Intelligence

Core idea

"Intelligence" is one of psychology's most contested constructs. Some researchers think it is a single thing — a general ability that shows up in everything you do. Others think it is many things — verbal, spatial, kinaesthetic, interpersonal — each with its own neural address. The fight matters because however you define intelligence is what your tests will measure, and what your tests measure is what schools and employers will use to sort people.

The history of the field is a long pendulum swing. Spearman pulled it toward a single g factor. Thurstone pushed back with seven primary abilities. Gardner pushed further still with eight (later nine) multiple intelligences. Cattell carved the g factor itself into two halves — fluid reasoning, which peaks in young adulthood and declines, and crystallized knowledge, which keeps growing through middle age. Each model is a different bet on how the mind is carved at the joints.

Why it matters

IQ scores predict school performance, job performance, and health outcomes better than almost any other single number. But the predictions are correlations, not destinies, and the score does not measure creativity, common sense, social skill, or wisdom. If you mistake the score for the underlying construct, you have made a category error with high stakes — the kind that The Bell Curve made.

Mental model

Single g vs many intelligences

The cleanest way to see the field's two big factions is to draw the structure they each claim the mind has.

Copy sourceZoom

Fluid vs crystallized — the age curve

The fluid/crystallized split (Cattell) is the most useful distinction for everyday life because the two curves go in opposite directions across the lifespan.

Copy sourceZoom

The bell curve and what it means

Copy sourceZoom

Practical application

Example

Imagine you are hiring a senior software engineer. You have two strong candidates. Candidate A scored 145 on an IQ test in their teens and breezed through the abstract-reasoning portion of your screen. Candidate B scored 120 (still high) but has 15 years of building production systems in the exact domain you operate in.

Spearman's model would tell you A is the safer hire — higher g, all else equal. Cattell's split tells a more interesting story. A has more fluid reasoning available right now; B has far more crystallized domain expertise. For a novel research problem with no precedent, A's headroom matters. For the messy reality of a brownfield codebase with five years of context-specific decisions, B's pattern library is the dominant factor.

This is also why a 145 from age 17 does not survive translation. Fluid intelligence has been declining for two decades since that test. If A is now 37, the relevant question is no longer their teenage peak — it is what they have crystallized in the interim. Asking that question filters out a lot of plausible but wrong hires.

Related lessons