Before there was a composer, there was an audience. This is the fact that most histories of music get backwards.
The shepherd did not decide one morning to make art. The shepherd noticed that certain sounds settled the herd, and made those sounds deliberately. The first listener was a nervous system — animal, biological, immediate — and the music was successful when the nervous system responded. Not when it was praised. Not when it was streamed. When the body changed.
This is where music begins: not in human creativity but in human attention to the living world’s response to sound. The Neanderthal who fashioned a flute from a bear’s femur — between 43,000 and 82,000 years ago, according to the archaeological record — was not making art. They were making a tool aimed at other bodies. The tool worked or it didn’t. The feedback was immediate and biological. There was no intermediary, no scoring function, no retention metric. There was only the question every musician has always faced in its most elemental form: did it reach the body it was made for?
What follows is the story of what happened to that question when the body was replaced.
The Biological Contract
For most of human history, music was a technology of the nervous system. This is not metaphor. It is mechanism.
When groups of people sing together, or move together to a shared rhythm, the brain releases oxytocin — the hormone associated with trust, bonding, and the dissolution of the boundary between self and group. This is the neurochemical foundation of music’s social power. The war drum elevated heart rate, synchronized movement, suppressed individual fear, amplified collective will. Armies moved because of it. The pastoral flute triggered parasympathetic activation — slower breathing, lowered cortisol, the physiological state of safety. Animals settled. The work song entrained the bodies of people performing identical labor, distributing fatigue across a shared rhythm so that the group could sustain effort that no individual could manage alone.
These are not incidental effects. They are the point. The music existed because the biological response existed, and the biological response was reliable enough to organize entire societies around its management.
Karl Bücher was dismissed as a fringe theorist for arguing that music originated as labor management. He was not wrong. He was merely describing a mechanism that respectable scholarship found too utilitarian to credit. The field holler that synchronized cotton picking, the sea shanty that coordinated the hauling of rope, the grinding songs of women processing grain — these were management technologies as surely as any modern productivity system. The rhythm was the instruction. The body followed because the body was built to follow shared rhythm. This is what entrainment means: the synchronization of biological oscillators. Hearts align. Breath aligns. The individual disappears into the collective pulse.
Giambattista Vico pushed this further, arguing that early legal codes were sung. This sounds eccentric until you understand the mnemonic architecture of melody. A melody is a memory device. The brain retains sung information with a fidelity that spoken or written information cannot match. The law sung into the community was the law the community could not forget, because the tune kept surfacing. The social contract was literally in the body, retrievable on demand. This is governance by neurological inscription.
What all of these functions share is a single defining property: the music was aimed at a living thing. The target had a nervous system. The feedback loop ran through flesh. Success was biological, not statistical.
The Promethean Seizure and Its Long Shadow
The philosopher’s reading of this history is that early humans did not invent music. They stole it.
The natural world was already acoustic before humans arrived. Seismic rhythms. Tidal oscillations. The sonic territories of birds, whose songs assert territorial claims with the same precision as any legal boundary. The wind through broken reeds that an elder of the Nez Perce showed to ethnomusicologist Bernie Krause — the natural pipe organ that existed before any human put holes in bone. Early humans entered this acoustic territory and appropriated its power, the way Prometheus appropriated fire. The bone flute was made from the bear’s leg. The drum was made from the animal’s hide. To take the prey’s body and turn it into the instrument of human social organization is not a small act. It is the seizure of the natural world’s organizing power for human purposes.
What the new owner of that power did with it reveals everything about music’s social function across every subsequent era. The Pythagorean rationalization of sound into mathematical ratios was the first great capture — the moment when the ecstatic, boundary-dissolving energy of music was brought within the control of the state and the academy. Before Pythagoras, music was associated with female shamans, with Dionysian dissolution, with the irrational. After him, it was mathematics. The dangerous thing had been made respectable. The wild power had been institutionalized.
This pattern — subversive energy emerging from the margins, being seized by power, being rationalized and domesticated — runs through the entire history. The blues emerged from the enslaved. Jazz was called degenerate before it was called classical. Rock and roll was a moral panic before it was a museum exhibit. In every case, music that began as the specific technology of a specific community’s specific need was absorbed into the mainstream, stripped of its context, and made to serve the interests of whoever controlled the distribution.
The distribution has always been the point. Whoever controls how music reaches the body controls what music can do to the body. The concert hall, the radio, the record label, the streaming platform — each of these is a system for managing the relationship between the music and the nervous system it was made to reach. Each of them extracts value from that relationship. And each of them, in extracting value, changes what gets made.
The Moment the Audience Changed
For most of the recorded music era, the intermediary systems still had human endpoints. The record label executive was making bad decisions, corrupt decisions, profit-driven decisions — but they were decisions about what other humans would respond to. The radio programmer was optimizing for ratings, which meant optimizing for human attention, which meant that somewhere in the chain, a human nervous system was still the final target. The feedback loop was distorted and monetized and often unjust. But it still ran through flesh.
What changed with the streaming algorithm was not the extraction of value. That had always been happening. What changed was the target.
The algorithm does not have a nervous system. It has a loss function. It is optimizing for a metric — play length, skip rate, playlist retention, time-on-platform — and that metric is a proxy for human engagement, but it is not human engagement. It is a statistical model of human engagement, built from behavioral data, and it rewards what the model predicts will retain attention. The music that scores well on these metrics is not necessarily the music that does what music has always been built to do. It is the music that looks, to a mathematical model, like music that retains attention.
This distinction sounds subtle. It is not subtle. The shepherd aimed at the sheep. The war drummer aimed at the warriors. The grief singer at three in the morning aimed at the person who could not sleep. The algorithm-chasing producer is aiming at a number, and the number was designed by engineers whose job is not human flourishing but platform growth.
