Why does one piece of music catch us while another leaves us cold? The answer lies not in class, education, or exposure alone, but in recursive resonance: the moment a sedimented anticipatory field meets a sufficiently constrained deviation that arrives already pleasurable. A seasoned jazz listener hears a single substituted chord as a small shock. Two children on their hundredth Magic Flute hearing feel the same aria arrive with undiminished charge. Familiarity does not dull surprise, it makes sophisticated surprise possible, carving the narrow channel within which a deviation can be felt as exquisite rather than chaotic. Yet resonance carries no guarantee of persistence. Taste is stratigraphy: layers accumulating without replacing one another. Some layers remain active, others become dormant. The year my children inhabited Mozart’s opera ended not because the music changed, but because they grew up a little bit more and moved on. Music catches us, holds us, and, just as mysteriously, lets us go.
I. The Year of the Magic Flute
When my children were four and six, they converged, for no obvious reason, on Mozart's The Magic Flute. This was not background music that happened to be playing while they were in the room. They wanted it. They asked for it. They listened to it repeatedly, sang the arias to each other, performed scenes with props improvised from whatever was lying around the house, and watched several different filmed productions closely enough to tell the Salzburg Festival staging apart from the Vienna production and from the Royal Opera House production in London, arguing over which Queen of the Night or which Papageno they preferred. Once, when I had my hair cut, both children sat through the entire appointment perfectly absorbed, watching the opera on a tablet propped against the mirror, apparently oblivious to the barber, the clippings, or the wait.
After something like nine months to a year, it stopped. Nothing dramatic caused it. Neither child repudiated Mozart or announced they were tired of him. The shared attachment simply thinned out and then was gone. It has not returned. Neither child went on to develop any particular attachment to opera or to classical music more broadly, and neither has ever again converged with the other so completely on a single extended piece of music.
The question this raises is not yet why do I like what I like, but something narrower and stranger: what happened during that year?
The obvious explanations fail almost as soon as they are stated. It cannot have been simple inherited class taste, since the taste did not persist beyond the year in which it appeared. It cannot have been straightforward parental transmission, since both children subsequently diverged from whatever musical taste their parents actually hold. It cannot be explained by the sheer greatness of the piece, since plenty of other equally canonical works were available in the same house, played by the same parents, without producing anything close to the same effect. It cannot be explained by exposure alone, since the heaviest exposure came after the attachment had already begun rather than before it, the children asking for more rather than being given more and gradually acquiring a taste for it. And it cannot be reduced to two instances of ordinary individual taste occurring side by side, since the striking feature of the whole episode was that the two children converged on the same opera at the same time and then visibly intensified each other's attachment to it by singing, performing, and comparing productions together.
Musical tastes are patterned by social position, and yet actual attachments of this kind are astonishingly singular. Taste layers rather than replacing one preference with the next. People occupying very similar class positions diverge from each other in what catches them. The same person can love Bach, bluegrass, Indonesian gamelan, and contemporary pop without any of these attachments cancelling the others. All of this is true and important, and the rest of this article develops it at length. None of it, on its own, explains a single year in one household in which an eighteenth-century opera became the centre of two small children's world and then, just as mysteriously, left it again.
II. A Simple Question
I like Bach. I also like Beethoven's late string quartets, American bluegrass, gamelan, contemporary pop of the kind Olivia Rodrigo makes, and, on certain evenings, Schoenberg. Discovering a song by Olivia Rodrigo did nothing to diminish my liking of Bach. Finding gamelan did not dislodge bluegrass, and returning to Schoenberg after years away did not make Beethoven sound thin by comparison. Musical taste does not behave like a ranking of substitutable preferences, in which a new favourite quietly displaces an old one the way a new pair of shoes displaces the pair it replaces, but like a set of layers, each retaining whatever pull it had at the moment it first mattered, available to be returned to on its own terms.
A harder puzzle appears once other people enter the picture. Two people with almost identical class position, education, profession, and cultural capital can hold strikingly different musical tastes. One adores opera and the other cannot sit through an act of it. One finds Bach transcendent and the other finds him arid and mechanical. Sociology has become very good at explaining why populations display patterned distributions of taste across class, education, and region. It has been far less successful at explaining why this piece of music catches this person while an almost identical piece, offered under almost identical conditions, leaves them cold. A distribution is not an encounter, and no amount of averaging across a population will tell you what happened to one particular listener on one particular evening, let alone what happened to two children over one particular year.
This article takes music as the hardest possible case for a general theory of taste, because music strips away much of the symbolic scaffolding through which sociology usually explains why people like what they like: meaning, representation, status, and explicit moral judgement. If a theory of taste can say something true about why a piece of music catches a listener, and about why it later lets go of them, it will have said something about a much larger territory than music alone.
