A couple of weeks ago, I wrote a post about cognitive diversity (and learning styles), essentially arguing that in the process of debunking the learning styles myth we may have lost sight of a bigger issue, that of cognitive diversity. That post, and the ideas underlying it, had been on my mind for a while though I wasn’t sure anybody else would care.

So when I posted the link to my post on LinkedIn I was not expecting much traction, partly because it seemed such a niche topic and partly because all the world is talking about today is AI and just AI. (To be fair, I am quite guilty of the latter). But for whatever reason that post struck a chord… clearly there is some latent interest in cognitive diversity.

In a kind of Baader–Meinhof effect, once the idea of cognitive diversity took root in my mind, evidence of it began appearing all around me—in articles, conversations, even casual remarks I’d once have overlooked. The one that stood out the most, and the one that drives this follow-up post is a remarkable article in The New Yorker about aphantasia—the inability to form mental images—and its opposite, hyperphantasia. I had never heard of either of these phenomena. The piece, by Larissa MacFarquhar, follows several people who made startling late-life discoveries about how their own minds work fundamentally differently from others.

For instance, a physicist was 35 years old when he realized that other people could actually see pictures in their minds. He had always assumed “visualizing” was just a metaphor. Similarly, an Irish artist discovered in her forties that other people could picture things when they weren’t looking at them. At the other end of the spectrum are people with hyperphantasia. For instance, Clare Dudeney, a hyperphantasic artist, watches entire Harry Potter films in her head during meditation retreats. There now exists the Aphantasia Network consisting of hundreds of people who share these cognitive quirks.

Here’s what struck me: these are highly accomplished, intelligent people—physicists, artists, writers—who didn’t know this basic fact about their own cognitive architecture until decades into their lives. If such fundamental differences can remain invisible to us about our own minds, imagine the complexity hidden beneath the surface in a classroom of thirty students.

The challenge of cognitive diversity reminded me of an earlier discovery about physical diversity. I had written previously about Gilbert Daniels, the gardener who changed our world. Daniels, a 23-year-old Air Force lieutenant in the late 1940s, discovered that not a single pilot among 4,063 measured fit the “average” on just ten physical dimensions. His insight, that if you design for the average, you design for no one, transformed everything from cockpit design to car seats. We all have “jagged profiles” that never smoothly line up with every average measure.

But Daniels was working with visible dimensions: height, weight, arm length. Cognitive diversity, on the other hand, is invisible, happening entirely in the space between our ears. The jaggedness is even more profound here, more hidden, more variable.

Consider what we now know: Some people see vivid mental images; others see nothing at all. Some can replay music in their heads; others cannot. Some people’s memories are like immersive virtual reality experiences; others remember facts about their lives with no accompanying imagery or emotion, knowing things happened without being able to feel what it was like.

And here’s the kicker: many (if not all) of these people assumed everyone’s minds worked like theirs. They heard others talk about “picturing” things and thought it was just a figure of speech. They heard people describe “seeing” memories and assumed it was metaphorical exaggeration. The profound mutual incomprehension was invisible because the processes were invisible.

This isn’t about “learning styles,” the debunked idea that people have fixed channels for receiving information. This is about fundamental architectural differences in how minds work, differences so profound that people can live decades without realizing their cognitive experience differs radically from others’.

What does this mean for education? For theories of learning? For any attempt to generalize about “how people think and learn”?

If Gilbert Daniels taught us that physical jaggedness makes the “average person” a myth, recognizing cognitive diversity suggests that this jaggedness may be far more extreme. We’re not just talking about variations along a bell curve. We’re talking about qualitatively different mental architectures that people can’t even articulate to themselves, let alone to us.

And if we can have such profound differences in something as seemingly basic as whether we can see mental images, what other invisible variations are we missing? What other fundamental cognitive differences exist that we don’t even have language for yet, that people haven’t discovered about themselves?

