urge producers and consumers alike to look at more than the "bottom line" in their market behavior suggests that neither normally do.
People don't forget their non-"bottom-line" interests as soon as they hit the office door or mall entrance: they cultivate real friendships (and animosities) on the job and look out for others (and indulge their prejudices) at the store. Market systems probably couldn't survive without other-regarding social norms-trustworthiness, responsibility, and so on-operating in the background (e.g., Sen 1999, 3-4). Still, despite their role in determining which transactions are available, nonpecuniary values have little influence on actual exchanges-when it comes time to hire a worker or write a check, self-interest usually comes to the fore.

II. Economic Accounts of Market Hegemony and Why They Fail
The fact that markets are hegemonic raises a problem for economic theory. 6 It is a foundational principle of standard economics that people act to satisfy their preferences, whatever they are. People form comparative preferences by reflecting on their values and interests. The vast majority of people-consumers, workers, managers, and owners alike-have 516 Philosophy of the Social Sciences 6. In this article, I focus on the tension between market hegemony and the account of human action in standard economics. There is also an important issue raised about the institutional framework of resource allocation. As James Wible (1995, 38-40) argues, modern economics takes markets to be the basic (even natural) institutions of resource allocation; other institutions (e.g., governments, firms, etc.) are secondary substitutes. Market hegemony suggests, however, that markets aren't up to the general task of resource allocation because they are geared to some interests and not others. To the extent that other institutions attend to interests neglected in the market, those institutions are better seen as compliments to markets rather than substitutes (see, e.g., Wible 1995, 52, 54-55). commitments to larger communities, principles, and causes. If, as economic theory holds, preferences drive market behavior then "bottom line" considerations can't be the only forces that motivate agents in the marketplace. Factors other than the use value of products ought to help shape preferences and so have economic influence (Freeman 1994, 80-84).
Economists usually try to explain market hegemony by appeal to the competitive pressures of free enterprise. Friedman makes the argument in an eloquent fashion: [A] free market separates economic efficiency from irrelevant characteristics. . . . The purchaser of bread does not know whether it was made from wheat grown by a white man or a Negro, by a Christian or a Jew. In consequence, the producer of wheat is in a position to use resources as effectively as he can, regardless of what the attitudes of the community may be toward the color, the religion, or other characteristics of the people he hires. Furthermore, . . . there is an economic incentive in a free market to separate economic efficiency from other characteristics of the individual. A businessman or an entrepreneur who expresses preferences in his business activities that are not related to productive efficiency is at a disadvantage compared to other individuals who do not. . . . In a free market they will tend to drive him out. (Friedman 1982, 109-10; also see Becker 1991, 299) 7 Friedman has in mind the pressures on price-taking firms that sell a homogeneous product. Such firms' profits are determined by their costs of production. A company that refuses to hire certain lower-wage employees or subsidizes community improvements forgoes opportunities to cut its costs. Higher-cost firms, however, fail as competition drives revenue, and so profits, down (also see Barry 2000, 106). 8 Ellis / Market Hegemony and Economic Theory 517 7. Karl Marx andFrederic Engels ([1848] 1978, 475) make much the same point when they claim that "the bourgeoisie, wherever it has got the upper hand, . . . has pitilessly torn asunder the motley feudal ties that bound man to his 'natural superiors,' and has left remaining no other nexus between man and man than callous 'cash payment.' . . . For exploitation, veiled by religious and political illusions, it has substituted naked, shameless, direct, brutal exploitation." 8. Gary Becker (1991, 299) thinks that this argument is "naïve" with respect to one nonpecuniary value: "Altruists receive psychic income in place of money income-they consume as they sell their products and services-and they can survive as well as money-income maximizers if they do not try to consume too much." He still expects market hegemony, however, "because altruism in market transactions is an inefficient way to produce psychic income. . . . Although efficient participants in market transactions may be highly altruistic, they act as if they are selfish and maximize their money incomes. They express their altruism through cash transfers not tied to market transactions" (Becker 1991, 300).