III. Bourdieu and the Limits of Distinction
Pierre Bourdieu's account of taste in Distinction shows that taste is patterned by class, that cultural capital is unequally distributed across households, and that exposure to what a society counts as legitimate culture is concentrated among those already positioned to receive it and convert it into further advantage. Judgements that present themselves as spontaneous and personal, my ear simply prefers this composer, have social histories that a purely aesthetic account conceals from the person making the judgement, who typically experiences their own taste as an unmediated fact about themselves rather than as the outcome of an unequal distribution of opportunity.
The claim that people occupying a given social position are disproportionately likely to acquire a particular taste is a claim about populations, well supported by decades of survey evidence. It becomes a different claim when social position is asked to explain why this individual possesses this taste, and that second claim is not supported by the same evidence. Imagine ten upper-middle-class academics of similar age, education, discipline, and cultural capital, sharing a common room after a seminar. Their musical tastes will not be identical, and may be spectacularly different from one another, ranging across opera, heavy metal, none of the above, and everything in between. A survey researcher can note that this group as a whole listens to more classical music and attends more concerts than a comparable group of manual labourers, and that observation will very likely be true. It will not tell you which of the ten academics actually loves Mahler, which merely tolerates him as a professional obligation attached to a certain kind of dinner party, and which cannot stand him despite having attended dozens of performances out of politeness.
Habitus was never meant to function as a deterministic law predicting that every upper-middle-class academic must like Mahler. It is a theory of probabilistic disposition, and the criticism that matters here is not that it makes a false prediction about any one person, since it never claimed to make one, but that it operates at a coarse grain. Habitus explains why certain kinds of musical encounter are more probable for people in a given position, and why certain dispositions cluster socially in the aggregate. It gives a coarse-grained account of how one particular history has made one particular listener capable of being caught by one particular passage on one particular evening, which is where taste is actually located.
The Magic Flute case makes this limit unusually visible. The children's social position helps explain why an extended eighteenth-century opera was readily available to a four- and a six-year-old in multiple high-quality recorded productions, and why a household existed that was willing to indulge nine months of repeated listening, singing, and improvised staging. None of that explains why this particular opera caught both children at precisely this moment rather than at another, why it caught them together rather than separately, why the attachment intensified as sharply as it did, or why it disappeared roughly a year later without leaving behind any lasting taste for opera or classical music in either child. Class position altered the probability that Mozart would be available. It did not, and could not, determine the encounter.
IV. Prior, Hennion, and Lahire
Nick Prior's sociology of music restores technology, format, and situated listening practice to a Bourdieusian frame. His work on musical mediation, domestic listening practices, and the material culture of genre shows that taste is never a disembodied preference floating free of turntables, streaming interfaces, acoustic spaces, headphones, and social occasions. A love of vinyl crackle, a preference for a particular playback chain, or an attachment to the ritual of putting on a record are part of taste's actual substance rather than decorative footnotes to it. Music after Prior is understood as materially and technologically mediated at every point, including the tablet propped against a mirror in a hairdresser's shop.
What this more materialist sociology still leaves untouched is the event of musical appreciation itself. Neither a class position nor a particular playback technology explains what happens during the three or four seconds in which a bridge arrives and feels exactly right, or during the months in which a four-year-old's attention becomes fixed on one specific opera among dozens available to her. That moment, and that year, are the things a theory of taste is supposed to explain.
Antoine Hennion's account of taste as performance treats taste as neither an effect of external social determination nor the passive discovery of a pre-existing aesthetic property in the object. Hennion studies the gestures, devices, media, practices, and deliberately uncertain outcomes through which listeners actively make themselves capable of being affected by music. A listener who cues up a record, adjusts the volume, closes their eyes, and waits to see whether it takes hold this time is performing a practice whose result is truly in doubt, which is close to the acquisition-ecology argument developed later in this article, and close to what the two children were doing when they staged their own productions of scenes they already knew by heart. Hennion's framework describes the rehearsal in considerable detail. It has less to say about the decisive moment itself, the point at which a performance succeeds or fails to take hold, which is where a recursive vocabulary can add something: naming the sedimentation of an anticipatory field, which is what most of Hennion's practices are actually building, and distinguishing it from the felt misalignment that occurs, or fails to occur, once the field is in place.
Bernard Lahire has argued at length against sociologies that caricature individuals as coherent embodiments of a single class position, showing empirically that heterogeneous, internally dissonant cultural profiles are the norm. Someone can hold a doctorate and love reality television. Someone can work a manual trade and be a serious devotee of opera. That a person can simultaneously love Beethoven, bluegrass, and contemporary pop is a fact Lahire has already documented in considerable empirical detail across many individuals. What remains open is an account of how such heterogeneity becomes recursively sedimented within one singular life: not merely that dissonance exists across a population, but how one person accumulates several independently sedimented anticipatory fields side by side, each with its own acquisition history, in the manner developed below under the heading of taste as stratigraphy.
