In certain interactions each person, acting on individually rational self-interest, chooses a defecting move that leaves both worse off than if they had cooperated. The trap is that cooperation yields a better joint outcome, yet unilateral cooperation is punished, so rational actors defect even knowing the shared loss.
Everyday conflicts of interest reduce to a small set of recurring game structures (such as the Prisoner's Dilemma, Chicken, and the Tragedy of the Commons) each defined by how the players rank the four possible outcomes. Identifying which dilemma you are actually in tells you why cooperation keeps failing and which escape route applies.
A division is made fair without a trusted arbiter by having one party split the resource and the other party pick first; the cutter's self-interest forces an even cut because they get whatever is left. The mechanism aligns individual incentives so that fairness emerges from procedure rather than goodwill.
In repeated encounters a strategy that cooperates first, then simply copies the other player's previous move, outperforms more complex or purely selfish strategies. It works by being nice, retaliating against defection, forgiving quickly, and being predictable enough for the other side to learn cooperation pays.
When a dilemma traps players in mutual defeat, the better move is often to alter the payoffs, add future rounds, or introduce communication rather than play the existing game harder. Shifting the structure, for example by making the interaction repeated or letting players make binding commitments, can convert a trap into an opportunity for cooperation.
Cooperation in dilemmas becomes sustainable when players can trust each other, because trust removes the need for constant verification and defensive defection. Trust is built incrementally through small reliable exchanges and signals, and once established it lets groups reach outcomes that pure rational calculation would block.