Emergence and complexity

Context

A rebellion from inside physics. In 1972 Philip Anderson — a condensed-matter physicist tired of hearing that particle physics was the only "fundamental" science — published "More Is Different" in Science. His argument: the reductionist hypothesis (everything obeys the same microscopic laws) does not imply the constructionist one ("the ability to reduce everything to simple fundamental laws does not imply the ability to start from those laws and reconstruct the universe"). At each level of organization — solid-state physics, chemistry, biology, psychology — new laws appear that are not deducible in practice from below, and doing science at level N is not "just applied" level N−1. The essay became the founding document of complexity science; Anderson's own broken-symmetry work is exactly the physics that made the point technical.

The idea acquires an address. Conway's Game of Life reached the world through Martin Gardner's October 1970 column — a toy universe with four rules that turned out to contain gliders, self-replication, and (later proved) universal computation, the cleanest demonstration ever built that trivial parts can generate unbounded complexity. Prigogine's dissipative structures and Mandelbrot's fractals gave the mathematics teeth. In 1984 George Cowan and Murray Gell-Mann founded the Santa Fe Institute, and "complexity" became a discipline with a campus. The current frontier: causal emergence (Hoel), assembly theory (Walker & Cronin, 2023), scaling laws (West), networks (Barabási).

The hygiene that makes this unit survivable. Three different things get called "emergence": weak emergence — properties derivable from the parts, but only by running the computation (uncontroversial, ubiquitous, and the deep connection to computational irreducibility); strong emergence — properties not derivable even in principle (a live philosophical claim, not settled science, and arguably never once demonstrated); and mysticism — "everything is emergent, man" (noise with a vocabulary). Almost every bad argument in this territory is a quiet slide from the first meaning to the second. Keep the three labeled and the unit becomes a pleasure instead of a swamp.

How to read it

Anchor. Melanie Mitchell, Complexity: A Guided Tour. The best general introduction; Mitchell is at Santa Fe, writes clearly, and is honest about what's proven versus speculative. ~15 hours.

Companion (essential). James Gleick, Chaos: Making a New Science — the narrative history: Lorenz, Mandelbrot, Feigenbaum. ~12 hours.

Companion (short, essential). Anderson, "More Is Different" (1972). Four pages, free online. Read it three times over the course of the unit.

Companions (optional). Kauffman, At Home in the Universe (partisan for self-organization — read critically); West, Scale (genuinely surprising empirical content); Hoel's causal-emergence papers (current, contested — only after the rest).