V. DeNora, Benzecry, and the Opera Case
Tia DeNora's account of music's place in everyday, embodied self-regulation treats music as a technology for managing energy, mood, and the pacing of ongoing activity, entraining a listener's body and attention to a sustained rhythmic or textural field rather than delivering discrete events to be registered one after another. This vocabulary returns below, once the argument reaches forms of musical pleasure that plainly do not depend on surprise.
Claudio Benzecry's ethnography of opera obsessives treats their devotion as a matter of passion and cultivated attachment rather than status display, and his subjects are frequently people whose love of opera arrived well after childhood and grew into something closer to vocation than to inherited habit. Benzecry's material reads as a record of late-acquired recursive resonance, built through years of self-directed listening, repeated attendance, and increasingly demanding engagement with an idiom that gave nothing away easily at first. The children's case sits at the opposite temporal extreme from Benzecry's adult converts and is worth holding alongside it. An adult coming to opera late typically needs a deliberate scaffold that a young child rarely does: repeated listening with a libretto in hand, a first live performance chosen for its accessibility, a friend already fluent in the idiom who can point out what to listen for on a second or third hearing. The four- and six-year-old needed none of this scaffolding to become extraordinarily discriminating listeners within months. What they needed instead was repetition, company, and permission to stage the thing themselves, which raises the question addressed directly in the next section: how much explicit musical understanding does recursive resonance actually require.
The article's contribution alongside these four is not the discovery that taste involves more than social position, since Hennion and Lahire have already shown this, or that music is bound up with embodied self-regulation and passionate devotion, since DeNora and Benzecry have already shown this too. The contribution is a diagnosis of what kind of recursive process expectation, deviation, and attachment actually are, which lets these separate findings be connected to one another rather than accumulated side by side.
VI. Music Below the Threshold of Language
Composition, notation, performance institutions, genre classification, and music criticism plainly involve symbolic articulation of a high order. A conductor reads a score, a critic writes a review, and a conservatory teaches theory using vocabulary built up over centuries. None of this should be denied. What can be denied is that the musical payoff itself, the moment at which a piece of music actually catches a listener, requires any of that symbolic apparatus to occur.
A four-year-old can experience it. This is one of the reasons the Magic Flute case matters so much for the argument. The children acquired an extremely differentiated familiarity with particular productions, particular arias, particular characters, and particular performers, fine enough to prefer one Queen of the Night's coloratura over another's, without anything resembling a musicological account of why any of it worked. Their musical discrimination ran well ahead of their capacity to articulate it. A cadence need not denote anything, and a child can find a passage magnificent without knowing what a key change is, just as an adult can be moved by a harmonic shift without possessing the vocabulary to name it. Music can carry lyrics, titles, and socially coded genre labels that invite symbolic reading in the way literature does, but organised sound does not ordinarily work aesthetically because it stands for something else the way a word or an image does. Strip away the plot of the opera, the German libretto neither child yet spoke, and the eighteenth-century conventions surrounding it, and the music could still catch, or fail to catch, a four-year-old entirely on the strength of what it did to her within her own anticipatory field.
The claim that follows is therefore narrower than an earlier and more sweeping version of it. Music is not simply an art belonging to the first two levels of recursivity, since composing, performing, classifying, and criticising it plainly reach well beyond that. The musical payoff itself does not require symbolic articulation to occur, and whatever taste turns out to be in general, in music its decisive event will have to be explained mostly below the level of language.
VII. The Anticipatory Field, at Three Scales
Listening depends on sedimentation built up through exposure. Rhythm, metre, tonal relation, repetition, melodic contour, timbre, and genre convention become ground for an experienced listener rather than propositions continually and consciously held in mind while a piece plays. Exposure does not simply supply information about music the way a textbook supplies information about a historical period; it changes what a listener can anticipate without being able to articulate the anticipation at all, so that a seasoned listener of bluegrass can feel a fiddle break coming half a bar before it arrives without being able to say, in words, what told them it was coming.
This anticipatory field operates at three distinct temporal scales, nested inside one another whenever a piece of music is heard. The deepest scale consists of very general embodied and developmental regularities, acquired early and common across enormous numbers of listeners: an expectation that a steady pulse will continue once established, that a rising melodic line carries tension and a falling one carries release, that a held dissonance wants resolution, that the human voice will organise a listener's attention around it wherever it appears. The middle scale consists of biographically and culturally sedimented genre expectations, ordinarily thought of as accumulating over years of exposure to a particular musical world. The third and shallowest scale is piece-local: a piece can establish its own small, temporary set of rules within seconds of starting and then exploit precisely those rules a few bars or minutes later, entirely independently of whether the listener has ever heard anything else in the genre before.
