Fallacies: E (Part 2)

Core idea

The second half of E targets misplaced authority and misread evidence: the elevator button that does nothing, the embassy that is not foreign soil, the expert who is confidently wrong, the statistician who extrapolates a river into absurdity. The thread is that a source's prestige — a control panel, a diplomat, an expert, a chart — is not the same as its reliability.

Why it matters

We outsource judgement constantly, and we outsource it to whatever looks authoritative. But authority is uneven. Experts are reliable about what they have observed and unreliable about what they predict. A button can exist purely to make you feel in control. A trend line can be technically real and still produce nonsense when stretched. Knowing where authority ends keeps you from following it off a cliff.

Key takeaways

Mental model

Copy sourceZoom

Practical application

Test whether the control actually controls

The placebo elevator button gives the feeling of agency without the fact of it. Before relying on any control — a setting, a process, a procedure — confirm it does what its label promises. Comfort is not function.

Trust experts on observation, doubt them on prophecy

Eminent scientists laughed at gas lighting, fast trains, and atomic power. They were excellent at describing what they had measured and poor at ruling out what they had not imagined. When an expert says "this is how it is," weigh it heavily; when they say "this can never be," weigh it lightly.

Refuse to extend a trend past its evidence

Twain's joke — that the shrinking Mississippi was once a million miles long — works because a real short-term rate becomes nonsense when projected. A line on a chart describes the range it was measured over and nothing beyond it.

Example

A team adopts a project tool whose dashboard shows a satisfying "team health" score. Morale feels managed. Months later they discover the score is computed from a single rarely-updated field — it has been frozen for weeks. They trusted the dashboard because dashboards look authoritative, never checking what fed it. Like the elevator button, it offered the sensation of control with none of the substance. The fix is dull but reliable: ask what every authoritative-looking signal is actually measuring.

Related lessons