As we have seen throughout the book, perpetrators may inflict violence on one person to constitute relationships with others. For example, a soldier may kill an enemy in obedience to an officer, or an initiate may kill a member of a rival gang. A perpetrator may kill one member of another group in retaliation for some other member of that group’s killing a member of the perpetrator’s group. A man may kill his wife’s lover to regulate his relationship with his wife, or, in an honor culture, kill his niece’s lover, and his niece, in order to regulate his relationships with his family and with everyone in the community. As we saw, the Trojan War was all about men fighting men to constitute relationships with other men, or with the gods. In all these and many other cases we have considered in this book, the motive to constitute one or more of the component relationships of a metarelational model morally requires violence in another of the relationships that compose the metarelational model. In general, relationships have moral implications for other relationships with which they are metarelationally linked.

But these moral links often work the other way around. As we mentioned briefly in Chapter 2, metarelational models may inhibit violence. If a soldier is ordered to kill a family member or his village chief, he may refuse. The gang initiate may avoid killing his sister’s boyfriend. In a feud, potential perpetrators may refrain from violence if the opposing group includes in-laws, co-members of an age set or secret society, blood brothers, or compadres. Conflict-restraining relationships such as these are “cross-cutting ties” that limit violence in many societies, including ones where there are no effective police, judiciaries, or chiefs (Colson, 1953; Cooney, 1998: 90–6; Evans-Pritchard, 1939; Gluckman, 1954, 1963; LeVine, 1961; Nader, 1990; Rae and Taylor, 1970; Ross, 1993). Such metarelational cross-cutting ties operate in all societies, including modern ones. For example, Varshney (2003) found that violence between Hindus and Muslims was less likely to occur following an instigating act of violence elsewhere in the country in cities where Hindus and Muslims were already working together on joint civic projects.