This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American
On December 13, 1994, a group of fifty-two experts in the scientific study of intelligence and allied fields provided the following unified definition of intelligence in the Wall Street Journal:
Intelligence is a very general mental capability that, among other things, involves the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly and learn from experience. It is not merely book learning, a narrow academic skill, or test-taking smarts. Rather, it reflects a broader and deeper capability for comprehending our surroundings—“catching on,” “making sense” of things, or “figuring out” what to do.
This is a reasonable definition of intelligence. It includes a description of behaviors relating to attention, perception, and learning that we would certainly want to include as key aspects of intellectual functioning. What’s more, this definition captures how most people—especially in Western cultures conceptualize a “smart” person. When we talk about someone being “smarter” than someone else, we tend to invoke the notion of quick reasoning and problem solving. This conceptualization pervades Western media, such as Jeopardy and The Big Bang Theory.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
Another claim of the Wall Street Journal definition, however, is that intelligence, as measured by IQ tests, is a “very general mental capability.” There is one sense in which this statement is surely correct, but a few other senses in which we reach very thorny territory.
It is empirically true that the cognitive abilities that are measured on IQ tests are positively correlated with each other, giving rise to a "general intelligence factor".* It is also true that the cognitive skills that are most strongly related to IQ (e.g., abstract reasoning, working memory, vocabulary, visual-spatial mental rotation) are highly general in the sense that they facilitate the speed and efficiency of learning novel and complex information across a wide range of contexts. It's difficult to imagine a situation in which one is conscious and processing information that doesn't draw whatsoever on these cognitive skills.
However, IQ test scores are just that-- test scores. The scores themselves don't actually have any causal properties. To say that IQ itself is a very general capacity reifies the IQ test. It is becoming increasingly clear in the field of intelligence that IQ is best thought of as an emergent property (not cause) of a range of cognitive mechanisms that are positively related to each other, and influence each other during the course of development. In other words, IQ is a summary score, not the driver itself of cognitive potential.
But there’s another sense in which I think the framers of the WSJ definition may mean “general”, which gets us into even more complex territory. They point out that “IQ is related, probably more so than any other single measurable human trait, to many important educational, occupational, economic, and social outcomes.”
This is statistically true. IQ is part of a large “nexus” of positively correlated societal outcomes. IQcorrelates positively with family income, socioeconomic status, school and occupational performance, military training assignments, law-abidingness, healthful habits, illness, and morality. In contrast, IQis negatively correlated with welfare, psychopathology, crime, inattentiveness, boredom, delinquency, and poverty. The correlations exist. What remains far more ambiguous, in my view, is the interpretation of these correlations.
Tellingly, all of these outcomes are also correlated with one another, forming a large interconnected web of positively related variables. For instance, income is not only positively correlated with IQ, but is also positively correlated with amount of schooling, parental income, and parental social status— all of which are individually also positively correlated with IQ. This makes it extremely difficult to disentangle all these variables from each other to figure out what’s causing what.
To be clear: IQ tests do not simply index family background. IQ and SAT scores are still correlated with important academic and societal outcomes even after taking into account socioeconomic status.** But again, let’s not forget: IQ or SAT test performance is alsoan outcome—a person’s score on a brief test measured at one moment in time. As convenient as it would be for a person’s IQ score to measure something completely divorced from any learning, experience, and background, that’s not how IQ tests work, and that’s certainly not how the world works.
Any single person’s IQ test performance is itself a measure of many things in addition to their intellectual functioning across a range of cognition, including the person’s duration of education, motivation, level of test anxiety, self-belief, history of traumatic experience and opportunities, language, and many other things that are strongly connected to the very outcomes they are trying to predict.
The truth is, many factors operate together to produce a broad positive manifold of “desirable” social outcomes (e.g., high IQ, life expectancy, health, occupational success, etc.).*** Also, the path from IQ to academic achievement to occupational achievement to high achievement is far from direct. The rich get richer and the poor get poorer. Small differences multiply over the years to create large differences.
Ten-year-olds with good fluid reasoning skills and high executive functioning are a pleasure to teach, and they receive more challenging materials. They are expected to go to a good university. The positive expectations and increased intellectual enrichment only increase their chances of graduating from high school and getting into a prestigious university, which in turn will increase their chances of obtaining a “prestigious” job. Meanwhile, the child who early on finds school difficult, unrewarding, and perhaps even boring, is fed a less enriched curriculum, which exerts downward pressure on both expectations and achievement. As a result, the child is likely to get poorer and poorer grades, and leave school at the first opportunity.
In fact, it’s quite likely that the cognitive skills that comprise IQ are positively correlated with each other in large part because society chose them to represent the foundation of education. In other words, we are the ones that laid the groundwork for them to be positively correlated as they are mutually developed in school.
Indeed, statistical techniques such as "bivariate latent change score models" hold a lot of promise for furthering our understanding of how this may operate during the course of cognitive development. Recent research by Rogier Keivit and colleagues suggest that the "mutualism model"-- which proposes that basic cognitive abilities directly and positively interact during development-- provides the best account of developmental changes compared to more traditional notions about human intelligence.
Therefore, when having such discussion about the relationship between IQ and societal outcomes, I believe it's important to recognize that we can't simply divorce the former from the latter, as this sets up a false dichotomy. Yes, IQ tests measure an important set of cognitive abilities for dealing with on-the-spot novel and complex problem solving. But IQ is also a societal outcome, deeply intertwined with a larger nexus of other positive societal outcomes. When it comes to relationships between this societal outcome and other important societal outcomes, the data don't simply speak for themselves.
© 2018 Scott Barry Kaufman, All Rights Reserved
*The general factor of intelligence is technically defined as the first factor derived from a factor analysis of a diverse battery of cognitive tests, representing a diverse sample of the general population, explaining the largest source of variance in the dataset (typically about 50 percent of the variance). See “Human Intelligence Differences: A Recent History”.
** The SAT is basically a thinly disguised IQ test.
*** Importantly, and often underappreciated, if you get a large enough number of participants and variables, then even the slightest positive correlation between one variable and the next will be statistically significant without necessarily being practically significant.
This article was adapted from Chapter 13 of Ungifted: Intelligence Redefined