At the beginning of the twentieth century, psychologists faced an intriguing pattern. Measures that seemed to test quite different mental abilities tended to correlate. A person who did relatively well on one kind of intellectual task was more likely than chance to do relatively well on another. Charles Spearman thought that pattern demanded an explanation. His answer became one of the founding ideas of psychometric intelligence research: performance on a cognitive task reflects something general, shared with other intellectual activities, as well as influences specific to that particular task. He called the general component g and the specific component s. That idea is often flattened into the slogan that Spearman believed intelligence was “one thing.” It is more interesting than that. Spearman’s theory was an attempt to explain why cognitive performances are simultaneously related and different—and the tension between those two facts still structures intelligence research today.
The puzzle Spearman inherited
At the beginning of the twentieth century, psychologists were accumulating measurements of sensory discrimination, school performance, and mental tests. Spearman asked whether apparently different intellectual performances were related systematically. His 1904 work used newly developing correlation methods and argued that intellectual activities shared a common function. Later formulations developed what became known as his two-factor theory. The basic idea can be written conceptually as:
performance on a test = general influence (g) + influences specific to that test (s)
This is a conceptual simplification, not a modern scoring equation.
General and specific at the same time
Imagine four tasks:
- identify a rule in an unfamiliar pattern;
- understand difficult vocabulary;
- mentally rotate an object;
- retain and manipulate information.
Spearman’s framework says that performance on each task partly reflects something general enough to contribute across all four. But each task also requires things the others do not. The vocabulary task depends on particular learned words. The spatial task depends on particular spatial operations. The memory task has its own task-specific demands. So the theory was not:
“Every cognitive task is exactly the same.”
It was closer to:
“Different cognitive performances contain a common component as well as specific components.”
Why did Spearman infer a general factor?
Because correlations among cognitive measures were systematically positive. If the same general source contributes to several tests, then people who differ in that source should tend to perform differently across all of them. In modern language, a reflective common-factor model represents this by placing a latent factor upstream of the observed test scores:
g → reasoning score
g → vocabulary score
g → spatial score
g → memory score
The arrows are a model of dependence. They should not be mistaken for direct observation of a biological substance.
What was s?
This is often explained too casually. In Spearman’s two-factor framework, s referred to specific influences associated with particular performances after the general component was considered. It does not follow that every modern broad cognitive ability is simply a Spearman s factor. Later factor-analytic work found systematic groupings among tests: verbal measures share variance beyond g; spatial measures do too; memory, speed, and other domains show additional structure. That evidence helped motivate hierarchical models with levels between a single general factor and individual tasks.
Was Spearman saying g was just a statistic?
No. His theorizing went beyond statistical description. Spearman did not merely observe a mathematical factor and refuse to interpret it. Across his work he developed psychological interpretations of general ability and at times invoked ideas resembling general mental energy. Those historical causal interpretations should be separated from the modern statistical fact that a general factor can be extracted from diverse cognitive measures. A researcher today can use a g factor without accepting Spearman’s original psychological mechanism.
What did later research change?
A great deal. Spearman’s simple two-factor framework did not adequately represent all of the systematic covariance among cognitive tests. Thurstone emphasized several primary mental abilities. Cattell and Horn distinguished broad abilities such as fluid and crystallized intelligence. Carroll’s reanalysis of a very large factor-analytic literature supported a hierarchy containing narrow abilities, broad abilities, and a general factor. The modern lesson is therefore not:
Spearman proved intelligence is one thing.
It is:
Spearman identified and modeled a general component of cognitive covariance that later models had to accommodate while also recognizing additional structure.
A worked comparison: Spearman versus a hierarchy
Suppose six tests are administered:
- two verbal;
- two spatial;
- two reasoning.
All six correlate positively. But the two verbal tests correlate especially strongly with each other, as do the two spatial tests and the two reasoning tests. A simple Spearman-style representation emphasizes:
all six tests → general component + test-specific component.
A hierarchical model can represent more:
individual tests → domain factors → general factor.
The second model can capture both what all tests share and what tests within a domain share beyond the general component. That is one reason modern intelligence structure is usually described as hierarchical rather than strictly two-factor.
Did Spearman discover the cause of intelligence?
No. He made foundational contributions to correlation, measurement, and the statistical study of general intelligence. But identifying a common dimension does not uniquely reveal its cognitive or biological cause. That remains the central boundary between Spearman’s durable psychometric legacy and stronger interpretations of g.
Spearman’s durable insight—and its limit
Spearman’s theory proposed that cognitive performance contains both a general component and task-specific influences. The positive correlations that motivated g remain fundamental. The simple g + s architecture was not the final map of cognitive abilities, and Spearman’s own psychological interpretation of g should not be confused with the statistical general factor used in contemporary models.
References
- Spearman, C. (1904). “General intelligence,” objectively determined and measured. American Journal of Psychology, 15, 201–293. DOI
- Spearman, C. (1927). The Abilities of Man. Macmillan. Source
- Carroll, J. B. (1993). Human Cognitive Abilities. Cambridge University Press. DOI
- Savi, A. O., et al. (2019). The wiring of intelligence. Perspectives on Psychological Science, 14, 1034–1061. DOI
