Chapter 1: Evolution of reciprocator preferences when agents can pay for information. A benchmark result in the evolutionary games literature is that a preference for reciprocity will evolve if preferences are observable (at zero cost), since reciprocators can cooperate with each other rather than with materialists, thereby achieving a fitness advantage. I investigate how a preference for reciprocity evolves if individuals can observe an opponent's preferences only by bearing a fitness cost. My main result applies when observing an opponent's type is cheap, but cooperating only gives a modest fitness advantage or the preference for reciprocity is intense. In this case, a preference for reciprocity cannot evolve from a small starting share in the mix of preferences, even if discovering an opponent's preferences is arbitrarily cheap. This is in sharp contrast to the benchmark result. Chapter 2: A theory of conditional cooperation on networks (with Julien Gagnon) Chapter 2 is a study of reciprocity on social networks. We model a group of connected agents who play a one-shot public good game. Some players are materialists and others are reciprocators. We characterise the maximal Nash equilibrium (ME) of this game for any network and a broad class of reciprocal preferences. At the ME, a novel concept, the q-linked set, fully determines the set of players who contribute. We show that influential players are those connected to players who are sufficiently interconnected, but not too much. Finally, we study the decision of a planner faced with an uncertain type profile who designs the network to maximise expected contributions. The ex ante optimal network comprises isolated cliques of degree k*, with k* decreasing with the incidence of materialists. We discuss an important application of our results: the workplace. Chapter 3: Ideological games Chapter 3 is a theory of ideology. I define a preference type to be a set of first-order preferences over the outcomes of a `game of life', together with a set of (`meta-') preferences over all players' first-order preferences. Players can influence each other's preferences via costly investment: if player A invests and B does not, B's preferences becomes those of A. Players may invest for instrumental reasons (i.e. to achieve better outcomes in the game of life) or `ideological' reasons (i.e. they want their opponents to have the same preferences they do). I characterise `strongly ideological', `weakly ideological' and `pragmatic' types. Weakly ideological types wish to preserve their own type, as do strongly ideological types, who also seek to convert others. A pragmatic player, in contrast, is willing to have her type changed if her new type would prefer the resulting equilibrium of the game of life to the status quo. I show that if two players of different ideological types meet, there is an equilibrium investment profile with lower aggregate welfare than the no-invest profile. If at least one type is strongly ideological, there is a unique such equilibrium. Finally, a `perfectly ideological' type is a strongly ideological type which, if held by all players, results in the best outcome of the game of life as judged by that type. If a perfectly ideological player plays a pragmatic player, aggregate welfare is always greater than in the no-invest profile.

No / This important and highly original book explores the application of economics to the subject of hate via such diverse topics as war, terrorism, road rage, witchcraft mania, marriage and divorce, and bullying and harassment.