Decades ago, I first read Stephen Jay Gould’s The Mismeasure of Man. The book was published first in 1981, then a revised edition in 1996 which included essays critical of the 1994 book The Bell Curve. It was this second edit that led a colleague to tell me is was “one of the most profound books ever written.” I reread the book, and agreed, but I wasn’t until I started listening to the book in 2020 that I really got the message.
I believe the ideas in the book influenced my thinking about learning and teaching since I first read it. In the time since generative AI has been causing me to think much more carefully about intelligence and the many roles the word and its definitions affect what we do.
Gould’s book is about intelligence as defined early in the 20th century. The psychologists who worked in the field were dedicated to making intelligence a “real” science in the way physics is with objectively defined quantities that describe how brains work in the same way physical laws affect inanimate objects. Part of this work led these scholars to advocate public policies based on their ideas; we now know those ideas are false, but they continue to influence society. Those influences are as disastrous today as they were back then. Probably more disastrous as we know better now.
According to Gould, we should reject intelligence that is:
- A single, abstractable entity—Gould contends that the dominant view treats intelligence as a “single entity” or a “central stuff” that can be isolated and measured. He refers to this tendency to convert an abstract concept like intelligence into a “hard entity” as reification. He emphasizes that this view is philosophically contrary to the idea of “richly multiple and independent intelligences.”
- Locatable within the brain—This reified “thing” or “central stuff” is presumed to have a physical substrate and a specific location within the brain. Gould critiques how factor analysis, particularly Spearman’s ‘g’, was used to provide a theoretical justification for this “unitary thing in the head.
- Quantifiable as a single number—A core aspect of the definition he opposes is the belief that intelligence can be quantified as “one number for each individual,” typically an IQ score. He highlights that Alfred Binet, the inventor of the original intelligence scale, explicitly denied that his tests permitted the “measure of the intelligence” as a single, superposable, or linear quantity like height.
- Used to rank individuals in a single hierarchy of worth—The quantified number is then used to “rank people in a single series of worthiness”. This often leads to a “dichotomization” of complex reality into divisions like “smart and stupid” or “normal vs. feeble-minded”. This ranking is frequently invoked to argue that “oppressed and disadvantaged groups—races, classes, or sexes—are innately inferior and deserve their status.”
- Innate and genetically based—A crucial element of the definition Gould argues against is that intelligence is inborn, inherited, and determined by genetics. H.H. Goddard, an American popularizer of the Binet scale, viewed intelligence as a “unitary mental process…conditioned by a nervous mechanism which is inborn,” and determined by chromosomes, suggesting it was “a single gene, undoubtedly recessive to normal intelligence”. Similarly, Lewis M. Terman, who developed the Stanford-Binet IQ test, believed in the “relatively greater importance of endowment over training as a determinant of an individual’s intellectual rank”. Cyril Burt also strongly advocated for intelligence as “innate, general, cognitive (i.g.c.) ability”.
- Fixed, immutable, and minimally alterable—The contested definition asserts that intelligence is a “fixed quantity, a quantity that cannot be increased” and is “but little affected by any later influences” or educational intervention. Gould points out the “brutal pessimism” of this belief, arguing it is “founded upon nothing”. He explicitly challenges the “false equation of ‘heritable’ with ‘permanent’ or ‘unchangeable'”.
In summary, Gould’s argument is primarily against the reductionist, hereditarian theory of intelligence that treats it as a single, innate, fixed, and quantifiable entity residing in the brain, used to create linear hierarchies of human worth.
Gould’s definition arises from the fact that intelligence as first conceived in psychology was a human trait. We know this definition has been rejected by careful study. Increasingly, I see intelligence as a meaningless term; as originally conceived it is false. It seems whatever large language models are doing, they are very useful, but seem to lack this thing we have called intelligence. Whatever we teach in schools, or more accurately whatever out student leave school with, is far more diverse than anything we can conceive as general intelligence.