As an application of economic theory, the foregoing argument for market hegemony is implausible. If consumers would prefer to see certain values practiced by firms then they would pay more for products from firms that have those values. Likewise for the capital market, "if investors care about ethical issues, they will find the stock of an unethical firm to be less desirable and its stock price will be lower" (Chambers and Lacey 1996, 93-94). 9 Where people have some unselfish interests, market hegemony seems inconsistent with economic theory. 10 One might take the existence of market hegemony to be evidence that people don't hold nonpecuniary, other-directed values. There is, however, plenty of evidence for non-"bottom-line" values from nonmarket circumstances. In experiments where subjects are given the task of dividing sums of money, for example, "only a quarter . . . [were] selfish moneymaximizers" (Andreoni and Miller 2002, 737). Even within the marketplace, consumers respond to values that aren't self-regarding when those value considerations are made explicit. Environmental labeling, for instance, can influence consumer behavior (Stecklow and White 2004;Bjorner, Hansen, and Russell 2004;Nimon and Beghin 1999). It was established as early as the 1970s that providing information relevant to an ecological issue in the 518 Philosophy of the Social Sciences 9. Becker's efficiency argument doesn't affect these points. It doesn't even purport to cover ethical principles, social norms, and so on. Furthermore, it considers a very limited sense of altruism. Cash transfers are only more efficient than subsidizing purchases in the sense that they allow beneficiaries to get more of what they want. Concern for others, particularly in families, rarely takes the form of a desire to see them have more cash to spend on anything (Hausman and McPherson 1996, 41-45, 71-83;Becker 1991, 298). Most of our altruistic values envision others as doing well, not merely having more resources. 10. Modeling agents as self-regarding allows economists to predict market trends. Some economists purport not to care about anything else-economic models are instruments for making predictions about a limited range of variables, not psychological accounts (see, e.g. , Friedman 1990). This sort of instrumentalism is controversial for a variety of reasons. It is sufficient to note here that indifference to the individual foundations of economic activity is hard to sustain given what we know (and are learning) about behavior. In fact, most instrumentalists are willing to hypothesize about the connection between aggregate outcomes and individual behavior. One view holds that economic models work because individual-level deviations from these models are randomly distributed and so cancel out at the aggregate level (see, e.g., Kreps 1990, 5-6;Posner 1998Posner , 1556. This is a particularly implausible amount for nonpecuniary interests. Many people care about the well-being of their families; few wish them ill. Many people want to be honest; few are averse to honesty. Instrumentalists also appeal to market selection pressure: people who fail to act in accordance with standard models are likely to wash out of economic roles where those features cost them much (Manne 2006, A16;Friedman 1990, 4;Becker 1962). Whatever its value as an argument for rational behavior or the shape of aggregate supply and demand curves, this is a poor argument for market hegemony. news can influence consumer demand for products (Henion 1972). People really do have other-directed values. Those values, however, rarely seem to influence market behavior.
Given that people have (some) unselfish values, market hegemony suggests that the impact of market transactions on such interests is somehow unclear. It is an important part of Friedman's account, then, that consumers don't know about production processes. It is only this ignorance that allows producers to disregard the values of their customers (see also Freeman 1994, 82-83). Of course, firms that share their customers' values have no incentive to leave them in the dark about that fact. Firms can profit by publicizing their commitment to popular values. Product differentiation of this sort is a perfectly standard and, more to the point, obvious economic strategy. Even where businesses would prefer to avoid disclosing activities related to certain value issues, it doesn't follow that such information would be unavailable. Given consumer interest, there is a monetary incentive for third parties to supply the relevant information (and in a consumer-friendly fashion). 11 The competitive-pressure argument for market hegemony makes sense only under a rare set of conditions that market pressures work to eliminate. 12 Economic theory suggests, rather, that market forces will serve as a mechanism for transmitting value considerations through society Ellis / Market Hegemony and Economic Theory 519 11. Researchers have developed bar code readers that link products with databases of various sorts-pollution complaints, ethics violations, product recalls, and so on (Wade 2003, G5). 