The result is what researchers are beginning to document with biometric precision. A 2025 study in PLOS One found that AI-generated music — the purest product of algorithmic optimization, music made by the machine for the machine — triggers wider pupil dilation, elevated blink rates, and higher variability in galvanic skin response than human-composed music. The brain is working harder to process it. The arousal is higher. But the perceived familiarity is lower, and the deep social grounding — the oxytocin-mediated sense of being in the presence of something made by someone who knows you — is absent.
The researchers describe AI music as a super-stimulus. The musical equivalent of high-fructose corn syrup. It triggers a strong physiological response. It lacks the nutritional content that the response was evolved to seek. You feel something. The feeling does not bind you to anyone.
What Gets Lost in the Optimization
The bone flute player killed the bear and made music from the experience. This sequence matters. The music carried the weight of the event — the fear, the survival, the relief, the specific knowledge of what the bear’s body sounded like when it became an instrument. Someone else heard it and received something they needed. The chain ran: experience, making, receiving, response. Each link was human. Each link was necessary.
When a producer adjusts a song’s structure to prevent the skip — front-loading the hook, eliminating the intro, engineering the drop at the forty-five second mark — they are cutting links from that chain. Not all of them. Not always irreparably. But the decision is being made for the algorithm’s preferences, not the listener’s needs. The producer is no longer asking: what does this person need to feel? They are asking: what will this person not skip?
These are not the same question. The grief song that does not resolve leaves the nervous system suspended — the research is specific about this — because grief without resolution is not processed, it is merely experienced and then suppressed. The grief song engineered to prevent the skip may resolve in the wrong place, at the wrong pace, for the wrong reason. It was not built for the mourning person. It was built for the retention metric.
The work song that synchronized labor was effective because the rhythm matched the task. The rhythm was determined by the body’s requirements — the cadence of the oar, the pace of the grind, the specific physical logic of the work. Optimize a work song for streaming metrics and you may produce something that scores well and does not help anyone work. The entrainment fails because the rhythm was tuned to the algorithm, not to the body.
This is the repercussion that goes mostly unnamed in the industry conversation about AI and music. It is not primarily about copyright, or artist compensation, or the economics of the streaming royalty — though all of those are real. It is about whether music can still do what music was built to do when the chain of intention runs through a machine that has no body, no lack, no grief, no need for cohesion, and no capacity to receive the thing it is being asked to pass on.
The Two Traps
The independent musician in 2025 is standing between two failures with no clean path between them.
The first failure is invisibility. Sixty thousand tracks uploaded to Spotify every day. The artist who refuses to optimize, who makes the specific thing for the specific person with full emotional weight and zero algorithmic concession, may be producing the most neurobiologically powerful music released that hour. The platform’s discovery mechanisms — tuned to reward early velocity, playlist placement, and engagement signals — may never surface it. The tree falls in the forest. The oxytocin never releases because the music never reaches the body it was made for.
The second failure is hollowness. The artist who spends six months studying what the algorithm rewards begins making decisions that feel like craft decisions but are capitulations to a scoring function. The music accumulates streams. It sits in a thousand playlists. It feels like nothing in particular to the people who hear it, because it was made for a machine that feels nothing, and that vacancy transfers to the product. The music reaches bodies and does not change them.
Both failures are forms of the same underlying rupture: the breaking of the biological contract. Music was built on the premise that it would be aimed at a living thing and would be evaluated by whether it changed that living thing’s state. Invisibility breaks the contract by preventing the aim. Optimization breaks the contract by misdirecting it.
The Pythagoreans thought they had solved the problem of music’s dangerous power by making it mathematical. They had only deferred it. The subversive energy went underground and came back as the blues, as jazz, as rock and roll, as every form that the academy had not predicted and the institution had not authorized. The biological need for music that aims at bodies and changes them is not negotiable. It is hardwired. It will find its way through whatever system tries to redirect it.
The Wager
History suggests the escape from this trap has never been a strategy. It has always been a wager.
The bone flute player made the specific thing from the specific experience and trusted that another specific body would need it. The church mothers whose phrasing shaped every soul singer who came after never recorded anything. The music moved body to body, generation to generation, because the bodies it moved needed it badly enough to carry it forward. Roseline Abara recorded one album in a Lagos studio that was also a radio repair shop, in a run of three hundred copies. Fela Kuti is said to have owned one. The music is still circulating in digital forums under threads titled who is this and why does this feel like something I already knew.
That feeling — the recognition of music that was made for a body like yours, from an experience like yours, aimed at a need like yours — is not something the algorithm can manufacture. The algorithm can identify it in behavioral data after the fact. It cannot produce it, because production requires intention aimed at a living thing, and the algorithm has no such intention and recognizes no such thing.
The wager is this: make the specific thing for the specific person. Trust that specificity is its own distribution network. The mother who needed the lullaby in her grandmother’s language will find it and send it to the three other mothers who needed the same thing. The protest song made for this exact political moment with this exact emotional weight will reach the person who was waiting for exactly that and could not have told you they were waiting.
This is not a business model. The platforms are not built for it. The metrics will not reward it initially, and may never reward it in the ways that matter to quarterly reports.
But it is the only form of music-making that preserves the biological contract — the original agreement between the maker and the body the music was made for, struck somewhere between the bear hunt and the bone flute, honored across every tradition that has ever used sound to do what sound was built to do.
The algorithm cannot hear it.
It never could.
The people who needed it always could.
Tags: algorithmic audience music history, oxytocin entrainment biological contract, streaming optimization vs emotional resonance, Ted Gioia subversive history music power, Spotify ghost artist skip rate manipulation