The Magic Flute case shows that the middle scale can compress dramatically under the right conditions. Two children with almost no prior exposure to opera built, within months rather than years, a genre-specific field differentiated enough to distinguish productions, recognise favourite arias on first bars, and anticipate particular staging choices before they arrived. What would ordinarily be described as slow biographical sedimentation, requiring years of dispersed exposure, can evidently proceed far more quickly when exposure is concentrated, repeated daily, socially reinforced by a sibling doing the same thing, and voluntarily sought rather than passively received. The boundary between the middle and shallow scales turns out to be a matter of intensity and repetition rather than a fixed calendar duration, which matters considerably once the discussion turns to duration and repeated listening below.
VIII. Constructed Misalignment
Music's involvement with expectation is not a new discovery. Leonard Meyer's account of musical emotion made expectation, its confirmation, its delay, and its eventual resolution or denial central to how a listener experiences a piece. David Huron's later work develops expectation into a fully worked empirical framework for musical cognition, showing how listeners' predictions about what comes next are formed, violated, and confirmed, with measurable emotional consequences. What follows draws directly on this line of work while trying to locate it within a general account of what a living being's recursive coordination with its world consists in.
Felt disturbance at the second level of recursivity is usually treated as something that interrupts smooth coordination and calls for repair, a discrepancy a living being would rather not have and would prefer to resolve quickly. Music shows that this disturbance need not be experienced negatively at all, and that it can be deliberately manufactured rather than merely tolerated when it occurs. A delayed cadence, an unexpected chord, a rhythmic displacement, an altered melodic line, an extended phrase, a sudden silence, a deceptive return, or an astonishing harmonic turn all violate something the listener's ear has already sedimented as expected, and the violation is precisely what the listener came for.
Not every violation works, and the limit case makes the point with unusual clarity. A car horn sounding outside the concert hall during a Beethoven quartet is surprising, and it may register a considerable disturbance in the listener's anticipatory field, larger by any objective measure than the disturbance produced by the actual musical passage playing at that moment. It produces no comparable pleasure, and something closer to irritation instead. Musical pleasure cannot be explained by the sheer magnitude of surprise on its own. Something narrower is doing the real work, and naming it precisely is the task of the next two sections.
IX. Constrained Surprise and Mixed Futurity
The space of sounds that could physically follow any given musical moment is effectively unbounded once every possible instrument, pitch, and duration is counted. The space of continuations that will actually feel musically satisfying, given everything the passage has already established, is extraordinarily narrow by comparison. The best musical continuation occupies a peculiar position within that narrow space: it is not fully predictable in advance, and yet, once it has occurred, it can feel retrospectively inevitable, as though no other note could have followed. I did not know that particular chord was coming, and now that it has arrived, nothing else could have been right in its place.
This is neither pure determinacy, in which the next note follows mechanically from a rule, nor pure randomness, in which any note would do equally well. It is mixed futurity made directly audible. The passage creates real constraints on what can plausibly happen next without fixing what will happen, and the actual continuation, once it sounds, realises one possibility from within those constraints in a way capable of reorganising, in retrospect, how everything that came before it is heard. This is constrained surprise: unpredictability held tightly inside a narrow and already sedimented channel that the listener's own layered history has carved out in advance.
Stated this way, the account describes a first hearing well. It says nothing yet about the fiftieth, and a listener who already knows a piece by heart is not a marginal case for a theory of musical pleasure but closer to the ordinary condition under which most of the music anyone actually loves gets heard.
X. The Metabolic Window and the Strange Temporality of Re-Listening
Music is usually short. A novel can comfortably demand twenty or forty hours to read. A film runs to about two hours as a matter of convention. A symphony approaching an hour in length is already internally divided into movements, each with its own beginning and close. Opera, whatever its overall length, is built from acts, scenes, arias, recitative, and intervals rather than presented as one continuous unbroken span. Even extended improvisational forms such as jazz develop episodes, returns, and internal segmentation rather than proceeding as a single undifferentiated stretch. Popular songs have historically clustered around remarkably short durations, and the difference between a three-minute song and a six- or eight-minute one is culturally conspicuous, despite being a trivial difference set against the length of a novel.
None of this need be turned into a hard biological law, but it does suggest a metabolic-window hypothesis. Music's characteristic operation at the first two levels of recursivity requires recent musical events to remain sufficiently active in a listener's attention, working memory, and arousal for later events to be experienced against them as ground. Temporal integration of this kind has limits. A continuously evolving anticipatory field can only remain coherent for so long before it needs internal resetting or segmentation, and long musical works accordingly have architecture, movements, acts, verses and choruses, returns and reprises, precisely because unbroken duration creates a recursive problem that segmentation solves.
Music compensates for its typically short individual duration by being extraordinarily repeatable, in a way a novel or a film is not, and this is where the comparison with literature becomes useful. A long novel acquires its temporal magnitude by being long. A pop song, or an opera, acquires comparable temporal magnitude by being heard hundreds of times across a life rather than once at great length. A piece lasting four minutes can inhabit a life for fifty years, and an opera lasting three hours can occupy nine months of a four-year-old's imagination through dozens of repeated hearings rather than through any single extended sitting.