12. Some hold that focus on the "bottom line" is legally required for public corporations because managers have a fiduciary obligation to stockholders. James Hazelton and Ken Cussen (2005), for example, claim that this legal obligation is "arguably even more powerful than competitive markets" in explaining the lack of corporate social responsibility. This view overstates the stringency of legal obligations. In common law, managers are given wide discretion. Corporate charitable giving, for example, has been found to be consistent with shareholder sovereignty on the grounds that it can be part of a long-term strategy to increase profitability (Springer 1999, 87-88, 93-94). Twenty-nine states have also supplemented common law with "constituency statutes" that permit corporate managers to consider nonshareholder interests in their decisions (Springer 1999, 85, 96-102). The experience with constituency statutes reinforces the point that people just don't consider all of their values in economic contexts. The law could accommodate people bringing their broader values to the marketplace. The primacy of shareholder wealth hasn't been overthrown, however (Springer 1999, 121-23). In fact, constituency statues "can be nominally deployed in the name of protecting non-shareholder constituent interests while simultaneously and effectively deployed in the pursuit of maximizing shareholder wealth" (Murray 2000, 659). (Chambers and Lacey 1996;Cloninger et al. 2000). The facts about market hegemony show, however, that it doesn't work out this way. 13 Market hegemony poses a serious difficulty for any account that is recognizably economic. Economists model the role a person's values play in causing her behavior with a utility function that numerically represents her preferences over possible outcomes. With rare exceptions, economic models treat utility functions as univocal and fixed. In the few cases where models are dynamic, utility functions are understood as evolving slowly and continuously (see, e.g., Kreps 1990, 133-46). The evidence suggests, however, that while unselfish interests are important in some circumstances, they don't impact preferences, or at best have little influence, in a certain range of situations. The weights of different value considerations are fixed in utility functions of the standard sort. This is inconsistent with the evidence that values have very different degrees of influence in different cases. A very basic fact about how markets work is, therefore, beyond the explanatory reach of the tools economists normally use.

III. Extending Economic Theory to Account for Market Hegemony
Standard economic theory arguably has a number of problems. Its failure to account for market hegemony, however, suggests difficulties that are rarely recognized. The problem isn't, for example, that economic theory 520 Philosophy of the Social Sciences 13. Economists sometimes try to account for (what is arguably) a species of market hegemony by appealing to subtle features of circumstances. "Crowding out" of intrinsic motivation occurs when the introduction of financial incentives for behavior decreases the amount of the behavior. The crowding out effect has been observed in cases as diverse as blood donation, day care pickup, and charitable contributions (see, e.g., Janssen and Mendys-Kamphorst 2004;Frey and Jegen 2001;Gneezy and Rustichini 2000). Economists usually try to explain such cases by claiming that financial incentives change the total payoff to behavior in ambiguous ways. Uri Gneezy and Aldo Rustichini (2000) hold that a set schedule of financial incentives change can resolve contract uncertainties in a way that reduces overall incentives to abide by the contract. Following a standard psychological line, some economists claim that financial incentives undermine either autonomy or self-esteem (Frey and Jegen 2001, 594;Deci, Koestner, and Ryan 1999, 628-29). Maarten Janssen and Ewa Mendys-Kamphorst (2004) hold that financial incentives can replace social rewards (e.g., prestige) in a way that reduces overall incentives. Whatever value the foregoing accounts have, they are insufficient to explain the general phenomenon of market hegemony. In fact, those accounts presuppose market hegemony. The uncertainty-resolution account assumes that only "bottom line" considerations matter in decision making. Likewise, monetary incentives seem controlling, and so to undermine autonomy, because the market is seen as a value free zone. Self-esteem and prestige attach to certain behavior only outside the market because it is assumed that people in the market act only in their own self-interest. focuses exclusively on self-interest-it doesn't. 14 The difficulty, rather, is that it fails to accommodate the changing influence of unselfish values in different circumstances. Concern for others, ethical values, social norms, and the like play important behavioral roles in many areas of human life. In market transactions, however, the influence of such motives drastically diminishes. Economic theory can't account for this disappearing act.
