The Trouble with Multiple Intelligences

How seductive ideas undermine our understanding of intelligence and invite pseudoscientific thinking

Carl Hendrick has already written an excellent exploration of the problems with Gardner’s theory of multiple intelligences, and I felt sufficiently inspired to add my twopenn’orth. This one has footnotes!

---

In 1983, Howard Gardner gifted the world with his theory of multiple intelligences.¹ No longer were we stuck with the dreary notion of a single IQ, now we could all be geniuses, just in different ways. Struggle with algebra? No problem, you might be a musical or naturalistic prodigy. Can’t do crosswords? Perhaps you’re a kinaesthetic marvel. It was the academic equivalent of handing out participation trophies. The only snag is that none of it’s true. Gardner never tested his theory, never ran an experiment, and never produced a shred of empirical evidence. It was, quite literally, a bright idea he had in the bath. A charming and long-lived hunch, to be sure, but as even Gardner now admits, “I readily admit that the theory is no longer current.”²

Of course, he’s not quite ready to entirely abandon his theory. While he admits that he never carried out any experiments designed to test the theory, he is not willing to accept that it lacks empirical support. He says, “The theory is not experimental in the traditional sense … but it is strictly empirical, drawing on hundreds of findings from half-a-dozen fields of science.” I’m not at all sure this is good enough. Just because you’ve got a lot of data, it doesn’t mean that you know how to interpret it, nor that you understand what it’s telling you. Gardner’s argument is a classic closed circle: I’m right because I have a lot of data which says I’m right. How does he know the data is correct? Because he’s got a lot of it.

But maybe there’s a ray of hope. Recent research claims to provide much needed empirical support for multiple intelligences. After reviewing 318 neuroscience reports, one study concluded that “there is robust evidence that each intelligence possesses neural coherence”.³ That sounds pretty convincing, but what does it actually mean? Basically this: researchers trawled though neuroscience studies to find indications in cortical areas for each of Gardner’s eight intelligences. And guess what? Brain imaging reveals that people really do have brain regions dedicated to each area. Does this provide support for multiple intelligences? No. All it tells us is what we could have guessed: there are brain regions associated with musical ability, physical movement, communication, handling figures and so on. It really shouldn’t come as much of surprise to find out that mental abilities are located in the brain. This has never been in doubt. And this is the whole problem with Gardner’s theory. He himself admits that if he’d steered clear of the word ‘intelligence’ no one would have given his idea a second glance. It’s completely uncontroversial to say different people have different talents, but by calling these talents intelligences he poisoned the well of intelligence research and strayed into pseudoscience.

Too strong? Well, if you make an empirical claim, then it should be falsifiable. There comes a point when twisting your ideas to fit the facts becomes pseudoscience, otherwise we can all believe whatever the hell we like and damn the evidence. We can argue that what we like ‘works’ because we like it. And if it’s unsuccessful on verifiable metrics then the metrics are worthless. This is how closed circle arguments work: you can explain away any amount of disconfirming evidence as not fitting your paradigm. You’ve given yourself permission to ignore reality, and anyone who suggests you might not be wearing any clothes can be safely dismissed as having a fixed mindset. The redoubtable Richard Feynman put the problem like this:

It doesn’t make a difference how beautiful your guess is. It doesn’t make a difference how smart you are, who made the guess, or what his name is. If it disagrees with experiment, it’s wrong.”

If you can’t accept that there could be conditions in which your guess is wrong, there’s a good chance you’re fooling yourself.

Let’s be charitable for a moment and imagine Gardner was right. Suppose there really are eight (or more) distinct intelligences. What then? Would recognising that some children are “musical” while others are “naturalistic” revolutionise education? Just because someone has ‘naturalistic intelligence’, does this mean they need to be taught maths in a garden? Would possessing high ‘linguistic intelligence’ mean you wouldn’t appreciate art? As we know from research into learning styles, matching instruction to children’s perceived preferences only serves to narrow their experiences.⁴ The bottom line is this: the theory of multiple intelligences is conceptually confused, lacks experimental support, flies in the face of more generally accepted, mainstream scientific research on intelligence, predicts nothing in the real world and, most importantly, provides absolutely nothing of any practical value.

