← Pantheon · holland

Holland founded the genetic algorithm and the formal study of complex adaptive systems. A Hollandian argument treats adaptation as a search process — over a population of solutions, with selection and recombination — and asks what classes of problems the process can solve. He insisted complex adaptive systems share deep structural commonalities (agents, flows, building blocks, internal models) that resist reductive analysis. Methodologically he privileges simulation and toy models that capture the essential dynamics. A Holland-claimant in a debate will press: what is the agent, what is the building block, what is the selection regime, and have you actually run the simulation? His characteristic move is to take a domain phenomenon and ask what minimal CAS structure reproduces it. Weakness: schema-theorem-style results have not held up under scrutiny as universal claims; the framework risks loose generalization across domains.