Analogical thinking maps structure from one domain onto another. Surface similarity ("this clogged drain is like that clogged drain") is easy. Deep analogical thinking — Dedre Gentner's focus — finds relational structure across domains that look unrelated on the surface. That is what wicked-world problem solving requires.

Why it matters

Kepler had no database for planetary forces. He chained analogies — light, odor, magnets, boatmen in a whirlpool — until he invented astrophysics. Duncker's radiation problem stumps ~90 percent of solvers until fortress and fire-chief stories prime "split weak forces that converge." One distant analogy tripled success; two disparate stories pushed it to 80 percent.

Human intuition poorly self-triggers analogies for ill-defined problems. We default to the inside view and familiar surface cues — Cold War framing made students choose war or diplomacy based on which historical analogy was primed.

Practice

  • Collect multiple analogies from different base domains before deciding (private-equity investors underestimated returns until they compared to a reference class of similar projects).
  • Distant analogies beat near ones for idea generation (Nike + McDonald's beat Apple + Dell for business strategy prompts).
  • Northwestern's Integrated Science Program trains cross-domain sorting; students who took breadth courses grouped phenomena by causal structure, not just by field.

Kevin Dunbar's lab meetings: breakthrough labs threw analogies every four minutes, increasingly distant as problems got weirder. Homogeneous labs stuck in one trench.

Sources