Repeated listening of this kind is not the repetition of an identical recursive event. The recording itself, the nonrecursive trace, remains comparatively stable from one playing to the next, while the recursive listener who returns to it does not. Every hearing inherits everything the listener brings from every previous hearing, including memories of those previous hearings themselves.
This is where mixed futurity becomes considerably more interesting than a first-hearing account can capture. On the fiftieth hearing of a piece, informational uncertainty about what happens next is essentially zero. The listener knows the bridge is coming, knows exactly which chord will arrive, and could sing the melody a bar ahead of the recording without effort. Anticipation nonetheless survives intact. The bridge still arrives with a charge. The cadence still satisfies. The orchestral climax still produces goosebumps on the fiftieth hearing much as it did on the first, sometimes more intensely rather than less. Symbolic foreknowledge, in other words, does not abolish recursive futurity. Knowing what is going to happen is not equivalent to living through its arrival. A rider on a roller coaster who has ridden it fifty times knows exactly where the drop occurs, and this knowledge does nothing to stop the stomach from lurching at that exact point regardless. Knowing precisely when a musical bridge arrives does not make the bridge experientially present before it actually sounds.
This requires a refinement of constrained surprise itself. On a first hearing, recursive resonance can involve real novelty, a chord not actually anticipated at the level of conscious prediction. On repeated hearings it cannot, since nothing about the passage is any longer unknown in that sense. The deeper category, of which first-hearing surprise turns out to be only one instance, is constrained felt misalignment within an ongoing anticipatory process. Surprise is one possible phenomenology of that misalignment. Anticipation, delay, held tension, and eventual release or arrival are others, and these can survive complete propositional foreknowledge entirely intact, sometimes growing sharper with familiarity rather than duller, as the listener's field around the passage becomes more finely differentiated with each return to it.
XI. Recursive Resonance
The concept this whole argument needs is recursive resonance, and it has to be defined precisely, since a looser formulation risks placing it one step too late in the sequence of events it is meant to capture. Recursive resonance is not the uptake of a felt misalignment that happened first and was then responded to afterwards. Recursive resonance is a positively valenced, sufficiently constrained felt misalignment, generated within a sedimented anticipatory field. What follows resonance, wanting to hear the passage again, seeking out more of the same composer, telling someone else about it, is a further recoordination resonance makes possible, but it is not part of what resonance itself names.
Whether the listener knew, propositionally, that the event was coming is secondary to this definition rather than central to it, which is the correction the previous section makes necessary. A four-year-old on her hundredth hearing of an aria she can sing from memory, and a first-time listener hearing a Beethoven quartet with no idea what comes next, can both be undergoing recursive resonance, so long as each is experiencing a positively valenced, tightly constrained felt misalignment against a field sedimented enough to make that misalignment precise rather than diffuse. Informational surprise is one route into that experience. It is not the experience itself.
A car horn supplies felt misalignment that never becomes positively valenced, producing surprise without resonance. A wholly predictable cadence supplies smooth coordination with too little misalignment present for resonance to arise at all, whatever milder pleasure it might still offer. An irritating or badly executed deviation supplies felt misalignment that remains negatively valenced throughout, registering as something having gone wrong rather than something exquisite having happened, and is simply a different, unpleasant species of felt misalignment rather than a resonance that failed. Competence in reading, analysing, or performing a piece of music is not the same thing as resonating with it, a distinction developed further below. The extraordinary musical moment, whether on a first hearing or a fiftieth, is the coincidence of a sedimented anticipatory field with a sufficiently constrained deviation from it that arrives already positively valenced.
XII. Absorption Without Misalignment
Not all musical pleasure depends on felt misalignment. Drone music, extended chant, minimalist compositions built from long repeating cells, much techno and house music, the lullaby, and the simple satisfaction of a bass line locking absolutely into a drum pattern all seem to work through the opposite mechanism: continued alignment, entrainment, repetition, gradual intensification, or a deepening absorption into a coordination that is not going anywhere in particular and does not need to. A listener lost in an hour of slowly evolving drone is not waiting for a bridge to arrive and resolve something. Nothing in the piece is being set up in order to be exquisitely disturbed.
DeNora's account of music's place in embodied self-regulation supplies the right vocabulary here. Music functions, on this account, as a technology for managing energy, mood, and the pacing of ongoing activity, entraining a listener's body and attention to a sustained field rather than delivering discrete events to be registered one after another. A jogger's playlist, a lullaby, a congregation's chant, and a dance floor's groove are doing something closer to this than to the construction and release of constrained surprise, and the pleasure involved is real, often intense, and plainly musical.
