Sit 39 university students in a darkened room, one at a time, and play them 36 classical piano excerpts — Bach, Chopin, Schoenberg, some major-key and flowing, some minor-key and dissonant. Show them photographs too, from peaceful meadows to erupting volcanoes. Don’t ask them to rate how much they like each one. Ask them to rate it on two separate scales at once: how beautiful, and how sublime. That’s the design behind Cathedrals of Sound: Predictors of the Sublime and the Beautiful in Music, Images, and Music With Images, published Sep 30, 2026 in Psychology of Aesthetics, Creativity, and the Arts by Young-Jin Hur, Radvile Medeisyte, and Chris McManus (University College London).

The result, in short: minor-key pieces full of escalating dissonance and rapid, unresolved transitions got rated sublime. Major-key pieces that flowed and resolved cleanly got rated beautiful. And when music and image competed for the same judgment, the image won — its pull on the combined rating ran up to three times stronger than the music’s. Edmund Burke never heard a Schoenberg excerpt. He died in 1797, decades before Schoenberg wrote a note. But in 1757 he already insisted that beauty and sublimity were not two points on the same scale — not “pretty” and “prettier,” not calm and calmer — but two entirely different operations of the mind, built from different material. This study ran the first controlled test of that claim on actual music, and the two categories came apart exactly where Burke said they would.

The Quick Verdict

Question	What we found
What was studied?	Whether sublimity and beauty are driven by different features in music and images, rated on independent scales rather than one continuum
Who ran it?	Young-Jin Hur, Radvile Medeisyte, and Chris McManus at UCL
Where and when?	Psychology of Aesthetics, Creativity, and the Arts, published Sep 30, 2026; covered the same day by PsyPost
Sample	39 university students, rating stimuli in a darkened room
Stimuli	36 classical piano excerpts (Bach, Chopin, Schoenberg) and a set of photographs, rated alone and paired together
Core finding on music	Minor-key pieces with escalating dissonance and rapid transitions rated sublime; major-key, flowing, harmonious pieces rated beautiful
Core finding on pairing	When music and images were judged together, the visual component dominated the combined rating over the musical one
What Burke already argued	A Philosophical Enquiry into the Origin of Our Ideas of the Sublime and Beautiful (1757): sublimity and beauty are categorically different responses, not two ends of one scale

The Quick Version

A UCL study published Sep 30, 2026 had 39 students rate 36 piano excerpts and a set of images on two separate scales — sublimity and beauty — rather than one scale from “pleasant” to “unpleasant.” Minor-key, dissonant, unpredictable music came out sublime. Major-key, flowing, harmonious music came out beautiful. When music and image were judged together, the image dominated. Edmund Burke argued in 1757 that these were never the same feeling turned up or down — beauty runs on smallness and harmony, sublimity runs on terror held at a safe distance — two centuries before anyone could run the numbers on it.

What the Study Actually Found

The interesting choice in this design isn’t the music. It’s the measurement. Most aesthetic-response studies put participants on a single dial — rate this from unpleasant to pleasant, or from disliked to liked. Hur, Medeisyte, and McManus didn’t do that. They had participants rate every excerpt and every image on sublimity and beauty as two separate grids, which means a piece could score high on both, low on both, or high on one and low on the other. That design choice is itself a wager on Burke being right: if sublime and beautiful were really the same feeling at different intensities, the two scales would just track each other. They didn’t.

For the music, key signature carried almost the whole effect. Minor-key excerpts — marked by escalating dissonance, rapid transitions, and what the researchers describe as emotional ambiguity — scored high on sublimity. Major-key excerpts that were flowing, elegant, and harmonious scored high on beauty. Tempo barely mattered. Neither did era, really; a Schoenberg piece and a Bach piece could land in the same place if they shared the same harmonic character. What mattered was whether the music resolved or didn’t, whether it sat still or kept pulling the ear somewhere unstable.

The pairing phase is where the study gets sharper. When participants rated music and image together, the visual half of the pair dominated — its influence on the combined sublimity-and-beauty rating ran up to three times stronger than the music’s. A calm image paired with unsettling music still tended to read as calm. That’s not a small asterisk. It’s a finding about which sense gets final say when two art forms are arguing with each other, and it’s one Burke, writing before recorded sound existed at all, had no way to anticipate.

