The cosmopolitan, Martha Nussbaum tells us, is “the person whose primary allegiance is to the community of human beings in the entire world.” The nationalist, by contrast, holds a primary allegiance to a narrower social sphere, based on the idiosyncrasies of birth, residence, and upbringing. The opposition between these two allegiances is stark, indeed excessively stark. Later she affirms the subtler Stoic formulation of the cosmopolitan ideal: To be a citizen of the world is not to “think of ourselves as devoid of local affiliations, but as surrounded by a series of concentric circles. The first one is drawn around the self; the next takes in one’s immediate family; then, in order, one’s neighbors or local group, one’s fellow citydwellers. . . . Our task as citizens of the world will be to draw the circles somehow toward the center. . . .” I shall use this as a definition of cosmopolitanism in the remainder of my remarks.

The weakness in the cosmopolitan view is its implication that, in the ideal, all our allegiances to individuals would be of equal importance, independent of the accidents of particular life histories. As Nussbaum puts it, “One should always behave so as to treat with equal respect the dignity of reason and moral choice in every human being.”

By contrast, I believe we are all better off if we take our social commitments as dependent on the accidents of our particular life histories. Accepting the “concentric circle” imagery, the alternative is to hold that the various annuli representing our social connections have their ideal positions and sizes, allegiance to which tends to promote the good of all.

I respond to the needs of some human beings more than to the needs of others. I strive to help satisfy the aspirations and needs of my family more than those of other human beings. I help a friend in a situation where I do not help a stranger. I strive to improve the health, safety, and cultural vitality of my community more than those of other communities. I devote more time and energy to making mine a better country than I do other countries of which I am not a citizen. I do this with reason, since I think it is morally correct to do so, and with emotion, since my family, friends, ethnic heritage, and country have made me what I am. I will call this view “communalism,” although I do not know to what extent my use of the term conforms to its general usage.

Communalism is cosmopolitan in two senses. First, it makes a place in its values for all human beings — the outer annulus of the planetary system of commitments. I do not know what precise weight to place on this annulus, but I believe that it bids us to dedicate part of our resources to empowering others to meet their communal obligations to families, friends, ethnic groups, and countries of citizenship.

Second, communalism holds that if all people acted in this manner, we would all be better off, and conversely, if we do not act in this way, we will all be worse off. Why? My ability to affect the well-being of others — either through altruistic acts on my part or through arrangements of reciprocal aid — diminishes with their social distance from me. I am in a better position than a random individual to act on behalf of my parents, wife, and children than I am on behalf of abstract individuals distant from me, whose personalities are opaque to me, and whom I do not love with the love that comes from personal association and blood relations. Thus it would take a considerably greater dedication of effort and resources for an abstract individual of arbitrary distance from my family to achieve the same effect that I can through my personal efforts and by the use of my personal resources. Moreover, when I fail in my duties to my family, they are in a better position to punish my actions and prevent their recurrence than are abstract individuals when I fail in my duties towards them. Similar remarks apply to my circle of friends, my local community, and my country of citizenship.

In short, if we all committed ourselves to Stoic cosmopolitanism, the result would be inefficient — a great deal of wasted effort, a lot of defection from our committed principles, and an inferior outcome for large numbers of persons.

But what about inequality? Suppose I am very poor and a person whom I do not know is very wealthy. This person may then in fact be better equipped to help my family than I am myself. Similarly, if my community or nation is very poor and another community or nation is very rich, the latter may be better equipped to help mine than mine is to help itself. If communalism is more efficient, isn’t cosmopolitanism more equitable?

I do not know the answer to the question if we assume that poverty and wealth are immutable characteristics of groups (I shall use the term “group” to refer variously to families, communities, and nations, and terms “wealth” and “poor” to refer to general measures of group welfare). But they are not immutable. Among groups that were relatively poor in the past, some remain poor and some have become relatively wealthy, and conversely. Communalism can be defended as an ethic for the promotion of the well-being of even the nonwealthy. The argument is straightforward, and depends upon analyzing the gains from defecting from the cosmopolitan as opposed to the communalist ethic.

