Intelligence: Its Structure, Growth, and Action

Cattell's triadic theory rejects the notion that a single general factor, g, sufficiently explains the positive correlations among cognitive tests. Instead, it proposes three levels of ability: constitutional capacities (including fluid intelligence and speed), provincial abilities tied to specific sensory-motor channels, and acquired agencies such as crystallized intelligence. The theory's distinctive mechanism is investment: fluid intelligence is expended through learning to produce crystallized skills, which accounts for both the correlations among abilities and their divergent developmental trajectories. This framework was Cattell's most systematic alternative to Spearman's unitary g and to purely environmental accounts of ability differences.

This 1986 volume is explicitly a revised edition of Cattell's 1971 Abilities: Their Structure, Growth, and Action. However, Arthur Jensen's 1989 review in Contemporary Psychology was sharply critical of the claim that the book was brought up to date. Jensen noted that the revision was essentially the same as the 1971 text, with a new typesetting but minimal substantive updating of the research literature. He also identified specific defects, including inferior reproduction of shaded figures that in one case produced an actual error regarding frontal lobe function. Jensen concluded that as Cattell's fullest statement of his position it remained a landmark, but as a current textbook it was already outdated.

The fluid-crystallized distinction has proven to be one of the most durable constructs in the field, though not in precisely the form Cattell proposed. John Horn, Cattell's student, extended the model by identifying additional broad abilities such as visualisation and short-term memory, and later Horn explicitly rejected a higher-order g factor. Carroll's massive 1993 reanalysis of 461 datasets found a three-stratum structure with g at the top, which is compatible with but not identical to Cattell's framework. The Carroll-Horn-Cattell (CHC) model now dominates psychometric intelligence research, but its relationship to Cattell's original triadic theory is mediated by these later theoretical disputes, particularly regarding whether g is a meaningful higher-order construct or an artefact of correlated second-stratum factors.

Investment is the causal bridge in Cattell's model between fluid and crystallized intelligence. Fluid intelligence is conceived as a general, constitutionally based capacity for perceiving complex relations; crystallized intelligence consists of acquired skills and knowledge domains. The theory holds that individuals invest their fluid capacity into particular learning enterprises, and the returns on that investment depend on personality factors such as interest, concentration, and persistence. This explains why fluid and crystallized abilities are positively correlated without being identical, and why educational and motivational variables moderate the translation of fluid potential into crystallized achievement. The investment notion was later incorporated into the CHC model and remains one of Cattell's most cited contributions.

Horn was Cattell's collaborator and the most important empirical extender of the Gf-Gc model. Their joint papers in the 1960s provided the developmental evidence that fluid and crystallized abilities follow different lifespan trajectories: crystallized intelligence tends to rise or hold steady into middle age, while fluid intelligence declines earlier. However, Horn later diverged from Cattell by rejecting the existence of a higher-order g factor altogether, favouring a model of correlated broad abilities. This book, written late in Cattell's career, incorporates Horn's findings but maintains Cattell's own theoretical commitments. The tension between Cattell's acceptance of a general factor and Horn's rejection of it remains a live issue in the CHC literature.

Cattell tentatively proposed a localization scheme in which fluid intelligence and provincial sensory-motor abilities depend on the sensorimotor and general association zones of the brain, while crystallized agencies depend on the effectiveness of storage and encoding-decoding regions. These claims were speculative for their time and were presented as hypotheses rather than established findings. Later neuroimaging research has substantially revised this picture, identifying distributed networks involving prefrontal and parietal regions for fluid reasoning and different systems for semantic knowledge. Cattell's neurological chapter is therefore best read as a historical document of mid-twentieth-century psychometric theorizing about brain function, not as a current account of the neuroscience of intelligence.