Cover of Emergence

Emergence

Steven Johnson

6 ideas

  1. Intelligence Without a Pacemaker

    Complex coordinated behavior can arise from local interactions among simple agents following simple rules, with no central controller directing the whole. Order emerges from the bottom up rather than being imposed from the top down, which means searching for a 'leader' in an emergent system fundamentally misreads how it works.

  2. Slime Mold Aggregation Mystery

    Slime mold cells were long assumed to be organized by specialized 'pacemaker' cells that commanded the others to assemble. The actual mechanism is that any cell can emit a chemical signal (cyclic AMP) and follow gradients of it, so aggregation self-organizes from purely local rules with no commander at all.

  3. Stigmergy via Pheromone Trails

    Ants coordinate by modifying their shared environment rather than communicating directly: each ant deposits and follows pheromone trails, and shorter paths accumulate stronger signals because they are traversed faster. The colony optimizes routes as a collective without any individual ant understanding or intending the result.

  4. Four Conditions for Emergence

    Emergent systems require enough agents (a critical mass of low-level units), local interaction and information exchange among neighbors, attention to signaling and feedback, and the capacity to encode patterns into adaptive behavior over time. Without sufficient density and feedback loops, the system stays dumb; cross these thresholds and macro-intelligence appears.

  5. Neighborhoods Solve Without Knowing

    Cities cluster like trades and economic functions into districts through the accumulated micro-decisions of residents who each act on local information, never intending the citywide pattern they produce. The neighborhood functions as a kind of distributed memory and problem-solving structure that 'knows' things no single inhabitant knows.

  6. Feedback as the Engine of Adaptation

    Treat any complex behavior as the product of feedback loops rather than fixed instructions: negative feedback stabilizes a system around equilibrium while positive feedback amplifies small differences into large patterns. Looking for where information loops back on itself reveals how a system learns and self-corrects without external guidance.

Save and mark ideas in the app