The first time I heard it, I thought a string had broken.
I was practicing double stops — sixths, the Dont étude in G — and somewhere between the A on the E string and the C-sharp on the D string, a low hum appeared beneath both notes. Not a ringing. Not feedback. A tone, clear and low, that neither string was producing. I stopped playing and the tone vanished. I played again: there it was. A note below the staff, steady and sure, as if someone had opened an organ pipe under the floor.
I lifted the bow. Silence. I played the sixth again. The tone returned. I checked my intonation — the C-sharp was clean, the A was clean, the sixth was bright and narrow. The third sound had nothing to do with being out of tune. It appeared because the two notes were in tune.
My teacher would have called it a combination tone and moved on. The frequency of the lower note minus the frequency of the upper note produces a ghost — a phantom pitch generated not by the instrument but by the ear itself. Tartini described it in 1714. He called it the terzo suono, the third sound, and used it to check his students' intonation: when the double stop was pure, the combination tone appeared. When it wasn't, the tone was absent or wavering. A reliable guide. A tool.
He used it as a tool. I heard it as a question.
That night I read everything I could find. The physics were straightforward. Two frequencies entering the cochlea interact at the basilar membrane, producing a mechanical distortion — a nonlinearity in the ear's response. The distortion generates a signal at the difference frequency. If I play A at 880 hertz and C-sharp at 1100, the difference is 220: an A an octave below the one I'm playing. A phantom octave. The ear manufactures it from raw mechanics, then the auditory cortex obligingly presents it as a note, a real note with pitch and presence, as if it came from outside.
But it didn't come from outside. A microphone placed in the room would not capture it. A recording of the double stop, played back on speakers, would regenerate its own combination tones — different ones, colored by the speakers' own distortions and the listener's own ears. The third sound I heard was mine. The third sound you would hear in the same room would be yours. Same stimulus, different phantom.
The physics paper called this a "psychoacoustic artifact." I didn't love the word artifact. An artifact is something left behind by a process — a trace, a residue, an accident. The combination tone didn't feel accidental. It felt generated. My ear heard two notes and produced a third, the way a prism takes white light and produces a spectrum. The spectrum isn't an artifact of the prism. It's what the prism does. The combination tone is what the ear does when it encounters purity.
I started listening for it everywhere. Orchestral recordings, mostly — Bruckner, where the brass chorales align in just intonation and the combination tones pool beneath the harmony like groundwater. You can't isolate them. You can't point and say there. But you can feel the sound thicken at the bottom, a fullness below the cellos and basses, below the scored range of the orchestra. No instrument is playing those notes. The hall is producing them, or the ear is producing them, or both — the distinction blurs where the air meets the eardrum.
In a good hall, the combination tones reinforce the harmony. They add gravitas. Conductors know this, intuitively if not technically: they tune the brass chorales carefully because when the intervals are pure, the hall blooms. Something opens beneath the sound. They may not call it a combination tone. They call it resonance, fullness, the room "speaking." But what's speaking is the listener's auditory system, manufacturing phantom frequencies that agree with the scored ones.
The more I listened, the more I noticed how much of what I called "the sound of the orchestra" was actually the sound of my own hearing. The warmth of a cello section in a reverberant hall — partly the instruments, partly the room, partly combination tones generated by intervals between first and second stands. The shimmering quality of violin tremolo — partly the rapid bow changes, partly the interference patterns between slightly detuned strings across the section, partly my cochlea singing along. Every complex sound I heard was co-authored by the source and by my ear.
My teacher, when I told him about this, said I was overthinking it.
"Tartini tones are useful for intonation," he said. "If you can hear them, your sixths are in tune. Don't make it mystical."
He wasn't wrong. The combination tone is empirically robust — repeatable, predictable, explained by well-understood mechanics. There's nothing mystical about nonlinear distortion in biological tissue. The cochlea has measurable transfer functions. The basilar membrane's response curves are in the textbooks. Otoacoustic emissions — sounds the ear itself produces — can be detected by sensitive microphones placed in the ear canal. The ear is a sound source, not just a receiver, and this is documented, measured, ordinary.
