Before we start arguing over how strong reciprocity is, perhaps we should agree on what it is, and what it has to do with equality and sharing. One of the poignant lessons of egalitarian theory is that unselfinterested terms often get defined in ways that reflect self-interest. In a sense, that's what ideology is all about.

I turn to my dictionary (American Heritage, third edition), which defines reciprocity without reference to equality. It emphasizes consensual exchange, which to my economist's mind evokes markets. As with markets, the outcome of reciprocal exchange may or may not be equal and fair. Whether it is or not depends on the relative endowments and the relative bargaining power of the two exchangers, and how these are acquired. The capitalist pays the worker a wage in exchange for labor power. Is this an equal and fair exchange? To my mind, this depends on how the capitalist acquires capital and what kinds of opportunities and alternatives are available to the worker.

Some simpler examples: The feudal lord provides military protection for the serf in return for a large share of what the serf produces. The slave owner provides for the basic subsistence of the slave in return for property rights over him or her. The husband, in the traditional marriage contract, agrees to honor his wife in return for her obedience to him. The Federal Reserve arranges a bailout for private speculators in return for promises to help stabilize the world financial system. The International Monetary Fund lends money to Korea in return for promises that investors from other countries will be allowed to buy up Korean businesses. These are all forms of "conditional generosity." But the conditions are, to say the least, suspicious.

Thus far, Bowles and Gintis would surely agree. So let me tackle one of their own examples. They argue that widespread opposition to Aid to Families with Dependent Children (AFDC) in the United States was based not on widespread selfishness or lack of concern for the poor, but on a more noble concern that the program violated principles of reciprocity and encouraged bad behavior. Given a choice between these two explanations, I would side with Bowles and Gintis. But there is a third, not entirely exclusive, but rather different explanation. Selfish elites who largely control our political and cultural institutions succeeded in redefining the concepts of reciprocity and bad behavior in ways that significantly disadvantaged poor mothers. Many egalitarians, including Bowles and Gintis, failed to speak out against this redefinition.

AFDC was initially set up on a principle of reciprocity. Single mothers would care for their children, helping produce the next generation of taxpayers and workers in return for extremely modest levels of support while their children were under 18. The notion that this was a fair exchange was partly undermined by the increase in paid labor force participation of mothers and by increasing acceptance of paid childcare as an alternative to mother care. But careful and compassionate efforts to encourage poor mothers to increase their participation in paid employment (which is after all in their own economic interests) were derailed by a draconian "reform" rhetoric that accused AFDC recipients of getting a free ride on the taxpayers' backs.

Bowles and Gintis never explicitly dispute this accusation. They point to welfare mothers' "bad behavior" of bearing children out of wedlock without mentioning fathers' "bad behavior" of resisting economic and personal responsibility for their offspring. Other dimensions of reciprocity are also at stake–if we want to discourage non-marital births why don't we provide birth control and abortion services for those who can't afford them? In urging egalitarians to pay more attention to the way ordinary people think about fairness, which is certainly a good idea, Bowles and Gintis legitimate the commonplace assertion that welfare mothers are not among the "deserving" poor.

The weaknesses of their political analysis of welfare reform exemplify an approach to reciprocity that ignores a large feminist literature documenting significant inequalities within families. Contrary to most sociobiologists' claims, familial altruism is limited. For a significant portion of the "hundred thousand years of sharing" that Bowles and Gintis describe, men have exercised patriarchal power over wives and daughters, restricting their women's economic and sexual opportunities. That men have also expressed a certain "basic needs generosity," seldom allowing their wives and daughters to starve and only occasionally beating them, is reassuring, but not quite adequate.

I almost titled this reply "Homo Patriarchus and the Future of Androcentric Politics." As an epigraph to counter Bowles and Gintis's Nordic gloss, I considered an old English folk saying: "A woman, a donkey, a walnut tree, the more you beat them, the better they be." I thought this kind of reply would be, well, reciprocal. I rejected it, not because it was too contentious, but because it implies that male domination necessarily defines our species. Like Bowles and Gintis, I want to challenge the all-too-common presumption that entirely selfish behavior is embedded in our genes. But I don't think that egalitarian reciprocity is embedded there either. If it were, relations between men and women would have evolved rather differently.