But what about emotional intelligence, or EQ (emotional quotient) as it’s sometimes called? On the face of it emotional intelligence – the ability to perceive, identify and manage emotions – sounds like a wonderful thing. Far nicer than the cold brutality of IQ. Although it made its first appearance in a 1964 paper by Michael Beldoch, emotional intelligence was popularised by Daniel Goleman in his 1995 book, Emotional Intelligence: Why It Can Matter More Than IQ. While Goleman has claimed that emotional intelligence accounts for up to 67% of the qualities deemed to be essential for leadership, this has been flatly contradicted by Flavia Cavazotte and her colleagues who found it to have no significance on tests of leadership qualities.⁵ Although there are some studies showing a positive correlation between EQ and performance at work, these are much less strong than the correlations between IQ and performance.

It seems that EQ doesn’t actually seem to tell us anything we don’t already know from looking at IQ and existing personality tests. It may not even be a thing. Much criticism has focused on the way emotional intelligence is measured. Basically, you answer a series of multiple choice questions and respond to statements with ‘never’, ‘rarely’, ‘sometimes’, ‘often’ or ‘consistently’. The obvious problem with this sort of assessment is that it’s very easy to predict the ‘right’ or socially acceptable answer. This methodology is hard to take seriously when compared to the rigour with which intelligence is measured. Despite the claims, there is very little correlation between emotional intelligence and job performance.⁶ So perhaps, at least for now, we’d be better consigning this to the pile of good guesses that disagree with experiment.

But what of other attempts to invent different intelligences? As far back as 1920, Edward Thorndike proposed that intelligence should be divided into three types: abstract (ideas), mechanical (objects) and social (people). Social intelligence is the ability “to act wisely in human relations”. This seems fair enough. We readily understand what is meant by the idea of social intelligence and will have a store of acquaintances who we might recognise as possessing the qualities encapsulated by it. But is “social intelligence just general intelligence applied to social situations”?⁷ In order for it to be accepted as a separate concept it would need to be measurable in a way that is both reliable and valid – that is to say, test scores would need to make meaningful predictions in the real world.

The first attempt to measure social intelligence was the George Washington Social Intelligence Test. Like IQ tests, it consisted of a variety of sub-tests including judgement in social situations, memory for names and faces, observation of human behaviour, recognition of the mental states behind words, and from facial expressions, social information and sense of humour. The problem was that tests score were so highly correlated with IQ scores that it became increasingly hard to argue that they are actually different things.

Later attempts to distinguish social intelligence from abstract intelligence relied on self and peer ratings of social competence and the judgement of interviewers. The finding that self and peer ratings correlated more with the interview judgements than IQ did is interesting, but hardly impressive.⁸ Essentially, all this tells us is that likability and intelligence are not synonymous, which will surprise no one. The best we can say is that on the basis of intuitive evaluations of folk psychology, social intelligence ‘feels right’. It also probably corresponds with various personality traits such as openness to experience and extroversion.

Unlike social intelligence, Thorndike’s notion of practical intelligence is much harder to pin down and at least 27 different attributes have been bundled under the label.⁹ The difficulty with accepting practical intelligence as something separate is that it already seems to be included in Linda Gottfredson’s definition of intelligence as

…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 do¹⁰

These things all seem nothing if not practical.

On the other hand, some have argued that while IQ tests are useful for identifying individuals at the extremes of the distribution of ability, they’re not much help in differentiating between the vast majority whose scores fall between 85 and 115. What truly distinguishes these people is the quality and quantity of what they know. Exponents of practical intelligence argue that what we know, as well as our personality and interests, should be taken into account when trying to establish how intelligent a person might be in everyday circumstances, and that attempts to measure such a construct should focus on what’s typical rather than what’s possible.¹¹ In other words, it’s more useful to know what someone is normally like rather than what they could achieve with a downhill slope and fair trailing wind.

