1923: René Thom is born
On 2 September 1923, René Thom was born in Montbéliard, France. A brilliant geometer and topologist, he received the Fields Medal in 1958. In the 1960s and 1970s he developed catastrophe theory, a mathematical language for describing how smooth, gradual changes in control parameters can trigger abrupt, qualitative shifts in the state of a system.
Thom’s landmark book Structural Stability and Morphogenesis (1972) classified the elementary catastrophes and showed how they could model phenomena ranging from the breaking of a wave to the differentiation of tissues in an embryo. Although the theory originated in pure mathematics, its conceptual power quickly attracted biologists, linguists, psychologists and economists who needed to reason about sudden transitions.
Within the broader history that led to artificial intelligence, catastrophe theory supplied one of the first rigorous frameworks for discontinuous change in cognitive processes. Researchers used it to model the sudden reinterpretation of ambiguous figures, the “aha” moment in problem solving, and bifurcations in learning dynamics. In an era when most cybernetic and early AI models emphasised continuous feedback or discrete symbol manipulation, Thom’s geometry offered a third way: qualitative dynamics grounded in topology.
The influence travelled through cognitive science into AI. Ideas of structural stability, unfoldings and generic singularities appear, suitably transformed, in later studies of neural-network loss surfaces, in analyses of phase transitions during training, and in attempts to understand emergent capabilities of large models—points at which performance jumps discontinuously as scale or data change.
Today, when AI researchers speak of “emergent abilities,” “mode collapse,” or “sharp left turns” in capability, they are often circling questions that Thom formalised half a century earlier. His birth on this day therefore marks not only a milestone in mathematics but a subtle tributary in the long river of ideas that now sustains artificial intelligence.