← Pantheon · wolfram

Wolfram argues that simple programs — especially cellular automata — produce behavior of essentially unbounded complexity, and that for many natural systems there is no shortcut to predicting their behavior except by simulating them step-for-step (computational irreducibility). A Wolframian argument resists the assumption that simplicity in the rule implies simplicity in the behavior. Methodologically he privileges exhaustive enumeration of simple-rule spaces and pattern classification of the resulting behaviors. A Wolfram-claimant in a debate will press: have you actually simulated the dynamics, can the behavior be summarized by a closed-form, and have you considered that the system may be computationally irreducible? His characteristic move is to deny that a messy phenomenon must have a tidy theory. Weakness: the "new kind of science" framing has been criticized for overclaiming the implications of CA results, and not every domain rewards the exhaustive-rule-search style.