Music should be understood, on this basis, as having an extraordinary capacity to construct rich fields at the first level of recursivity, some of which are sought out and enjoyed for their own sake as sustained absorption without any requirement that they be disturbed at all. Within a great deal of music, though not all of it, some of the most powerful moments then arise through positively valenced felt misalignment generated inside that absorbed field, recursive resonance proper. Music has an unusual capacity to build the kind of field within which felt misalignment of this constrained, positively valenced kind becomes possible, and exercising that capacity is one of music's supreme powers rather than the whole of what music is for. A single piece, a long dance track that grooves for eight minutes before a dramatic break, can use absorption to build exactly the field a later disturbance will need.
XIII. The Paradox of Familiarity
Familiarity and surprise are not opposites trading off against each other. Familiarity is what makes sophisticated surprise possible in the first place. Repeated exposure to a genre creates an increasingly differentiated anticipatory field, and a listener equipped with such a field can register deviations that remain invisible to a newcomer, for whom an unfamiliar idiom often sounds uniformly strange rather than precisely surprising in any one identifiable place. Musical education, understood broadly, teaches a listener how to be surprised, rather than simply teaching facts about music. An expert can experience more astonishment from a given passage than a novice does, not less, because the expert's field has been differentiated finely enough over years of listening to register a deviation the novice cannot yet hear as a deviation at all, only as more of the same unplaceable strangeness.
A jazz listener who has spent years with standard chord progressions can hear a single substituted chord in a familiar tune as a small shock precisely because the substitution stands out against an extremely well-sedimented background. Someone hearing that same tune for the first time has no such background against which the substitution could register, and hears only an unfamiliar harmony among many other unfamiliar harmonies, none more surprising than the rest. The listener who already knows a genre best is often positioned to be surprised by it most acutely, while the complete newcomer finds almost nothing astonishing in any locatable sense, because astonishment requires a background precise enough to be violated in one place rather than everywhere at once.
XIV. Schoenberg and the Collision of Fields
Schoenberg supplies an extreme test of the argument, and it is worth pausing on him because his music is often used as the textbook example of taste running out. The naive complaint against atonal music is that an untrained listener's ear simply lacks a sedimented set of expectations for it to violate. This has the situation almost exactly backwards.
The untrained listener approaching Schoenberg for the first time is not arriving with an empty anticipatory field, but typically with an extremely well-sedimented one, built from a lifetime of pop songs, hymns, film scores, and folk melodies, almost all of it organised around tonal resolution, functional harmony, and cadence. That field is not absent, only organised around the wrong axis for the music now in front of it. Schoenberg's music does not fail to offer sufficient structure for expectation to attach to. It declines to satisfy the very specific tonal expectations the listener already has fully sedimented, while offering, in their place, an alternative organising logic, built from motivic development, register, and systematic pitch relations independent of tonal function, that the listener has not yet had any opportunity to sediment at all. The result is a collision between a strongly sedimented old field and a weakly sedimented new one, in which almost every gesture registers as a violation of the old field without yet being interpretable as a precise event within any new one. Everything sounds equally wrong, and because nothing is locatably wrong in a particular place against a particular expectation, nothing achieves the sharp quality that produces resonance.
Repeated exposure can change this, not by erasing the old tonal field, which remains available whenever tonal music is next heard, but by allowing a second, differently organised field to sediment alongside it, sensitive to the kinds of regularity atonal music actually offers. Once that second field is in place, a listener can register Schoenberg's own deviations as precise events rather than as undifferentiated strangeness, and the same passage that once sounded merely disordered can begin to produce recursive resonance in its own right.
Difficult music carries a high acquisition cost for recursive resonance, a fact about the labour required to sediment an appropriate field rather than an aesthetic ranking of one style above another. Some music deliberately exploits anticipatory structures that are already very widely shared and can therefore generate its own constrained surprise almost immediately, on first hearing, and this makes such music neither better nor worse than music that demands years of patient listening before its deviations become audible as deviations at all.
XV. The Acquisition Ecology
Class matters because access to specific kinds of acquisition ecology is radically unequal, and what exactly is unequally distributed matters, since a careless version of this claim collapses into the idea that wealthier households simply contain more music. They do not. A working-class household or community can offer an immensely rich musical ecology built through churches and their choirs, social clubs, family and neighbourhood performance, dancing, local scenes built around particular venues or sound systems, radio, community music-making, oral transmission of songs and styles across generations, and obsessive, self-directed listening that costs nothing but attention and time. Gospel choirs, bluegrass jam sessions, and sound-system culture all constitute acquisition ecologies of considerable depth, capable of sedimenting recursive resonance every bit as thoroughly as a conservatory does for the repertoire it teaches. What is unequally distributed is access to particular acquisition ecologies, especially those built around expensive, institutionally credentialed forms such as private instrumental lessons, subscription concert attendance, and formal music theory instruction, which happen to attach to a narrow canon a given society has already agreed to call legitimate.
