Order the psychometric theories by how much structure they claim intelligence has, from the simplest to the most elaborate.

1. One general factor accounts for performance across every cognitive test
2. Two broad kinds: reasoning with novel material, and accumulated knowledge
3. Seven primary abilities, related but worth distinguishing
4. A hierarchy: one general factor above broad abilities above narrow ones

Laid out this way the century-long argument looks less like a series of contradictions and more like a question of resolution. Spearman saw one thing; Cattell and Horn split it in two along a line that predicts different age trajectories; Thurstone distinguished seven.

The hierarchical model is last because it is not a rival to the others but an arrangement of them. Narrow abilities sit at the bottom, broad ones like fluid and crystallized in the middle, and g at the top - and each earlier theory turns out to have been describing one storey of the same building.

Which is also why the data never settled the original dispute by refuting anyone. Thurstone's seven abilities correlate, so a general factor can be extracted from them; Spearman's g does not preclude meaningful distinctions beneath it. Both men were reading a real structure at the level of detail their method encouraged.