It is often the case that intelligence tests do not measure what they purport to measure, although it is commonly assumed that every test with a different name is a test of different skills, and that tests with the same name test the same skills. Spearman attempted to study the relationship between tests that supposedly measured different mental abilities, and upon finding substantial correlation between different tests, posited a general factor of intelligence ("g"), to explain these relationships. The present study questions this approach. If "g" turned out to be indistinguishable from language proficiency, for example, questions would arise as to the implications for intelligence testing and the use of IQ scores, and for school tests that claim to measure constructs other than language proficiency but rely heavily on language-based performances. These questions are considered from three angles: (1) the pattern of interrelationships among tests that claim to measure different aspects of intelligence and language tests that require discourse processing is examined; (2) intelligence tests themselves are examined to analytically define what they require of learners; and (3) evidence from brain physiology and the processing of information is considered. (CLK)