The varying importance of nonpecuniary interests in different situations shows that market hegemony isn't just another empirical anomaly of the sort that motivates behavioral economics. Psychological evidence establishes that individual choices don't always fit well with standard consumer theory (see, e.g., Tetlock and Mellers 2002). Behavioral economics responds to this evidence by accepting standard consumer theory as a normative benchmark, but proposing descriptive models meant to capture the way behavior deviates from it. 15 These models retain the basic structure of consumer theory while replacing simple accounts of subjective probabilities and utilities with more psychologically plausible (but still mathematically tractable) alternatives (Camerer 1999). The problem of market hegemony, however, isn't that behavior doesn't reflect beliefs and values in precisely the way that consumer theory predicts. The issue, rather, is that behavior reflects different sets of values (and beliefs) at different times. From the perspective of standard economic theory, there is nothing amiss with market behavior, given the assumption that only selfregarding values are in play. What is strange is that only self-regarding values seem to be in play. The "curve fitting" strategy of behavioral economics isn't the right approach to market hegemony because it doesn't help with the nowyou-see-them-now-you-don't quality of various motivations. 16 Ellis / Market Hegemony and Economic Theory 521 14. That people have many sources of motivation has been acknowledged by virtually all economists, although it still needs to be internalized by some (McCloskey 1997). If anything, market hegemony suggests a cautious approach to economic modeling: why build in unselfish motivation if it rarely seems to matter? 15. Among its more prominent claims are that people evaluate changes in situations rather than states of affairs, that decision weights in uncertain situations are distinct from probability assignments, and that people discount future prospects exponentially (Tetlock and Mellers 2002;Camerer 1999). 16. People do exhibit certain psychological regularities at variance with economic models. A fixed set of judgment and evaluation biases can't, however, explain the variable influence of nonselfish interests. Stereotyping about the traits and capacities of certain groups (a cognitive shortcut) can, for example, result in racial or sexual discrimination (Reskin 2003, 9). The ultimate effect of such stereotypes still depends on the (variable) motives of the biased actors: discrimination, altruism (albeit of a paternalistic sort), or even attempted exploitation are all compatible with a given stereotype (Reskin, McBrier, and Kmec 1999, 339). Furthermore, it is important to note that the influence of stereotypes and other cognitive biases is notoriously variable itself (Reskin 2003, 9-10).
The existence of market hegemony shows that particular interests can play different roles in a person's behavior at different times. To account for this we must recognize that people can act on proper subsets of their values. 17 A person is disposed to act in a situation that poses a problem (challenge, opportunity, etc.) for her. She responds to the problem by trying to solve it (see, e.g., von Wright 1983, 50-51; Martin 1991, 129-30;Schick 1997, 17-18;Schick 1991, 55-60). 18 Aristotle (1978, 701a32-4) provides a simple example: "I want to drink, says appetite; this is drink, says sense or imagination or thought: straightaway I drink." In this case, some situation prompts the agent's thirst-heat, exercise, and so on. A person can be thirsty, of course, without paying attention to that fact. Here, however, she does pay attention. The agent's view of her situation activates her desire for a drink. Her perception of her circumstance brings thirst to the fore, making it seem like a problem. The agent operates in the context of her problem. The claim here isn't merely that thirst prompts the agent to think generally about how she isn't doing as well as she could be. Rather, her mind focuses on thirst in particular as her problem. Aristotle's agent might well have a number of other reasons for acting (e.g., she wants to give a sacrifice, finish a tragedy, etc.), but in this situation she acts on her thirst. 19 Following Frederic Schick (1997,(14)(15)22), we can say that a set of values and beliefs is a ground for the action it recommends. Understood this 522 Philosophy of the Social Sciences 17. The observation that values go "on-line" and "off-line" might also be accommodated for by "multiple self" accounts of behavior. There are two main versions of this approach. One tries to capture behavioral differences over time by positing that people are made up of distinct subagents, each of which takes a turn in the agency "driver's seat" (e.g., Schelling 1984;Margolis 1982;Etzioni 1986). The other treats different values as competing "advocates" who control the behavior of an agent through "political" procedures (Kavka 1991). The first approach would model people as more single-minded than they actually are. With regard to the second, the outcomes of political processes won't have the same features as the preferences of univocal agents. "Political control" views are, in fact, subject to an intrapersonal version of Arrow's Theorem (Kavka 1991;Ellis 2006, 322-24). The market behavior of most people, however, is perfectly well behaved apart from the fact that it doesn't reflect all of their values (Hayden and Ellis 2007, 653-54). 18. What matters is not that the agent actually has a problem but that she thinks she does. 19. This idea is similar to the idea of framing, as it is discussed in prospect theory. On both views, there is a step in the decision-making process where agents code or edit the information they have in order to use it (see, e.g., Kahneman and Tversky 1979, 274). Framing in prospect theory, however, doesn't involve censoring or highlighting particular interests. Instead, people are thought to conceive of their options as (overall) gains or losses relative to some reference point (see, e.g., Kahneman and Tversky 1979, 274, 275, 277-80). The claim that situations prompt a kind of (more or less severe) "tunnel vision" distinguishes the view considered in this article from framing as it is traditionally understood. way, grounds are potential reasons for actions. Each of us has grounds for a number of incompatible actions. A particular ground is activated, and so leads to action, in virtue of the person's understanding of her situation (Schick 1997, 21-27). In Aristotle's case, the agent's ground for drinking (and not her other grounds) is activated by her understanding that thirst is her problem now, not her religious duties or her play deadline.