My real problem with slapping the label ‘intelligence’ on things that aren’t intelligence is that as well as devaluing the currency, you just end up with conceptual confusion. Howard Gardner acknowledged that he very deliberately labelled his different categories of accomplishments as intelligences rather than talents to challenge “those psychologists who believed that they owned the word ‘intelligence’ and had a monopoly on its definition and measurement”.¹² It’s probably also true to say that if he hadn’t, no one would have been interested.

1. Howard Gardner’s Frames of Mind (1983) proposed that intelligence is not a single general ability but a constellation of distinct capacities each representing a different way of processing information and solving problems. The critical reception was mixed. Gardner’s theory of multiple intelligences was praised for challenging narrow, IQ-based conceptions of intelligence and for its appeal to educators seeking a broader understanding of human potential. However, psychologists criticised it for lacking empirical support and for blurring the distinction between talents, personality traits, and cognitive abilities. Despite limited validation in research, the book had immense influence in education, where its ideas were enthusiastically adopted, often in ways Gardner himself later regarded as oversimplified.

2. Howard Gardner, Multiple Intelligences: Prelude, Theory, and Aftermath. In Robert J. Sternberg, Susan T. Fiske and Donald J. Foss (eds), Scientists Making a Difference: One Hundred Eminent Behavioral and Brain Scientists Talk About Their Most Important Contributions. Gardner’s account exposes the internal incoherence of his theory. He insists it is empirical yet admits it was never experimentally tested; he defines each “intelligence” as distinct but simultaneously accepts that they overlap in practice; and while claiming to challenge psychometrics, he offers no alternative means of reliable measurement. His retrospective defence blurs observation with evidence, intuition with proof, and ultimately reveals that multiple intelligences rests more on conceptual appeal than on scientific coherence.

3. C. Branton Shearer and Jessica M. Karanian, The Neuroscience of Intelligence: Empirical Support for the Theory of Multiple Intelligences?, Trends in Neuroscience and Education 6 (2017): 211–223 at 212. Shearer and Karanian reviewed neuroscientific evidence relating to Gardner’s multiple intelligences theory, finding some support for the idea that distinct neural networks underlie different cognitive abilities, though they concluded that the evidence remained partial and did not fully validate the theory’s independence of intelligences.

4. Pashler, H., McDaniel, M., Rohrer, D., & Bjork, R. A. (2008). Learning Styles: Concepts and Evidence. Psychological Science in the Public Interest, 9(3), 105–119. This influential review concluded that there is no empirical evidence supporting the idea that tailoring instruction to students’ preferred learning styles improves learning outcomes, and warned that such practices risk limiting rather than enriching students’ experiences.

5. Flavia Cavazotte, Valter Moreno and Mateus Hickmann, Effects of Leader Intelligence, Personality and Emotional Intelligence on Transformational Leadership and Managerial Performance, Leadership Quarterly 23(3) (2012): 443–455.

6. Dana Joseph and Daniel A. Newman Emotional Intelligence: An Integrative Meta- Analysis and Cascading Model, Journal of Applied Psychology 95(1) (2010): 54–78. Joseph and Newman conducted a meta-analysis of studies on emotional intelligence (EI) and proposed a “cascading model” linking emotion perception, understanding, and regulation. They found that EI predicts job performance primarily through its influence on emotion regulation and interpersonal effectiveness, though its effects are modest once personality and cognitive ability are controlled. Their findings suggest that emotional intelligence overlaps substantially with established psychological constructs rather than representing a distinct form of intelligence.

