“When I see equations, I see the letters in colors — I don’t know why. As I’m talking, I see vague pictures of Bessel functions from Jahnke and Ernde’s book, with light-tan j’s, slightly violet-bluish n’s, and dark brown x’s flying around. And I wonder what the hell it must look like to the students.”— Richard Feynman

The learning styles theory suggests that people have preferred ways of receiving information (visual, auditory, or kinesthetic, aka VAK) and that instruction should be matched to these preferences for optimal learning. It’s an appealing idea: identify each student’s learning style, tailor your teaching accordingly, and watch learning improve. The only problem? Decades of research have failed to find evidence that this matching actually works. Study after study shows that matching instruction to supposed learning styles doesn’t improve learning outcomes. The matching hypothesis is dead.

This point has been made over and over again. Articles debunking learning styles appear regularly, each one hammering the same nail. And yet some forms of the myth persist, and it’s even baked into AI systems, as I demonstrated in 2023 when I showed how language models cheerfully generate learning styles content, complete with perfectly formatted tables matching teaching strategies to VAK categories.

And I’ve noticed something else: there’s a certain glee researchers and instructional designers take in debunking all this. A triumphant “gotcha!” aimed at teachers who’ve been duped by pseudoscience. I get it. The research is clear. The debunking is deserved. Message received.

But here’s what bothers me about this repeated, gleeful dismissal: we might be throwing out something genuinely important along with the pedagogical snake oil of learning styles.

We can reject pseudoscience without becoming reductive about how minds work.

I should know. I’ve been guilty of this myself. Back in 2009, I wrote a blog post titled “Teaching to Learning Styles: What Hogwash” where I enthusiastically joined the debunking bandwagon. I wasn’t wrong about the research. The matching hypothesis remains bunk. But I think in our rush to debunk learning style we may have thrown out the baby with the bathwater. We may have in the process, missed certain critical aspects of what it means to learn.

Some context setting is needed.

A few years ago, my colleagues and I studied the four mathematicians who won the Fields Medal in 2014, the closest thing mathematics has to a Nobel Prize, in an article titled “Creativity in Mathematics and Beyond – Learning from Fields Medal Winners.” What we found was striking: each of them had developed a completely different way of engaging with mathematics. Manjul Bhargava finds mathematics in tabla rhythms and Rubik’s Cubes, turning to his drums when stuck because “Indian classical music, like number theory research, is largely improvisational.” Maryam Mirzakhani sprawls giant sheets of paper across her floor, sketching geometric structures creating “elaborate stories” with mathematical characters that evolve and surprise her. So much so that her daughter thought she was a painter. Martin Hairer hears “music in noisy equations” and created award-winning DJ software. Artur Avila calms chaos through rapid visualization. (You can read more about their approaches in our full article.)

In short, these are not just different “preferences,” they represent fundamentally different cognitive approaches to mathematical problem solving, what my colleagues and I have called transdisciplinary cognitive skills, ways of thinking that cut across traditional boundaries through perceiving, patterning, embodied thinking, and more.

If this is true of mathematics, what about other domains?

This perspective is not new. Einstein imagined riding beams of light. Feynman saw equations in colors because of a book he had read in his childhood. Darwin sketched crude evolutionary trees in his notebooks, working out species relationships visually with his famous “I think” diagram. In addition, he walked his daily “thinking path” at his estate, using the slow, methodical rhythm to work out evolutionary theory. Barbara McClintock developed what colleagues called “a feeling for the organism,” talking to her corn plants and intuiting genetic patterns decades before molecular biology could explain them.

These weren’t quirky personal details; they were integral to how these minds worked.

When we dismiss individual differences in thinking as “learning style myths,” we risk losing sight of these cognitive realities. Yes, there’s no such thing as a “kinesthetic learner” who can only learn through movement. But there absolutely are people who think more effectively through movement, spatial relationships, auditory patterns, or visual representations.

Here’s the thing: the people who speak of learning styles got the mechanism completely wrong, but they were responding to something real. When people insist that they’re “visual learners,” they might not be describing fixed learning channels. They might be recognizing certain genuine differences in how they process some kind of complex information, differences shaped by their backgrounds, experiences, and the cognitive approaches they’ve developed.

People really do have different thinking approaches, processing strengths, and problem-solving strategies, based on their experience, natural talents, and propensities. The original learning styles theory’s flaw wasn’t recognizing differences, it was oversimplified categorization and misguided instructional matching.