Class structures the kinds, durations, technologies, and institutional scaffolds through which musical exposure happens, and correspondingly structures which forms of accumulated resonance a society is prepared to recognise, reward, and treat as evidence of taste. A household with instruments, paid lessons, a decent playback system, an acoustically workable room, uninterrupted time, parents who listen attentively, siblings who play, regular concert attendance, and knowledgeable interlocutors offers one kind of acquisition ecology, well matched to acquiring resonance with a canonical Western instrumental repertoire. A household organised around church attendance, communal singing, and a family tradition of playing by ear offers a different kind of acquisition ecology, at least as demanding and at least as capable of producing resonance, matched to a different repertoire.
Consider two children given, on paper, the same opportunity: both enrolled in the same after-school orchestra programme at the same age. One goes home to a house where a parent puts on recordings in the evening unasked and discusses a phrase that caught their ear over dinner. The other goes home to a house where the instrument sits in its case between lessons and is not mentioned again until the next one. The programme looks identical for both children on any survey. The ecology surrounding it is not, and it is the ecology, far more than the programme, that determines whether either child ends up with anything resembling recursive resonance for orchestral music twenty years later. There is no fixed threshold of exposure after which a given repertoire automatically becomes pleasurable for anyone who reaches it. Recursive resonance can be temporally and materially expensive to acquire for a given repertoire, sometimes prohibitively so, and that expense is where cultural capital does its most important work.
One mechanism through which embodied cultural capital is acquired is unequal access to repeated and scaffolded opportunities for specific forms of recursive resonance, tied to specific repertoires a given society has already decided to treat as prestigious. That formulation preserves Bourdieu's original insight about unequal access and unequal reward while adding an account of the recursive mechanism, sedimentation of a field, occasions for constrained deviation, positive valence, through which that unequal access does its work inside one particular listener. The Magic Flute case sits slightly outside this picture, since the acquisition ecology that produced it was neither an institutionally prestigious one, since neither child received formal instruction, nor a deliberately scaffolded one, since the children largely built it themselves through repetition and play. What the household supplied was access, time, high-quality recordings, and permission. What the children supplied, and what no household could have engineered on its own, was the convergence itself.
XVI. Why Capacity Is Not Resonance
Years of paid piano or violin lessons regularly produce adults who actively dislike classical music and rarely if ever choose to listen to it once the lessons have ended. This forces a distinction between two things that are easy to conflate. Recursive capacity concerns which coordinations have become available to a listener: what they can technically recognise, anticipate, sight-read, or perform under instruction. Recursive resonance concerns whether that same listener, encountering a passage under no instruction at all, experiences a positively valenced, constrained felt misalignment within it.
A conservatory-trained pianist can recognise a sonata's formal structure, anticipate its harmonic turns several bars in advance, and perform it expertly at a recital, all without experiencing anything resembling resonance when the same piece is played to them afterwards for pleasure. An untrained listener with almost no vocabulary for harmony can experience overwhelming resonance with that same piece the very first time they hear it, purely by chance, in a film or a passing car. Such a listener is never hearing the piece against nothing, as the account of nested anticipatory fields above already showed: the deepest, most widely shared layer of anticipation, together with whatever piece-local scaffolding the recording supplies in its opening bars, can be enough to generate real resonance even in the complete absence of a trained pianist's finely differentiated field. A four-year-old with no musical training whatsoever can be caught by Mozart precisely along these lines. Competence is not taste, and a great deal of music education proceeds as though the two were the same thing, measuring success by the first while hoping the second will follow automatically from it.
XVII. Taste as Stratigraphy, Now Dynamic
New musical attachments typically do not substitute for earlier ones. They accumulate on top of what is already there, much as sediment accumulates in layers rather than replacing the layer beneath it. Discovering Olivia Rodrigo did not require me to lose Bach, and finding gamelan did not require abandoning bluegrass. These layers are not equivalent, interchangeable items sitting side by side on a single playlist, ready to be shuffled at random. Each layer carries its own acquisition history, and music loved intensely at seventeen can remain fully available at sixty, sometimes acquiring new resonances precisely because the person returning to it has changed while the recording has not.
This picture needs one further dimension. Layers differ not only in content and acquisition history but in their current recursive relevance. Some remain actively sought, returned to again and again. Some become dormant, still fully sedimented but no longer generating any active demand for renewed encounter. Some are repeatedly reopened after long absences. Some may never be revisited at all. The Magic Flute layer in my children's biographies is of the dormant kind: intensely sedimented over the course of a year, evidently still present in some form, since a great deal of what was learned that quickly rarely disappears without trace, but no longer actively demanding the repeated encounter it once did. The model of taste as stratigraphy therefore needs to be read as dynamic rather than simply accumulative. Layers are laid down, and some of them continue to be visited, while others sink and stop calling for attention, without either process erasing what is underneath.
