Skip to content

Evolutionary Constraints and Contingency

Why It Matters

Evolutionary outcomes aren't inevitable or optimal - they're path-dependent and constrained. This combats deterministic thinking: "if it exists, it must be optimal." Often things exist because of historical contingency, not because they're best. Recognizing constraints and contingency reveals why some solutions never appear and why optimization is bounded.

The Key Move

When encountering evolutionary outcomes (biological, technological, social), ask: what constraints shaped this? What accidents locked in this path? What alternatives were foreclosed? This prevents mistaking "what happened to occur" for "what had to occur" and reveals opportunity for redesign.

Classic Application

Vertebrate eye. The retina is "installed backwards" - photoreceptors behind neurons, creating a blind spot where nerves exit. Cephalopod eyes (evolved independently) don't have this flaw. The vertebrate design is a historical constraint from early developmental architecture, not an optimal solution. Contingency matters.

Surprising Application

QWERTY keyboard. Designed to prevent mechanical typewriter jams by separating common letter pairs - a constraint that no longer exists. But switching cost and installed base lock in the design despite superior alternatives (Dvorak). Historical contingency, not current optimality, explains persistence. Path dependence creates constraints.

Failure Modes

  • Constraint apologetics: not all features are constrained accidents. Some are actually well-designed solutions. Don't use constraint as excuse to avoid seeking function.

  • Ignoring that constraints change: developmental constraints can be overcome by genetic changes; physical constraints by new materials. Constraints are real but not eternal.

  • Assuming contingency means randomness: contingent doesn't mean uncaused or unpredictable. It means path-dependent with sensitive dependence on initial conditions.

Go Deeper

  • Gould, Wonderful Life - contingency in evolution (though somewhat overstated)
  • Maynard Smith et al., "Developmental Constraints and Evolution" - types and roles of constraints