To have a certain understanding is to represent a situation with certain propositions involving certain properties. 20 Attitudes toward those propositions are thereby activated and so employed in her reasoning. Other propositions involving those properties are likely to spring to mind as well, activating desires and beliefs that involve them. Imagine, for example, a high school student who wants to buy a soda after school. Cans of soda are $0.10 cheaper at the nearby Local-Mart than they are anywhere else. Our thirsty student, however, doesn't like the way the Local-Mart management treats her peers: they limit the number of teens allowed in the store at one time, watch them like a hawk, and so on. If she understands going to the Local-Mart in terms of getting a soda at a discount, she is likely to buy her drink there. If she understands going to the Local-Mart as patronizing an unfriendly establishment, however, she is likely to go elsewhere. 21 The latter understanding is not an attitude toward the proposition I am patronizing an unfriendly establishment. The idea, rather, is that our student sees the circumstances surrounding her proposed action through the lens of group identity and fairness: those considerations loom large in her thinking. Her beliefs about being a teen, the troubles teens face, the adversarial relationship between teens and adults, her desire to be treated respectfully and fairly, and so on come to mind when she thinks about her situation, and so they will guide her actions. The understandings account is essentially an associationist story. 22 New mental states become activated because they are either environmentally cued (e.g., our student sees a "No Loitering" sign at the Local-Mart) or related in certain ways to desires, beliefs, and so on that are currently activated (e.g., our student has been reflecting on how unfair adult rules are.) Ellis / Market Hegemony and Economic Theory 523 20. A person has one understanding at a time-the one implicit in the values and beliefs she is actually using. This understanding cannot always be simply characterized, of course. It might involve more than one dominant theme; for example, someone might understand a situation in terms of both loyalty and justice. 21. If she understands going to the Local-Mart as both betting a soda at a discount and patronizing an unfriendly establishment, she will feel torn unless and until she can compare the value of cheap soda to the disvalue of supporting teen-unfriendly business (Ellis 2006, 313-16 A person's understanding of her particular circumstances plays an important role in determining how she comes to act. On this account, reasons for actions aren't necessarily overall evaluations of the sort that utility functions are supposed to provide. There are often values (interests, desires, etc.) that aren't considered. 23 The application of this account to market hegemony is straightforward. In market transactions, people focus on the products, services, and so on that are on offer and the money they must give up to acquire them. Other interests are likely to be "off-line." Understandings are able to play the role they do in the foregoing account because people have a number of values and interests. Recognizing the existence of multiple objectives raises a host of important but generally neglected issues (Ellis 2006). For the purposes of this article, however, we don't need to examine them in all in detail. Nothing in economic theory denies the existence of multiple human objectives. On the standard account, the motivating force of all of a person's interests and values is integrated into one preference ranking. Market hegemony provides evidence, however, that people sometimes act from a proper subset of their interests. Nothing in the understandings account denies that activated values can still be integrated. Someone who understands a potential purchase in terms of both use value and social concerns might well balance both considerations when she decides what to do. Some philosophers, however, deny that (certain) interests can ever be (reasonably) balanced or integrated (e.g., Berlin 1969;Williams 1981;Kavka 1991;Kekes 1993;Wiggins 1997;Anderson 1997;Raz 1997;Taylor 1997;Lukes 1997;Sunstein 1997). On this view, many values are incommensurable: there is no way to bring their evaluations together to form all-of-those-things-considered preferences of the sort of envisioned in standard economic analysis. Someone who understands a situation in terms of two (or more) incommensurable interests can't form a joint assessment, so this view rules out (most) more-than-one-thingconsidered evaluations.