But there may also be a structural issue at play, one that explains why the myth persists despite being thoroughly debunked. Most teaching and learning happens (at least in schools) via lectures, auditory, passive, often monotonous. When instruction moves beyond the lecture format to forms that involve movement or visuals, it may just be the novelty of the experience that enhances learning and recollection for everyone, not just supposed “kinesthetic” or “visual” learners.

Students, in turn, may attribute this improvement to their “learning style” when it’s actually about breaking cognitive monotony. The brain responds to variety, novelty, and multi-modal engagement. When a student who’s been stuck in lecture mode for hours finally gets to physically manipulate something, their engagement spikes, not because they’re a “kinesthetic learner,” but because they’re a human learner whose brain lights up with novel, embodied activity.

This may explain both why the myth persists and why educators who diversify their instruction see real improvements. They’re just misattributing the mechanism.

You might be thinking: “So you’re just saying use varied teaching methods. How is this different from standard good practice?”

Fair question. And to be clear, varied teaching methods are genuinely valuable. In situations of uncertainty, when you don’t know what will resonate with learners, trying multiple approaches is a legitimate strategy. There’s nothing wrong with throwing ideas at the wall when you’re navigating complexity. This is particularly important when faced with a variety of learners in a classroom context, each with their own backgrounds, talents, interests, passions and commitments.

There’s also something else at play: the content itself often demands certain forms of representation. Geometry requires visual-spatial reasoning, you can’t really understand geometric relationships without seeing them, manipulating them, experiencing them spatially. Chemistry’s molecular structures need three-dimensional manipulation to grasp bonding and spatial arrangements. Music requires auditory processing and temporal pattern recognition. The visual arts demand, well, vision and spatial awareness. This isn’t about accommodating “visual learners” or “auditory learners,” it’s about the inherent nature of the disciplinary content. The mathematics Mirzakhani was working on was deeply geometric; her floor-scattered visual doodles weren’t personal preference, they were appropriate to the spatial nature of the problems she was solving.

But it goes even deeper than representation. Different disciplines fundamentally structure thinking differently. Years ago, I wrote about Janet Donald’s research showing that disciplines shape the very processes of description, inference, synthesis, and verification. Verification in engineering is pragmatic—does it work? In literature, it’s about interpretive coherence. Problem description in physics differs fundamentally from problem description in history. As Donald argues, there is no “royal road,” no single instructional approach that works equally well across all disciplines. The knowledge, methods, and ways of thinking that characterize each field demand different pedagogical approaches.

But what I’m arguing for goes deeper than either of these points. It’s about recognizing that learners come with particular cultural and experiential resources that make certain entry points more accessible at this moment in time, when they are facing some new ideas. It’s about understanding that engagement isn’t just nice-to-have; it’s the necessary gateway to learning. It’s about respecting that different cognitive approaches can yield genuinely different insights, not just different routes to the same destination. And it’s about building from where learners actually are, not where we think they should be.

Consider Bhargava’s musical mathematics. His approach didn’t emerge from some innate “auditory learning style.” It emerged from deep cultural grounding in Sanskrit poetry and Indian classical music, experiential resources that made certain mathematical patterns accessible to him as productive entry points in ways they might not be for others. An educator working with a student from a similar background might recognize: “Ah, rhythmic pattern could be an entry point here.”

That’s different from saying “this student is an auditory learner, so I’ll lecture at them.” It’s about discerning where students are, what resources they bring, and using those as legitimate starting points, not because they’re the only way that student can learn, but because that’s where engagement and meaning making can begin.

The Fields Medal winners suggest something different from fixed styles: not categories determining how someone should be taught, but flexible approaches individuals develop for engaging complex disciplinary material. Mirzakhani’s visual method wasn’t a stepping-stone to “real” mathematics, it WAS her mathematics, a sophisticated tool she developed for abstract geometry. She didn’t need to graduate to standard techniques; her floor-scattered doodles were generative and powerful on their own terms.

There is a reason why disciplines develop ways of representation, verification and more. Over time, learners might develop appreciation for standard techniques, for their power, parsimony, elegance. Or they might not, because they’ve found approaches that are fecund and powerful for them. There’s no single correct developmental trajectory. We are complicated. Learning is complicated.

The learning styles myth persists not because people are scientifically illiterate, but because it touches something true about cognitive diversity that our educational frameworks don’t adequately capture. People do experience thinking differently. They do find certain approaches more engaging. They do have genuine cognitive strengths shaped by their experiences and backgrounds.