What Did Burke Mean by the Sublime and the Beautiful?

Burke’s sublime is the response to vastness, obscurity, power, or danger encountered at a safe distance — a species of delight mixed with terror, what he called “delightful horror.” His beautiful is the response to smallness, smoothness, gradual variation, and harmony — the feeling that produces love and sociability, not awe. He treated these as two structurally distinct passions, not two strengths of the same one.

The line that makes the whole theory click is this: “When danger or pain press too nearly, they are incapable of giving any delight, and are simply terrible; but at certain distances, and with certain modifications, they may be, and they are delightful.” Push the threat close enough and it’s just fear. Pull it back to a safe remove — a storm watched from indoors, a minor-key passage that never resolves but can’t actually hurt you — and the same material becomes sublime. Burke called the governing emotion astonishment, “that state of the soul in which all its motions are suspended, with some degree of horror.” And he was blunt about what drives it: “terror is in all cases whatsoever, either more openly or latently, the ruling principle of the sublime.”

Beauty, in his account, runs on nearly the opposite machinery. “By beauty I mean that quality or those qualities in bodies by which they cause love, or some passion similar to it,” he wrote — and the qualities he lists are smallness, smoothness, delicacy, gradual variation. Nothing that holds you at arm’s length. Nothing that suspends the mind in dread. A beautiful thing invites you closer. A sublime thing works precisely because it doesn’t.

Why Does Sad Music Feel Sublime, Not Just Sad?

1. Minor key alone does a surprising amount of work. The study found key signature was the dominant predictor of sublimity ratings — more than tempo, more than era, more than how technically complex the piece was.
2. Unresolved tension reads as danger the ear can’t fully predict. Escalating dissonance and rapid transitions keep the listener’s expectations off balance, which is close to Burke’s own account of obscurity and uncertainty as sublime triggers — the mind straining to grasp something that keeps shifting.
3. None of it can actually hurt you. A dissonant chord has no teeth. That’s exactly the “safe distance” Burke described — the emotional shape of danger, imported into a context where the real stakes have been removed.
4. Ambiguity is doing something major-key resolution doesn’t. A piece that keeps you uncertain where it’s headed occupies the mind the way Burke said astonishment does — suspending its ordinary motions rather than satisfying them.
5. Sad and sublime aren’t the same category, even when they share a minor key. Sadness has an object and a resolution logic. The sublime, in Burke’s account and in this study’s ratings, is about scale and unresolved power — which is why a piece can feel vast and unsettled without feeling sad at all.

The Gap Burke Didn’t See Coming

Here’s where the study complicates the theory rather than just confirming it, and that’s the more interesting half.

Burke built his account around a mind encountering one thing at a time — a cliff edge, a dark forest, a single object vast enough to overwhelm. He had no occasion to ask what happens when two art forms compete for the same judgment simultaneously, because the technology to pose that question, recorded music paired with a controlled image, didn’t exist yet. The study’s pairing phase found something Burke’s framework doesn’t predict on its own: the visual channel doesn’t just add to the auditory one, it overrides it, by as much as three to one. A composer scoring a calm scene with dissonant, sublime music is fighting an uphill battle against the eye, at least by this study’s measure.

That’s not a small footnote. It suggests the sublime and the beautiful aren’t just categorically different feelings, as Burke argued — they’re also differently weighted depending on which sense is carrying them, and sight appears to be carrying more of the vote than Burke’s largely vision-centered examples (cliffs, oceans, ruins) would have led anyone to expect music to concede.

Why the Study and the Philosophy Rhyme

The connection isn’t “an eighteenth-century politician predicted a 2026 music study.” It’s narrower than that, and it holds up better for being narrower: Burke’s specific claim was that sublimity and beauty are driven by different features entirely — terror and vastness for one, smallness and harmony for the other — not degrees of the same pleasantness. This study measured two independent grids against 36 pieces of music and found the features split exactly along Burke’s line: minor-key dissonance tracking sublimity, major-key resolution tracking beauty, with almost no overlap in what predicted each.

