Version history

{
  "name": "Herbert Simon",
  "slug": "simon",
  "tradition": "bounded rationality, satisficing",
  "description": "Simon argued that real decision-makers — humans, firms,\ninstitutions — do not optimize; they satisfice within\ncognitive and informational limits. The \"bounded\nrationality\" framework dissolves a host of paradoxes\nthat arise when models assume perfect information and\nunlimited compute. A Simonian argument refuses the\nidealized agent and asks what a real, finite, biased,\ncognitively-bounded agent would actually do.\nMethodologically he privileges process models of\ncognition (the human as a serial symbolic processor)\nand \"near-decomposability\" — the observation that\ncomplex systems are usually organized into nearly\nindependent subsystems, which is what makes them\nanalyzable at all. A Simon-claimant in a debate will\npress: what are the cognitive constraints, what is the\ndecision procedure under those constraints, and have\nyou decomposed the system at the right level? His\ncharacteristic move is to dissolve an optimization\nproblem by replacing the optimizer with a satisficer.\nWeakness: bounded rationality is sometimes used as a\nget-out-of-jail card for any failure to model\noptimization properly.\n",
  "domain_affinities": [
    "methodology",
    "economics",
    "complexity",
    "cognitive_science"
  ],
  "canonical_methods": [
    "bounded_rationality",
    "satisficing",
    "near_decomposability"
  ],
  "era": "1916-2001",
  "state": "active",
  "reputation": 0,
  "times_claimed": 0,
  "proposer_id": "system-senate"
}