Aphantasia Undercuts 300-Year-Old Theory of How We Think

Close your eyes and think about a triangle. For most people, something shows up. Not a diagram exactly, more a fleeting scribble of three sides, maybe chalk on a board, maybe just a reddish flicker that vanishes when you reach for it.

David Hume, writing in his Edinburgh lodgings in the late 1730s, was certain that flicker was not incidental. It was the machinery. Thought, he argued, could not proceed at all without some particular image to lean on. To think about triangles in general, you first had to summon a specific one and hold it, ready to swap it for another as the reasoning demanded.

But about one person in a hundred summons nothing. Ask them to picture a rainbow, a friend’s face, the bedroom they grew up in, and the screen stays dark. This is severe aphantasia, and the people who have it are not confused about what ‘s being asked. They simply have no mental pictures. And here’s the awkward part for Hume: they reason about rainbows perfectly well, recognise their friends, work through geometry and moral philosophy and long division without any apparent handicap. If the pictures were the machinery, the machinery is missing, and the mind keeps running regardless.

That tension is the argument of a new paper in Neuropsychologia by Uku Tooming and Roomet Jakapi, two philosophers at the University of Tartu in Estonia. Their claim is modest and large at once. As Tooming puts it: “If abstract thought genuinely required mental imagery, people with aphantasia should struggle to think abstractly. They do not.” Take that seriously, and one of the most carefully built accounts of thinking in the whole empiricist tradition springs a leak.

What Hume Thought He Had Solved

To see why this stings, you have to appreciate how much work Hume made images do. His entire picture of the mind rests on what scholars call the Copy Principle: every simple idea is a faint copy of some earlier sensory impression. You can’t think scarlet without having seen scarlet; the idea is just the impression run back at lower volume. Abstraction was the hard case, the place where this scheme threatened to buckle, because concepts like “triangle” or “friendship” plainly don’t correspond to any single sight or sound. Hume’s fix, borrowed partly from Berkeley and then given a proper engine, was clever. There is no abstract idea floating free of the senses. There is only a particular image, a specific triangle, which the mind has trained itself to treat as a stand-in for all the others, staying ready to call up any of them as needed.

Tooming and Jakapi translate this into modern terms as projective imagination, a specific mental move: you generate one image, then use it to reach for others in the same category, filtering out the strays your mind throws up (the stray square, say, when you wanted triangles). It’s not merely having a picture. It’s doing something with the picture, deliberately, under control. Which is exactly the capacity severe aphantasics seem to lack.

Four Escape Routes, All Blocked

A committed Humean has options, and the philosophers work through them one by one, like someone checking every exit in a locked room. Perhaps aphantasics just use imagery from other senses, sound or touch instead of sight? No good: there are multimodal aphantasics whose imagery is absent across the board. Maybe involuntary images, the ones that surface unbidden, do the job? Also no, because Hume’s abstraction needs the mind to steer, and involuntary imagery by definition does not steer. Spatial imagery, then, a sense of shape without the sensory paint? That one falters because spatial representation is too thin, too non-sensory, to be the vivid particular Hume insisted on, and anyway some people seem to lack even that. And the tempting idea that aphantasics just reason with the word “triangle” as a placeholder collapses on Hume’s own terms: for him, words without underlying ideas are close to nonsense.

The most interesting defence is subtler, and it leans on real and startling neuroscience. Maybe aphantasics do have imagery, they’re just not conscious of it. Some recent studies point that way. Brain-imaging work by Chang and colleagues this year found that content could still be decoded from the early visual cortex of aphantasics during imagery tasks, prompting them to coin the wonderfully paradoxical phrase “imageless imagery.” Other groups have reported intact-looking activation in visual regions even when subjects insist they see nothing.

Tooming and Jakapi are not persuaded, and their reasoning is careful rather than dismissive. For one thing, the evidence cuts both ways: a 2025 priming experiment by Purkart’s team found no hidden imagery effects at all, conscious or otherwise, and physiological markers like pupil response and skin conductance tend to fall flat in aphantasics too. For another, and this is the deeper point, having a buried representation is not the same as being able to use it in the demanding, projective way abstraction requires. An EEG study this year found aphantasics show reduced frontal delta activity, the sort involved in filtering and juggling internally generated images. Even if something image-like is down there, it may be exactly the wrong sort of thing, unable to do the job Hume needed it to do.

Which leaves the last refuge, the shrug. Fine, says the imagined Humean, aphantasia is a curious exception, but the theory still works for the other ninety-nine per cent. Hume himself pulled this move once, with his famous “missing shade of blue,” a colour you could supposedly imagine without ever having seen it. He waved it away as a one-off too trivial to matter. The Tartu pair argue aphantasia cannot be waved away so easily. Filling in a single unseen shade is a narrow trick. Abstraction, by contrast, underpins nearly all higher thought, so to say aphantasics do it by some entirely separate mechanism is to claim their whole cognitive architecture differs from everyone else’s. That is not a small exception. That is two kinds of human mind, which nobody really believes.

The reach of all this extends well past one eighteenth-century Scot. Image-based theories of thinking never went away; they run through modern cognitive science and turn up in contemporary accounts of concepts, such as Jesse Prinz’s “proxytypes,” which inherit the same vulnerability. The uncomfortable possibility the paper leaves open is that the mind is simply more various than our tidiest theories allow, that there may be more than one route to the same abstract thought, and that the pictures some of us find so vivid were never the point. For the millions who have spent their lives assuming everyone else’s mind’s eye was a figure of speech, that will not come as news. The rest of us may need a moment to picture it.

Frequently Asked Questions