The Evolution of Cooperation

Darwin’s world is supposed to be a bloodbath: every organism clawing for its own survival, every gene shoving its rivals aside, no room for charity. The cold logic of the one-shot prisoner’s dilemma agrees — when helping you costs me, the rational move, and the evolutionarily stable one, is to defect. And yet look around. Vampire bats regurgitate blood to feed hungry roost-mates. Cleaner fish and their “clients” run a trusting little service economy on the reef. Strangers return lost wallets, nations sign treaties and keep them, your flatmates (usually) do the dishes. The world is drenched in cooperation among agents who owe each other nothing. So which is it — is selection a war of all against all, or a nursery for altruism? That contradiction is the puzzle this course exists to resolve, and its resolution is one of the most beautiful results in all of science: cooperation is not the opposite of self-interest — under the right conditions it is what self-interest evolves into.

This is an advanced course spanning biology & evolution and strategy, and it stands on two prerequisites. From natural selection it borrows the engine — variation, selection, heredity — that decides which behaviours spread. From Nash equilibrium it borrows the game — the prisoner’s dilemma, best responses, and the crucial trick that changing the game changes the equilibrium. If those aren’t solid, take them first; everything here composes them.

From that base we build the whole model floor to ceiling. We’ll state the paradox sharply — why blind selection seems to forbid costly helping, and why a naive do-gooder gets eaten alive. Then we’ll turn the key that unlocks everything: the shadow of the future. Repeat the dilemma with the same partner and a defection today costs you all of tomorrow’s cooperation, so when the future looms large enough, cooperating becomes the self-interested move — no saints required. We’ll relive Axelrod’s computer tournaments, where the simplest program in the room, four-line tit-for-tat, quietly beat every fiendish schemer, and extract the four traits that made it win: nice, retaliatory, forgiving, and clear. We’ll then get honest about noise — one mistaken move can lock two tit-for-tat players into an endless revenge spiral — and meet the refinements that survive a messy world: generous and contrite tit-for-tat, and win-stay, lose-shift. We’ll go beyond direct reciprocity to the other engines that grow cooperation: indirect reciprocity and reputation (I help you, and someone who heard about it helps me — the machinery of gossip and status), kin selection and Hamilton’s rule rB > C (why you’d sacrifice most for your closest relatives), network reciprocity (why cooperators survive by clustering together), and an honest word on the group-selection debate. Finally we’ll collect the conditions that make cooperation win into a practitioner’s checklist, and name the ways the model lies to you — cooperation’s fragility, the one-shot-versus-repeated confusion, tit-for-tat’s vulnerability to noise, and the ever-tempting naturalistic fallacy (what evolves is not the same as what is good).

Throughout, you’ll run a live iterated dilemma and watch a whole population of strategies evolve — turning the length of the future up and down and seeing cooperation sweep the population or collapse into universal defection. By the end you’ll be able to look at any standoff — between firms, countries, colleagues, or cells — and say precisely which levers would turn a war of all against all into a stable, self-enforcing truce.