My full view about value incommensurability is spelled out elsewhere (Ellis 2008). For the purpose of this article, it is enough to note that there is no good reason to think that anyone's values are really incommensurable. 524 Philosophy of the Social Sciences 23. How is such a partial evaluation sufficient to support an action? An activated ground shows that a course of action is desirable insofar as it advances toward the objective the agent actually considers. Given the agent's understanding, she doesn't see anything to give her pause before acting. If a person has grounds for an action and she understands her circumstances in a way that activates exactly the mental states that make up those grounds, then (in the absence of external restraint or infirmity) she does the action.
There are, of course, many cases where agents don't know what to do because their interests pull them in different directions. Difficulty, however, doesn't imply impossibility: it is important to distinguish between uncommensurated values and incommensurable values. There is real incommensurability only if there is something unreasonable about balancing some considerations against others. While values are often in conflict, in the sense that they can't be jointly attained, it never is the case that values themselves resist integration (Ellis 2008, 35-38). Recognizing the pull of one interest is compatible with recognizing the pull of others, so people can be in a position to look at their values side-by-side. People certainly try to balance their different objectives (e.g., they worry about work and family trade-offs). Appeal to experience plays an important role in these efforts (e.g., How well did I fare when I emphasized this ends?). As a matter of fact, people make a variety of many-things-considered evaluations that suggest that joint evaluation can succeed. 24 In many cases preferences seem to have enough structure to make economic sorts of models reasonable. Adding understandings to standard models can help make further sense of cases where people have reasonably complete preferences. 25 It is fairly easy to adapt economic theory to accommodate Schick's insights about understandings. Standard consumer theory starts with overall preferences. It doesn't say much, however, about how preferences result from an agent's various interests. It is reasonable to start by assuming that people can evaluate outcomes with respect to their basic interests. Formally, we can represent this by supposing that each of a person's basic goals induces an interval ranking of possible actions. Let A = {a i |a i is a proposition that the agent thinks she can make true} be the set of available actions; D = {d|d is a basic desire} be the set of desires for basic ends; R d be the complete and transitive binary preference relation provided by desire d; and u d (⋅) be an expected utility Ellis / Market Hegemony and Economic Theory 525 24. It is clear, of course, that people don't fully commensurate their values. People learn about how their values fit together (when they do) from hard-won experience, but life isn't always a clear and comprehensive teacher (Ellis 2006, 329-31). Commensuration is also a matter of degree: someone might decide that avoiding unfriendly establishments is worth paying tens of dollars but not worth thousands of dollars; she might be undecided about hundreds of dollars. Any degree of commensuration is an achievement and people can differ in their levels of accomplishment. Uncommensurated values will lead to incomplete all-thingsconsidered preferences. Economists, of course, have usually admitted that complete preferences are an idealization. 25. If you find my views about incommensurability unpersuasive, replace the following model with either a "serial control" or "political control" model of the sort discussed in note 17. Each spells out a way of combining an understandings view with value incommensurability (Ellis 2006, 317-22). function that numerical represents R d . 26 Each person has a set of basic ends and so a set of evaluations with respect to those ends. In different situations, different subsets of those evaluations will be activated. Again, we can represent all of this formally. Let C = {R d | d ∈ D} be the set of preference relations provided by basic desires, N = {N | N is a possible situation} be the set of possible situations, Pow(C) = {C i |C i ⊆ C} be the power set of C and S: N → Pow(C) be a function that specifies the set of preference relations activated by a person's understanding of her situation. For convenience, let S(N) = C N . The function S captures the way in which environmental cues and associationist principles bring some values into the foreground of thought.