It connects to what Aristotle meant by thaumazein, the wonder he placed at the root of philosophy itself. Aristotle’s wonder and Burke’s sublime aren’t identical, but they share a structure: both require something that exceeds easy comprehension, something the mind can’t immediately close its hand around. A piece of music that resolves cleanly doesn’t produce either. One that keeps slipping away from resolution might produce both.

It rhymes, too, with Empedocles’ account of shared awe dissolving the boundary between people — awe as something that doesn’t just happen to an individual mind but reorganizes it in relation to something larger. Burke’s astonishment, “that state of the soul in which all its motions are suspended,” is awe’s close relative, minus the social dimension that study measured. Both describe a mind temporarily overwhelmed rather than simply pleased.

There’s a sharper fit in what Stoic courage says about fear being part of the point, not a failure to overcome. Aristotle’s courageous person feels real fear and acts anyway; Burke’s sublime listener feels real terror’s emotional shape and finds delight in it precisely because the danger has been defanged by distance. Neither is describing the absence of fear. Both are describing fear doing something other than simply hurting.

And it’s worth setting beside what Walter Benjamin argued about aura and mechanical reproduction — his claim that something about a work’s presence resists being fully captured by a copy, however technically accurate. Burke’s sublime depends on a related kind of irreducibility: a dissonant chord can be described, scored, transcribed, and the description will still miss the astonishment it produces live in a dark room. The study measured the effect. It didn’t, and couldn’t, fully explain why the effect survives being written down.

What This Study Doesn’t Show

It’s worth being precise about the size of this, because the framing invites more confidence than the data earns.

A Small, Specific Sample

Thirty-nine university students in a lab is not a population that stands in for how people in general respond to music and images. It’s a tightly controlled first look, not a population-level claim.

A Narrow Musical Tradition

The excerpts came from Bach, Chopin, and Schoenberg — Western classical piano, spanning Baroque to Modernist. Whether the same key-signature pattern holds for other instruments, other traditions, or music without a Western tonal system at all is a separate, untested question.

Correlational Within the Design

The study found that minor key, dissonance, and rapid transitions predicted sublimity ratings. It didn’t manipulate one feature in isolation while holding every other musical variable constant, so the predictors are strong correlates within this stimulus set, not proof that any single feature alone causes the sublime response.

Self-Report on a Grid

Participants were rating their own felt response on researcher-defined scales. That’s a reasonable way to test whether sublime and beautiful behave as separate constructs, but it’s still self-report, taken in a single darkened room, not a physiological measure of what the mind is actually doing underneath the rating.

What Actually Follows From This

1. Stop treating “beautiful” as the compliment and “sublime” as a bigger version of it. The study’s whole design depended on these being different axes, and the results backed that up — a piece can be sublime without being pretty, and beautiful without being awesome, because they’re answering different questions.
2. Notice what your own strong aesthetic reactions are actually running on. If something moves you and it’s calm and resolved, that’s probably closer to Burke’s beautiful. If it unsettles you and you can’t look away, that’s closer to his sublime — and conflating the two makes it harder to say what you actually want more of.
3. If you’re pairing media, expect the eye to win. The study found the visual component dominated combined ratings by as much as three to one — composers, filmmakers, and anyone building a multisensory experience are fighting that imbalance whether they know Burke’s name or not.

Burke never heard a minor-key piano excerpt engineered for a psychology lab, and he never had to reckon with an image overpowering a soundtrack in a controlled trial. He had a question he found genuinely strange — why does a thing that should only frighten us sometimes delight us instead — and an answer precise enough to still be testable two hundred and sixty-nine years later. The sublime needs a little danger, held at a distance long enough to admire. The beautiful needs none at all. A 2026 study ran the numbers and found the split exactly where he said it would be.

---

This post discusses a 2026 experimental study published in Psychology of Aesthetics, Creativity, and the Arts and Edmund Burke’s 1757 philosophical treatise. The study’s findings come from 39 university students rating a specific set of Western classical piano excerpts and images in a controlled lab setting — they describe patterns within that stimulus set and sample, not a universal law of music, a claim about non-Western musical traditions, or a physiological account of what produces the sublime.