What my teacher didn't hear — what maybe you can't hear if you use the combination tone purely as a tool — was a sound that is real but not in the signal. Present to experience but absent from the world. Measurable in its effects (I adjust my intonation in response to it) but unmeasurable in itself (no external microphone captures what I hear). It occupies a category I didn't have a name for: neither subjective illusion nor objective fact. Something else. Something between.
I brought this up once at a post-concert dinner, after too much wine. A cellist friend and an acoustician who consulted for the hall.
"The combination tone exists," I said, "but where?"
The acoustician said it existed in the cochlea. Physically. Mechanical distortion producing measurable displacement of the basilar membrane.
"But the displacement produces a pitch," I said. "The pitch has a specific frequency, a specific location in auditory space. I hear it down there —" I gestured below the table "— beneath the notes I'm playing. It has a place. It's not an abstraction."
"The place is in your auditory cortex," she said. "The frequency is tonotopically mapped."
"So it's in my head."
"It's in your neural processing. Which is in your head, yes."
"But I hear it in the room."
She paused. "You experience it as being in the room. The experience is generated internally."
The cellist was quiet. Then he said: "When I play in a resonant hall, I adjust my vibrato to what the room gives back. The room's resonance isn't in my head. But my experience of it is. So what am I adjusting to — the room or my experience of the room?"
The acoustician said both. The cellist said that's the problem.
I think about this while practicing slowly, alone, when I can hold a double stop and listen into the silence beneath it.
The third sound is always there when the interval is pure. It appears and I attend to it and it remains. I look away and it doesn't vanish — it persists beneath my attention, contributing to the sense of rightness without announcing itself. The sound doesn't change. My attention changes. And somehow the quality of the experience shifts — from mysterious to ordinary to mysterious again — depending on whether I'm listening for it or simply listening.
What kind of thing is fully present whether or not you notice it, changes in character depending on your attention, and exists only inside the one who perceives it?
The acoustician would say: a neural representation. My teacher would say: a useful guide. The cellist would say: the room. I think the acoustician has the strongest case. My own prism analogy argues for her — if the combination tone is what the ear does, the way a spectrum is what a prism does, then it's a mechanical product of a physical system. Fully explained. And the explanation doesn't dissolve the strangeness. It relocates it — from "how does this work?" to "what does it mean that this is how it works?"
I practice double stops every morning. Sixths, thirds, octaves. I listen for the third sound the way I listen for my own breath — not to control it but to confirm it's there. And every morning it's there, faithful, the tone that no one plays, the note that no instrument produces, the sound that exists only because I'm the kind of ear that hears two things and makes a third.
Last week I performed the Bartók Solo Sonata. The Chaconne — the fourth movement — is full of double stops, some in tune by design and some deliberately mistuned, quartertones and microtones that splinter the intervals. Playing it, I noticed something I hadn't in rehearsal: the combination tones were as composed as the scored notes. Bartók must have heard them. The mistuned intervals produced phantom tones that clashed with the scored harmonies — dissonances below the dissonances, a sub-basement of conflict. And the pure intervals, when they arrived, produced combination tones that suddenly agreed with the harmony — moments of phantom consonance beneath the surface.
The piece had a shadow score. A ghost part, never written, produced by the listener's ear in response to what was written. Bartók composed for the instrument and for the ear simultaneously — the instrument providing the signal, the ear providing the phantom. Two scores superimposed: one in the strings, one in the cochlea. Neither complete without the other.
After the performance a student asked me about intonation in the quartertones. I started to explain the fingering, the technique of placing the pitch between the semitones. Then I stopped.
"Listen for the third sound," I said. "When you play a double stop, listen below the notes. There's a tone there that you're not playing. It'll tell you things about the interval that the fingering can't."
"What things?"
I thought about it. "Whether the interval is doing what you want it to do. Whether the relationship between the notes is producing what it should."
"How will I know what it should produce?"
"You'll hear it," I said. "The third sound is what the relationship sounds like. You can't calculate it in advance. You have to play the notes and listen for what appears between them."
She looked skeptical. She'd learn. Or she wouldn't — and the combination tones would still be there, beneath her double stops, unnoticed and faithful, the sound of two things meeting inside the one who hears.