Suppose there are a number of groups of different wealth, and suppose the wealth of a group can be affected by the actions of its members. In a cosmopolitan world, the differences in wealth among groups are minimized by the charitable acts of the wealthy. Moreover, all groups have an incentive to increase their own wealth, whether or not they are donors or recipients of aid, since all care as much about the well-being of others as they do about their own well-being.

But suppose groups can choose to defect from the cosmopolitan ethic by not sharing if they are potential donors, and not improving their wealth if they are recipients. Under plausible conditions a wealth-producing group has an incentive to defect, since it will gain from not sharing, and a recipient group has an incentive to defect by not improving its wealth-producing capacity, for by improving it would not only lose subsidies from wealthier groups, but would also become a donor to other groups. Thus cosmopolitanism is not a socially stable ethic.

The same defect does not apply to a communalist strategy. Clearly groups cannot “defect,” since they are the architects of their own wealth. What about individuals in groups? The narrower the group within which an agent operates, the more the agent’s personal well-being depends on that of the group, and the greater are the pressures that the group can impose on the agent to fulfill group obligations. Therefore, groups can be designed to forestall defection of their members. In short, communalist groups have an incentive to improve their wealth, and ceteris paribus, in a communalist world, all are better off.

But ceteris are never paribus, which is why a concentric circle view of our social obligations seems better than a purely myopic view. Small groups often do better when they cooperate with others to form larger groups (economies of scale), and the vagaries of nature bid groups to form alliances for mutual insurance and sustenance. And the location of small groups within larger groups allows individuals to defect from their communal obligations by exiting to groups where they might expect to be better off. Consider a pertinent example: “open borders” policies allowing unrestricted international migration equalize world income.

The general progress in the improvement of living standards in the world had been dramatic in the past thirty years. Average real GNP per capita has grown by 1.73% per year, average daily caloric intake has increased by 0.55% per year, and the mortality rate for children under the age of five years has declined at a rate of 2.51% per year.1

But “uneven development” has been equally dramatic. Averaging various measures, the percent of world income inequality that is between-nations as opposed to within-nations is 75%: that is, if all inequality within countries were eliminated, world inequality would remain at about 75% its current level.2 Moreover, inequality has increased dramatically since the mid-19th century, and recent trends show no signs of improvement; if anything, they indicate a worsening in the past three decades.

Clearly, inequality rather than the general pace of development has been a problem in the world economy. Wouldn’t the cosmopolitan want, then, to promote massive migration from poor to wealthy nations?

Here are some facts: (a) from 1989 to 1992, the number of people living outside their country of origin doubled, from 50 million to 100 million; (b) most are escaping poverty, and almost half of the migrants are women seeking better economic opportunities; (c) about 20 million have fled violence, drought and environmental destruction.3 Although very little of this movement is defection of individuals from their families, much is defection of families from their communities and nations (although migrants often remit large amounts of money to their communities of origin, and often return under auspicious conditions). Doubtless neither communalists nor cosmopolitans will fault migrant families for choosing “exit” over “voice” in improving their lot. But the ideal cosmopolitan would approve unconditionally of the recipient country’s acceptance of migrants, while the communalist will not.

The communalist argues that a successful strategy of “open borders” is likely to decrease world living standards and increase the degree of equality of living standards in the medium and long run. To be sure, successful, open borders would significantly equalize the wealth of nations in the short run, as families move from lower to higher wealth-producing regions. As long as all regions maintain a cosmopolitan commitment to increasing their wealth-producing capacity, there is no problem. But then lower wealth-producing countries have a strong incentive to defect to a selfish or communalist strategy of exporting people and accepting remittances, rather than improving the wealth of all. Similarly, a higher wealth-producing country has a strong incentive to defect to a selfish or communalist strategy of closing its borders and following an internal development strategy rather than improving the wealth of all. In sum, the cosmopolitan ethic is not self-enforcing.

A communalist alternative may do better. It would require that poorer countries identify the forces underlying their failure to share in the general rate of economic improvement, and correct these failures, and that wealthier countries determine their complicity in world inequality and modify their behavior accordingly. It would also be just and desirable if the wealthy countries transferred material resources to the less wealthy with viable development policies. But their doing so is probably not a precondition for successful development policies in the less wealthy countries.