This is actually a fairly obscure topic in the history of psychology. According to a recent paper (starting with its conclusion), in the early 1940s Cattell mostly reforumulated an idea of Hebb, which the latter obtained by observing the differential effect (on verbal vs non-verbal abilities) of some brain injuries (lobotomies of various kinds actually). After a 20-year hiatus, Cattell published on this again (together with his student Horn) but this time with focus on the fluid intelligence as being "culture-fair".
“Cattell's” theory of fluid and crystallized intelligence is Hebb's theory of Intelligence A and Intelligence B, given another name and popularized by Cattell. Cattell's theory was Hebb's idea. [...]
In his 1943 paper, Cattell (1943, p. 179) stated that:
Hebb (1941c, 1942) has independently stated very clearly what constitutes two thirds of the present theory, for he says that “intellectual power may be needed for the first appearance of the qualitatively superior response, but not necessarily for its persistence” (Hebb, 1942), and “in any test performance there are two factors involved, whose relative importance varies with test: one factor being the lasting changes of perceptual organization and behavior induced by the first factor during the period of growth.”
Hebb's presentation (Hebb, 1941c) discussed the differences in effects of brain injury on adults and children and the finding that “certain kinds of abilities are less affected by late than by early injury.” While I do not have a copy of Cattell's (1941) APA presentation, I do have the typed version of Hebb's (1941c) in which he presented his ideas on Intelligence A and Intelligence B as follows:
It may be proposed that intellectual development includes two distinct things: (A) direct intellectual power, by neural maturation, and (B) the development of qualitative modifications of perception and behavior. The first factor is what reaches a peak somewhere around the beginning of adolescence, declining slowly thereafter; the second is the product of the first factor. [Hebb, 1941c, p. 5]
Hebb (1942, p. 287) [...] concluded by saying “The contrast is not between intelligence and knowledge, but between capacity to develop new patterns of response and the functioning of those already developed.”
The two pages that Cattell sent to Hebb for his comments include the definition of fluid and crystallized intelligence as follows:
Adult mental ability is of two kinds, which may contingently be called ‘fluid’ and ‘crystallized.’ ‘Fluid’ ability is a truly general ability to discriminate and perceive relations between any fundaments, new or old. Crystallized ability consists of discriminatory response habits built up in a particular field, through the original operation of ‘fluid’ ability, but no longer requiring a true insightful perception for their successful functioning. Intelligence tests test, at all ages, the combined action of fluid and crystallized ability, recording the level of whichever is higher.
He continued to say that: “the level of the ‘crystallized’ abilities has been largely determined by the original level of the all-round, ‘fluid’ ability. ‘Crystallized’ ability is thus a dead coral formation revealing by its outlines the limits of the original living process. Loss of any system of discriminatory habits, e.g., through brain injury, is now likely to remain more localized, i.e., to affect general ability less, and to be less remediable. For as the tide of ‘fluid’ ability recedes after adolescence the ‘crystallized’ discriminatory powers persist at a level above that at which the fluid ability is capable of rebuilding them.” (Hebb, 1941-1942, R.B. Cattell to Hebb, September 22, 1941)
Hebb showed Cattell's letter to his Head of Department, George Humphrey, and Humphrey then wrote to Cattell that “I do not think however that you should present this as your own hypothesis. It is in essence the theory on which Hebb has been experimenting all the year.” He continued to say that “Some hundreds of A.P.A. members heard Hebb put out the theory at the Evanston meeting, and I hope you will forgive me if I say that I am sure you would do your scientific reputation harm by publishing the paper in the form in which I have seen it, i.e., without giving him credit.” (Hebb, 1941-1942, G. Humphrey to R. B. Cattell, September 24, 1941)
[...] Cattell also replied to Humphrey to explain how he developed his theory. He said that “I made what seemed to me a tenable, though possible premature, generalization on his data differentiating the adult abilities which do and do not deteriorate and asked in my last letter if he approved of the passage in which I described this as implicit in his writings.” Cattell went on to say that “The further theories are however quite independent, and to the best of my knowledge were neither in Hebb's paper nor in our after dinner conversation. My views were based on two pieces of evidence not in Hebb's field of research, and I had already discussed them with Thurstone and with Miles. Hebb's third source of evidence fitted in beautifully with these and completed the picture, but two thirds of the picture was already there. Naturally Hebb and I in discussion gained clarifications, but this would have happened if I had discussed the same matter with any other man of Hebb's intelligence and liking for the subject.” (Hebb, 1941-1942, R.B. Cattell to G. Humphrey, September 26, 1941)
Finally, Cattell agreed to credit Hebb, but suggested that he was being generous in doing so: “However, I trust that the revised statement which I have sent to Hebb will be agreeable to both of us, though I confess that on re-reading it I think it goes further in acknowledging priority of parts of the theory to Hebb than my reason alone indicates that it should.” (Hebb, 1941-1942, R.B. Cattell to G. Humphrey, September 26, 1941).
As noted by Horn (1966, p. 554), Cattell did not do any follow-up research on the theory of fluid and crystallized intelligence, but “left it on the shelf” until 1963, when he showed that two general ability factors could be derived from factoring culturally embedded and culture fair intelligence test results from a sample of 277 children in the 7th and 8th grades (Cattell, 1963). In this paper he concluded that one of these factors “fits the crystallized ability factor measured in traditional intelligence tests and the other a fluid general ability measured in culture-fair intelligence tests” (p. 20). Cattell's student John Horn continued to develop the theory of crystallized and fluid intelligence (Horn and Cattell, 1966, 1967).