Returning to the conclusion of Part II, reciprocal co-operators do best in the game of two-player Prisoner’s Dilemma (PD). 1 They solve the compliance problem because they are both rational and arguably moral. But this game is not the only or even the most difficult test for morally constrained free agents. While the PD is a deep and difficult problem, it turns out to be especially favourable to a moral solution by means of communication, commitment and an opportunistic minimal morality. Therefore I should consider other situations that may challenge my conclusions. 2 Aggression provides a good test case for my minimally moral agents. It is a commonplace that bullies test one’s moral mettle. There are three reasons that aggression provides a particular challenge to my justification of morality. First, aggressors make use of my favoured devices, communication and commitment, for the purpose of amoral advantage. They threaten using the devices moral agents use for promising. Second, opportunistic agents—the indirect maximizers of Chapter 7 —have good reasons not to resist such threats; it may simply be better to be red, or robbed, than dead. Third, morality seems to demand that we resist threats in favour of impartially fair outcomes.
I use the game of Chicken to model the problem of aggression. This chapter develops a theory of substantive rational strategies for Chicken. I contrast Chicken and the Prisoner’s Dilemma with special attention to the devices of threats and promises. Then I shall show that the presence of committed threateners complicates the situation similar to the way committed unconditional co-operators complicated the PD. Once again we are pulled in different directions: now, to acquiesce or to resist. Again I argue that rational morality takes the low road: the substantially rational strategy is to acquiesce to unreasoning threateners. This is a disturbing moral conclusion.