Cargo Cult

Cargo cult describes copying the outward form of a successful practice while the mechanism that made the original work is missing or never understood. The term comes from Melanesian movements that, after wartime airfields brought cargo, built runways, control towers, and headsets out of straw and wood and waited for planes that had no reason to land — an imitation nowhere near exact, but close enough in outline to show that the function was never in the form.

That source image is contested. Anthropologists largely treat "cargo cult" as a pejorative that lumps unrelated movements together — Lamont Lindstrom on the phrase's career as a Western fantasy, Martha Kaplan arguing to drop it altogether — and read the practices as attempts to create and renew social relationships under colonial pressure rather than as naive confusion about how aviation works. The term has largely fallen out of use in the discipline. The metaphor still names something real in engineering; it does so at the expense of the people it borrows from.

Richard Feynman borrowed the phrase for cargo cult science: research that follows every visible ritual of the scientific method — the paper, the citation, the statistics — while skipping the part that actually enforces honesty, a genuine attempt to prove yourself wrong. The same shape recurs in engineering process: a team adopts standups, story points, and a retro board because an admired team had them, without the discipline — acting on the retro's findings — that made them do anything. Rituals persist because they are the part that is easy to see and copy; the mechanism is invisible from outside, which is Chesterton's Fence run in reverse — the risk here is keeping a form nobody understands rather than discarding one.

Calling something a cargo cult is only useful if it is followed by naming the missing mechanism — what the ritual was supposed to cause, and why this instance of it does not cause that. Without that step the label is just Plausible Mechanism's failure mode aimed at other people: a confident-sounding explanation with the actual explaining left out. A Code Review that checks for approval rather than for correctness, or a Test-Driven Development cycle where the test is written to already pass, are the same failure with the paperwork intact.