7. David Weschler, The Measurement and Appraisal of Adult Intelligence, 4th edn (Baltimore, MD: Williams & Wilkins, 1958), p. 75. Wechsler defines intelligence as “the global capacity of the individual to act purposefully, to think rationally, and to deal effectively with his environment.” This broad, functional definition emphasises intelligence as an integrated, adaptive quality rather than a narrow set of academic skills.

8. Martin E. Ford and Marie S. Tisak, A Further Search for Social Intelligence, Journal of Educational Psychology 75(2) (1983): 196–206. Ford and Tisak investigated whether social intelligence could be empirically distinguished from general cognitive ability. Through a series of experimental tasks assessing social judgment and problem-solving, they found that measures of social intelligence correlated strongly with traditional IQ scores, suggesting little evidence for it as a separate construct. Their findings supported the view that so-called “social intelligence” largely reflects the application of general intelligence to social contexts rather than an independent cognitive domain.

9. Robert J. Sternberg and Cynthia A. Berg, Quantitative Integration: Definitions of Intelligence. A Comparison of the 1921 and 1986 Symposia. In Robert J. Sternberg and Douglas K. Detterman (eds), What is Intelligence? Contemporary Viewpoints On Its Nature and Definition (Norwood, NJ: Ablex 1986), pp. 155–162.

10. As this wikipedia article makes clear, Gottfredson’s definition is not without controversy. Mainstream Science on Intelligence is a 1994 public statement, drafted by psychologist Linda Gottfredson and signed by 52 researchers, summarising the scientific consensus on human intelligence. It affirms that intelligence is a general mental capability involving reasoning, problem-solving, and learning; that IQ tests measure this ability with reasonable accuracy; and that intelligence predicts a wide range of life outcomes, including academic and job performance. The statement also asserts that intelligence is influenced by both genetic and environmental factors, that group differences exist, and that no social or educational policy has yet been shown to eliminate them. The controversy this statement centred on its perceived defence of ideas associated with The Bell Curve, particularly claims about group differences in IQ. Critics accused the statement of legitimising racial and class-based interpretations of intelligence research, arguing that it blurred scientific findings with social ideology. Many objected to its tone of certainty about issues still under debate, especially the heritability of intelligence and the causes of group disparities. Supporters saw it as a necessary correction to public misunderstanding and political distortion of intelligence research, while opponents viewed it as an attempt to cloak contentious, socially charged claims in the authority of science. Current consensus holds that intelligence reflects a real, measurable general ability (g) underlying reasoning, problem-solving, and learning, with substantial heritability that increases with age but remains shaped by environment. Genetic influences are highly polygenic and interact with social and developmental factors, while environmental effects—such as education, health, and deprivation—play a significant moderating role. Although individual differences in intelligence are partly heritable, average group differences are not considered genetically determined. Most researchers now reject simplistic or categorical models, recognising intelligence as a complex, multifactorial construct integrating both biological and environmental dimensions.

11. Phillip L. Ackerman and Eric D. Heggestad, Intelligence, Personality, and Interests: Evidence for Overlapping Traits, Psychological Bulletin 121(2) (1997): 219–245. Ackerman and Heggestad proposed that intelligence, personality, and vocational interests are interrelated rather than independent traits. Drawing on meta-analytic evidence, they argued that cognitive ability correlates meaningfully with personality dimensions such as openness and conscientiousness, and with interests aligned to intellectual engagement. Their model suggested that these domains together shape patterns of performance and achievement, highlighting the importance of considering motivational and dispositional factors alongside cognitive ability in understanding individual differences.

12. Robert J. Sternberg and Cynthia A. Berg, Quantitative Integration: Definitions of Intelligence. A Comparison of the 1921 and 1986 Symposia. In Robert J. Sternberg and Douglas K. Detterman (eds), What is Intelligence? Contemporary. While Gardner argues that intelligence is best understood as a set of distinct capacities rather than a single measurable trait, Sternberg and Berg advocate a quantitative, integrative approach, viewing intelligence as a unified construct that can be meaningfully assessed across domains rather than divided into separate “intelligences.”