As I noted above, it seems at least possible for someone to bring together a set of evaluations with respect to basic ends to come up with a single, joint evaluation with respect to all of these ends together. To see what this sort of commensuration amounts to, suppose an agent in situation N faces two options, a i and a j , and has two activated preference rankings, R m and R n (i.e., S(N) = {R m , R n }). Suppose further that the agent's experiences convince her that the advantage of a i over a j with respect to R m is exactly balanced by the advantage of a j over a i with respect to R n . With respect to desires m and n together, therefore, she is indifferent between a i and a j . Knowing this allows us to calibrate our agent's preference ranking with respect to m and n. Formally, u m (a i )-u m (a j ) counts the same in her joint evaluation as u n (a j )-u n (a i ). In general, of course, u m (a i )-u m (a j ) ≠ u n (a j )-u n (a i ), so to reflect the fact that these two intervals count the same we can weight them: w m (u m (a i )-u m (a j )) = w n (u n (a j )-u n (a i )). We can rearrange the equality above to get w m u m (a i ) + w n u n (a i ) = w m u m (a j ) + w n u n (a j ). It is natural to interpret w m u m (a i ) + w n u n (a i ) as the value of a i with respect to desires m and n together since it is the appropriately weighted sum of the values of a i with respect to each desire alone. Let u mn (·) = w m u m (·) + w n u n (·). u mn (·) numerically represent a person's preferences with respect to m and n together (R mn ). 27 526 Philosophy of the Social Sciences 26. Interpret "a i R d a j " as "a i is at least as good as a j with respect to desire d." R d is complete (for any a i , a j ∈ A, a i R d a j , or a j R d a i ) and transitive (for any a i , a j , a k ∈ A, if a i R d a j and a j R d a k then a i R d a k ). I assume that interval preferences over actions are derived from ordinal preferences over outcomes and subjective probabilities in the standard way (see, e.g., Resnick 1987, 13, 22-25, 88-98). u d : A → ℜ (the real numbers) is an expected utility function that numerical represents R d . u d (a i ) ≥ u d (a j ) iff a i R d a j . u d (⋅) is unique up to positive linear transformation, that is, where b > 0, bu d (⋅) + c represents R d as well as u d (⋅). 27. Note that the weights w k are tied to the particular numerical representations being employed. The constraint on weighting schemes is that the weights must make equally attractive intervals equal, no matter the scale of the intervals. It is important, for example, that w n /w m = (u m (a i )-u m (a j ))/(u n (a j )-u n (a i )), but of course that ratio holds if we set w m = 1 and solve for w n = (u m (a i )-u m (a j ))/(u n (a j )-u n (a i )).
Standard economic theory, of course, supposes that all of a person's values are not only commensurable but also commensurated. While this is clearly an idealization, it can be a reasonable one, for example, when an agent has a clear joint ranking of options that are on the table. Our soda-drinking high school student, for example, might decide that taking everything together, she prefers to not buy her sodas from the Local-Mart.
In general, u(·) = Σ i∈C w i u i (·) represents a person's most considered evaluation, her preferences when she attends to all of her interests. The claim that a person's overall utility function can be a weighted sum of the numerical representations of preference rankings with respect to specific ends is formally analogous to a result proven by Harsanyi (1955, 312-14). We must make allowances, however, for particular understandings in particular situations. If a person attends to only some attributes, then her overall ranking of options, and so her behavior, will represent that focus. In a particular situation (N) a person will have a particular set of activated rankings (S(N) = C N ). Her ranking of available actions in that circumstance will be her joint evaluation with respect to the ends in that set. Formally, her preferences will be represented by u CN (·) = Σ i∈CN w i u i (·), and she will choose an action a i ∈ A such that u CN (a i ) > u CN (a j ) for all a j ∈ A.
With the modified account of consumer utility in place, the rest of economic theory proceeds in the usual fashion. Market demand for products is built up by aggregating consumer demands. Consumer demands, however, will depend on the understandings of consumers. 28 What sorts of understandings are they likely to have? The evidence suggests that, when acting as consumers, people tend to focus on their own "bottom line." Is this merely an accident? If someone is looking to consume goods or services, self-regarding values are almost certain to be activated. No other value(s) will be cued in such a regular way. Recognizing that self-interest is the most reliable motive in market circumstances, people who think about the market will tend to emphasize it. This emphasis, however, has influence on how people behave. Reflection on the role of self-regarding interests in markets strengthens the association between the marketplace and those values.
