The regression of listening, articulated in Adorno's 1938 essay 'On the Fetish-Character in Music and the Regression of Listening,' describes how the culture industry's standardized musical products restructure the listener's perceptual apparatus. The listener habituated to popular music's harmonic progressions and rhythmic patterns does not merely prefer those patterns—she loses the capacity to perceive deviations from them. The regression is not a failure of taste but a restructuring of perception itself, accomplished through repetition, through the relentless provision of stimuli calibrated to satisfy without challenging. The ear is trained to expect smoothness, and the expectation of smoothness destroys the capacity to hear what smoothness conceals: silence, dissonance, rhythmic asymmetry, the formal resistance genuine art deploys to crack expectation and force encounter.
Adorno's analysis was rooted in his training as a composer and his lifelong engagement with the Second Viennese School—Schoenberg, Berg, Webern—whose atonal and twelve-tone compositions resisted the harmonic expectations that tonal music had naturalized over centuries. The resistance was not arbitrary difficulty but a formal strategy: by refusing to resolve harmonic tension in expected ways, the music forced the listener into a different mode of attention. It demanded active engagement rather than passive reception. The culture industry's products reversed this demand: they resolved tension predictably, provided harmonic closure on schedule, and trained the listener to consume music in the manner one consumes a commodity—expecting satisfaction, receiving it, moving on.
The concept extends beyond music into every domain where AI generates cultural content. AI-generated prose is fluent, contextually appropriate, tonally calibrated—delivering exactly what statistical patterns predict a reader expects. The reader habituated to this fluency experiences it as quality. She does not perceive the absence of roughness, because roughness—the sentence that resists easy assimilation, the paragraph that forces rereading, the formal difficulty that signals genuine thought—has been progressively eliminated from her reading environment. Her capacity to perceive and value difficulty atrophies through disuse. Edo Segal's confession that he almost kept an AI-generated passage because it 'sounded like insight' exemplifies the danger: the smooth surface seduces precisely because the reader's ear, trained on smooth surfaces, has lost the capacity to hear what smoothness conceals.
The regression operates cumulatively and irreversibly. Each smooth encounter reinforces the expectation of smoothness. Each reinforcement weakens the tolerance for roughness. The weakening proceeds below conscious awareness—the listener does not notice the degradation, because the faculty being degraded is the faculty that would perceive the degradation. A generation raised on AI-generated content will not experience smoothness as limitation, because the ear that could detect the limitation has never been developed. The atrophy of critical perception is invisible from inside the atrophy.
The concept emerged from Adorno's analysis of 1930s popular music—jazz, swing, the products of Tin Pan Alley. His dismissal of jazz was controversial and reflected his European art-music bias, but the mechanism he identified—standardization training perception—has proven more durable than his specific musical judgments. The 1938 essay was written in Oxford, during his first exile from Nazi Germany, and it laid the groundwork for the culture industry critique he would develop with Horkheimer six years later in Los Angeles.
Perceptual restructuring. Standardized cultural products do not merely fail to challenge perception—they actively degrade it, training the ear/eye/mind to expect patterns that preclude surprise.
Invisibility of degradation. The regression operates below awareness—the faculty being degraded is the faculty that would perceive degradation, making the loss self-concealing.
Cumulative and irreversible. Each smooth encounter reinforces smooth expectations; capacity for roughness atrophies through disuse and cannot be restored by simple re-exposure to difficulty.
AI-generated smoothness. Language models optimized on statistical patterns produce outputs of such consistent fluency that they train readers to mistake fluency for depth, adequacy for excellence.
Generational transmission. Children raised on AI-generated content will inherit ears/eyes structurally incapable of perceiving what smoothness conceals—a loss that will not register as loss because the baseline has shifted.