This isn’t a call for despair about the possibility of educational research or theory. But it is a call for humility. For recognizing that when we make pronouncements about “how people learn,” we may be describing only one kind of cognitive architecture among many. For acknowledging that the invisible space between our ears is far more variable, far more jagged, than we’ve ever imagined. After all, if a 35-year-old physicist can discover he’s been living without mental images his entire life, what else don’t we know about how minds work? The question isn’t whether cognitive diversity exists. It’s what other dimensions of it we haven’t even noticed yet and what it means for us as educators and educational researchers.

And, mind you, we have just been speaking about the cognitive—what happens in individual brains. Learning, as we know, is so much more. We haven’t even begun talking about what happens when these complex internal worlds collide with each other (in classrooms and playgrounds and online) and with the broader social, economic, and cultural contexts within which learning happens—and the ways in which our genetics interact with culture and context and environment.

Is it any surprise, as David Berliner so often reminded us, why educational research is the hardest science of all.

---

Note 1: On AI Tutors and Autodidacts

I couldn’t help but notice (and yes, I promised myself I’d get through one post without mentioning AI, but here we are) that many of the most vocal proponents of AI as “the ultimate tutor” tend to be autodidacts—brilliant people who learned largely on their own, often despite rather than because of formal education. They’re extrapolating from their own experience, assuming their cognitive architecture is universal.

But here’s the thing: most of these folks likely have no idea how their own minds actually work. They’re successful autodidacts precisely because their particular cognitive setup—whatever it happens to be—meshes well with self-directed learning. How many AI enthusiasts know whether they’re aphantasic or hyperphantasic or even if these constructs exist? Whether they think in words or images or spatial relationships? Whether their memory works through vivid re-experiencing or abstract facts?

Introspection can only take you so far. We’re notoriously unreliable narrators of our own cognitive processes.

This is the complexity that teachers work with every day—rooms full of invisible cognitive diversity, jagged profiles they can’t see, mental architectures that even the students themselves don’t understand. When a teaching approach works beautifully for one student and falls flat for another, it’s not necessarily because the teacher did something wrong. It might be because one student has a mind that generates vivid mental imagery and the other has a mind that doesn’t, and neither the teacher nor the students know this fundamental fact.

Perhaps we could extend a bit more credit to teachers navigating this invisible complexity, and be more cautious about simplistic solutions (technological or otherwise) to highly complex problems.

And here’s the deeper irony: How did Nick Watkins, Isabel Nolan, and thousands of others finally discover how their minds worked? Not just through introspection. They discovered it by engaging with other people, reading someone else’s casual description of their mental experience and having that shock of recognition: “Wait, you can do what?”

It was human conversation, human writing, that revealed these invisible differences. People comparing notes, sharing experiences, being vulnerable about how their minds work, and discovering profound differences they’d never suspected. Genuine human engagement is what will help us discover the astonishing variety of ways minds can work. The richness we find there will probably complicate every simple theory we’ve ever had about learning.

---

Note 2: On Language Models and Models of Mind

I came across this piece after writing the post but had to mention it since it resonates with much of what I have written above. The author, who has synesthesia and dyscalculia, questions how much we can trust LLMs just because they use language. Here is an extended quote

Most of us produce language, and we assume others who produce language produce language in similar ways. When we assume language reflects thinking, we may also assume that all thinking reflects our thinking [italics in original]. This can lead us to the faulty conclusion that language reflects the journey of thought.

This is why making room for a neurodivergent understanding of “how we think” matters: even though I think in ways radically different from most other people in the world, I had no language to express this. Coming to grips with the ways we think, as opposed to the ways we are “expected” to think, can help unravel universalist assumptions about there being any one way to think at all.

This is exactly the trap. LLMs model one way of processing information—through language—and we’ve mistaken that for modeling thought itself. But as both the aphantasia research and this author’s experience show, vast domains of human cognition happen outside of or orthogonal to language: in images, in spatial relationships, across modalities (music or numbers as color) and more.

“Whatever we attach to language,” the author warns, “merits careful attention, because it turns out it doesn’t mean nearly as much as we think it does.”