What I’m arguing is that cognitive diversity is real, valuable, and worth understanding—not as fixed categories, but as the messy, contingent, culturally-grounded reality of how different minds engage challenging material. And that material itself is often better understood through multiple representational formats. We should create multiple pathways not because people have fixed learning channels, but because complex material can be approached from multiple angles, and different approaches yield different insights.

Yes, learning styles theory was pseudoscience. But we can reject bad science without embracing bad philosophy, the reductive view that cognitive differences don’t matter or are mere illusions.

The solution isn’t dismissing these experiences as myths. It’s developing more sophisticated understanding of cognitive diversity, recognizing the rich complexity of how different minds engage challenging material without relying on oversimplified categories or misguided matching theories.

When we reduce this to “learning styles are debunked,” we lose sight of what makes human cognition fascinating: there are multiple paths to understanding, multiple ways of thinking about problems, multiple cognitive strengths that can be developed and leveraged.

Einstein rode light waves in his imagination. Feynman saw colored equations. The 2014 Fields Medal winners found mathematics in music, stories, and floor-scattered doodles. Maybe it’s time we stopped treating cognitive diversity as a myth and started treating it as one of the most interesting, and challenging, things about how humans think and learn. We need to keep asking ourselves what Feynman wondered about his own thinking as he saw colored equations swirling around in his head, and wondered: “…what the hell it must look like to the students.”

---

Two Notes

1. On Media and Tools

There’s yet another layer of complexity worth noting: the media and tools we use to engage with content aren’t neutral containers—they prefigure and shape what’s cognitively possible. McLuhan’s “the medium is the message” applies here: different technologies and representational tools enable different ways of thinking and being. I’ve written about this here and here, but that’s a topic for another day.

2. On Feynman’s Cognitive Experiments

Since we started with Feynman’s quote about seeing equations in color, here’s another story he shared about discovering individual differences in thinking. In this fascinating video, Feynman describes experimenting with his internal time sense, learning to count exactly to one minute. He discovered he could read while counting but couldn’t speak, because his counting relied on an internal voice saying “one, two, three.”

When he shared this with mathematician John Tukey, Tukey was incredulous: “I don’t see why you would have any difficulty talking whatsoever, and I can’t possibly believe that you could read.” It turned out Tukey counted by visualizing a tape with numbers that went “clink, clink, clink”—a visual system that let him speak freely but prevented him from reading while counting.

Feynman’s conclusion: “That’s where I discovered—at least in this very simple operation of counting—the great difference in what goes on in a head when people think they’re doing the same thing.”

---

Note: On Studying “Exceptional” Minds

One criticism this post has received is that I’m drawing conclusions from special people—Einstein, Feynman, McClintock, Fields Medal winners—whose brains work differently, and there’s only so much we can learn from them that applies to ordinary learners.

This objection misunderstands why we study highly creative individuals. An earlier version of this post addressed this directly, but I edited it out for flow. Here’s what I had written:

“Before you dismiss this as irrelevant because these are exceptional minds, consider this: we don’t study highly creative individuals because they have “miraculous” powers, but because they offer us a window into cognitive processes available to all of us. Their thinking patterns, while perhaps more developed or refined, perhaps having a stronger foundation of knowledge and expertise, illuminate the diversity of ways human minds can engage with complex material.”

Let me be even more explicit: The point isn’t that every student can or should think like Feynman. The point is that if even Nobel laureates approach the same problems through fundamentally different cognitive pathways (visualizing vs. verbalizing, spatial vs. auditory, embodied vs. abstract), then it’s reasonable to expect similar diversity, perhaps less developed, perhaps more modest in scope, across all learners.

Feynman and Tukey’s different counting methods aren’t about genius, they’re about the most basic cognitive operation (counting to 60), yet even there, their minds worked completely differently. That’s the lesson: cognitive diversity exists at every level, not just at the heights of creative achievement.

The Fields Medal winners are useful precisely because their approaches are visible and articulated. They’ve reflected on and can describe how they think. Most of us haven’t developed that metacognitive awareness, but the underlying diversity in how we engage with material is still there. These exceptional thinkers help us see what’s often invisible in everyday learning.

Moreover, the cognitive approaches I’m describing aren’t mysterious gifts of genius. Robert and Michele Root-Bernstein, in their foundational work on creative thinking (in their book Sparks of Genius) further extended by Danah Henriksen and I, identified a set of transdisciplinary thinking tools (perceiving, patterning, abstracting, embodied thinking, modeling, playing, and synthesizing) that creative individuals use across all domains. We argue that these cognitive skills are available to all learners and can be developed through education. The Fields Medal winners simply offer vivid examples of what these thinking tools look like when fully developed.