The suggestion that understandings can be self-reinforcing connects in an interesting way with Cristina Bicchieri's important work on social norms (see, e.g., Bicchieri 2000Bicchieri , 2006. On her account, a social norm is a certain sort of behavioral rule. It is important to see that such rules are "situationally contingent" (Bicchieri 2000, 153). As Bicchieri (2000, 160-70;2006, Ellis / Market Hegemony and Economic Theory 527 28. Owner utilities matter to supply in much the same way, but for the sake of simplicity I'll confine myself to the consumer side of the market. 56-76) notes, a given norm will only apply in some situations; different norms will apply in different situations. Bicchieri (2006, 2) accounts for social norms "in terms of the expectations and preferences of those who follow them." Very roughly, a social norm exists when a sufficient number of people are willing to apply a behavioral rule to certain situations because they share certain desires and beliefs (Bicchieri 2006, 11-20). 29 Bicchieri's account of norms fits nicely with Schick's story about understandings. The existence of social norms can, for example, help explain which understandings people are likely to form. The associationist account allows that people can have very idiosyncratic understandings. A social norm, however, provides a set of "pre-packaged" beliefs and values that people can employ to categorize, schematize, and so understand the situations they face (Bicchieri 2000, 171-73;2006, 81-99). On the other hand, any desire-belief account of social norms must appeal to something such as understandings to explain how norms go on-line and off-line as situations change. Bicchieri holds, for example, that "particular preferences and beliefs many not always be activated" (2006, 57) but also that "whenever the environment 'focuses' us on a particular norm, it makes salient certain behaviors appropriate to the situation, which we in turn follow" (2000,154). The story about understandings grounds notions such as activation, focus, and salience that are required to make sense of social norms.
The associationist account of understandings allows that if someone spends a good amount of time thinking in a certain way, that pattern of thought can become ingrained in her mind. A set of interests (a social norm or personal commitment) that is routinely activated by common circumstances will be easily accessible in many situations. As Bicchieri notes, people can bring familiar cognitive schemes to make sense of one-off situations. 30 In this way focus on the "bottom line" can become a kind of ideology (Blinder 2000). Market ideology emphasizes self-interested, consumptionoriented values not because it prioritizes those values over others but rather because those interests are regularly activated in a wide range of circumstances. A very interesting line of research suggests, for example, that people trained in economics tend to act on a particular set of market-oriented norms. In experimental situations, economics students act from self-interest to a 528 Philosophy of the Social Sciences 29. This, of course, provides very little insight into Bicchieri's rich analysis. It is enough for my purposes here, however, to note that there is a connection between norms and shared expectations and preferences. 30. Of course, other environmental cues may trigger other values or norms. Associationist principles help explain why things tend to look like nails to someone used to hammering. greater extent than other students (Marwell and Ames 1981;Carter and Irons 1991;Frank, Gilovich, and Regan 1993). Other experiments suggest, however, that economics students respect property rights to a greater degree than do other students (Yezer, Goldfarb, and Poppen 1996). As Laslo Zsolani (2003, 707) notes, "Economics students seem to represent a special pattern of moral behavior that is characterized by respect for property rights and strong self-interest motivation simultaneously." None of the foregoing suggests that anyone acts solely on self-interested, market norms. In certain circumstances, even hard-nosed economists will attend to their nonpecuniary values, altering the situation-specific utility functions that drive their behavior. All of this is exactly the result that the market hegemony view leads us to expect: prolonged exposure to the principles of the market leads people to attend to, and so ultimately act on, their own "bottom line" in any situation that looks remotely like a market. Still, even economists act on different interests at school or in the park than they do in the market. This latter point suggests that there is nothing inevitable about market hegemony either. While self-interested considerations are always likely to be activated in market situations, other interests can be activated as well, but only where those interests can be highlighted. This is more likely where adherence to those interests is a central part of people's lives. 31