A child who is good at maths is often a good reader too and learns poems quickly — but not always. Is there one “general intelligence” behind all these successes, or several independent abilities? This question kept psychologists busy for almost a hundred years. Their main tool in the search for an answer was factor analysis: a statistical method that shows which tests “move together” and how many shared sources lie behind them.
Spearman and the general factor g
In 1904 the English psychologist Charles Spearman noticed that the results of all mental tests are positively related: someone who does well on one test usually scores above average on the others too. He explained this with two kinds of factor: a general factor, g, present in every test, and specific factors, s, unique to each test. Tests differ in how heavily they draw on g.
Spearman likened g to a “mental energy” that, like an electric current, drives different machines. In his view, the heart of g is grasping relationships, as in the item “bird — nest, bee — …”. But Spearman's belief that g is inborn and cannot be changed by teaching is now outdated: schooling and the environment have a clear effect on results.
Thorndike and Thurstone: many abilities
Edward Thorndike rejected the general factor. In his view, learning is the building of connections between stimuli and responses, and intelligence depends on the number and complexity of these connections. He distinguished three kinds of intelligence: abstract (working with ideas and symbols), mechanical (working with objects and tools) and social (understanding people and getting along with them).
In 1938 Louis Thurstone gave dozens of tests to university students, studied the results with factor analysis and identified about seven primary mental abilities: verbal comprehension, word fluency, number, space, memory, perceptual speed and reasoning. He suggested looking at a profile of abilities instead of a single number: a child can be strong in one area and average in another.
But in 1941, when Thurstone studied children, it turned out that these abilities are also related to each other. If you analyse the links between them, a general factor appears again — this time as a “second-order” factor. So the debate reached a reconciliation: there is a general factor, and there are separate abilities.
Hierarchies: from Vernon to the CHC model
Later researchers combined the two ideas in a hierarchy. Philip Vernon put g at the top; below it, two large groups — verbal-educational and practical-mechanical; then narrower abilities; and specific factors at the bottom.
In the 1960s Raymond Cattell, working with John Horn, distinguished two sides of intelligence. Fluid intelligence is the ability to solve new problems for which there is no ready-made knowledge: spotting patterns, reasoning in an unfamiliar situation. Crystallised intelligence is the knowledge and skills a person gathers through learning and experience: vocabulary, general knowledge, familiar and tested ways of doing things. Crystallised intelligence can keep growing for as long as a person keeps learning.
In 1993 John Carroll re-analysed more than 460 data sets and proposed a three-stratum model: general intelligence at the top; broad abilities in the middle (for example, fluid and crystallised intelligence, memory, visual-spatial processing and processing speed); and dozens of narrow abilities at the bottom. Combined with the work of Cattell and Horn, this became the Cattell–Horn–Carroll (CHC) model, which is today the mainstream view in the field.
Where do Gardner's intelligences fit on this map? Several of them are close to the broad abilities of CHC: linguistic intelligence to crystallised intelligence and verbal abilities, logical-mathematical to fluid intelligence and mathematical knowledge, spatial to visual-spatial processing. In 2006 Visser and colleagues showed that tests of most of Gardner's more thinking-based intelligences are linked to each other and to g. Gardner's theory is useful as a window onto different strengths and as a reason to vary teaching, but as a scientific theory it remains debated.
What this means for parents
A century of debate offers parents a few practical conclusions: - A child has a general level of learning ability: it shows how quickly they pick up new things across different areas. This level is not destiny — good schooling and a rich environment develop it. - At the same time, every child has an uneven profile of strengths: stronger in one area, weaker in another. This is normal, not a flaw. - Both matter. A single IQ score is an average of many abilities and can hide both strengths and weaknesses; a profile, in turn, is incomplete without the general level. - Use strengths to support weaker areas: a child with strong spatial imagination often understands a maths problem more easily through a drawing. - Discuss test results with a consultant and look at the profile, not just the overall number.