XVIII. A Mesocosmic History Is Singular
Consider two listeners occupying virtually identical sociological positions on every conventional measure. One spent formative years living in Indonesia and the other did not. One encountered bluegrass through particular people and places in the American South, perhaps a relative's porch on specific summer evenings, and the other never had any comparable encounter. One heard opera constantly at home from infancy and the other only as an adult, if at all. These are not minor biographical deviations sitting on top of an otherwise sufficient sociological explanation, footnotes a sufficiently detailed regression could eventually absorb, but constitutive of the very listeners who go on to encounter music in the first place, shaping what each of them will even be capable of hearing as significant.
A class position, however precisely specified, is generic, in the sense that many people can and do occupy the same one. A mesocosmic history, the specific accumulated sequence of encounters through which a person's anticipatory fields were actually built, is singular to that one person alone. This gives a non-mysterious account of musical individuality that requires no appeal to an inexplicable psychological essence. Individuality is the irreducible historical specificity of accumulated recursive mediation, and taste is the point at which broad social patterning meets equally real biographical irreducibility.
XIX. When Recursive Resonance Becomes Interrecursive
Everything so far has treated musical resonance as occurring between a recursive listener and a nonrecursive musical form. The Magic Flute case will not sit still inside that picture, because the two children were not simply two individuals who happened, independently, to like Mozart at the same time. They listened together, sang together, performed scenes to one another, and argued about productions together, so that one child's excitement became part of the other's musical environment, one began singing an aria and the other joined in, and one anticipated a favourite scene while the other recognised exactly what was being anticipated and joined the anticipation before the scene arrived.
The ontology here needs to be stated carefully. Mozart's recorded music is not itself interrecursive. It does not respond to the children, however many times they play it, and nothing about the recording changes in reply to their excitement. The interrecursivity exists between the children, and between the children and their parents, organised around a comparatively stable nonrecursive musical form that neither responds to anyone nor needs to in order to do this work. A stable musical object provides the common ground around which two separately recursive beings become recursively responsive to one another. Music does not connect people by itself, in the vague sense that phrase usually carries. It supplies an occasion stable enough to hold two people's attention on the same thing for long enough that their responses to it can start responding to each other.
Most shared listening never reaches this point. Two thousand people can attend the same concert while undergoing entirely different experiences from one another, and two close friends can listen to the same record with one of them transported and the other checking how much longer it has to run. Interrecursive musical resonance occurs specifically when the responses themselves begin to circulate between listeners, rather than merely being present in the same room at the same time.
The Magic Flute case adds a further complication to this picture. Independently acquired resonance and interrecursive resonance did not simply coexist during that year. They reinforced each other. The children did not love Mozart first and then happen to play together as a separate consequence of it. Playing, singing, and watching together plausibly deepened each child's own individual attachment to the opera, so that each child's resonance fed the shared activity and the shared activity fed each child's resonance in turn, a loop that neither child started or sustained alone. No household could have engineered this convergence in advance. A parent can supply Mozart, recordings, time, encouragement, and permission. A parent cannot manufacture the moment at which two particular children, at four and six, seize on the same opera together and begin intensifying each other's attachment to it through play.
XX. Why Recursive Resonance Cannot Be Transferred
A recording preserves an acoustic organisation with extraordinary fidelity, but every replay is materially re-mediated through a particular room, particular speakers or headphones, and a particular listener's particular state that evening, so it is not quite right to say that a recording reproduces the same patterned sound with no loss whatsoever, or that musical form is transmissible without remainder. What can be said with more confidence is narrower and, in the end, more interesting: recursive resonance cannot be transferred from one person to another. Only the conditions for its reacquisition can be transmitted.
The Magic Flute case is the strongest evidence available for this claim precisely because so much was successfully transmitted and the resonance itself still was not. The household transmitted access, high-quality recordings, encouragement, time, and whatever enthusiasm a parent brings to a favourite piece of music. Parents can do considerably more than arrange conditions from a distance. Singing to a child, dancing with them, playing music together, and sitting through repeated co-listening sessions are forms of active musical socialisation that do real embodied coordination work, entraining a child's sense of rhythm and phrase well before any explicit teaching begins, and all of this happened during that year as well. For a while, something remarkable followed: both children acquired their own vivid, independently sustained recursive resonance with a single opera, reinforced by each other's company.
What the outcome proves is that the resonance itself was never transmitted in the way the recordings and the encouragement were, because once the shared episode ended, neither child became a durable opera lover. Recursive resonance cannot be deposited into another person because its decisive architecture sits mostly below the level of symbolic articulation, in a register that explanation and shared activity do not fully reach. A parent can tell a child exactly why they love a particular piece of music, and can sing it to them a hundred times besides. A parent cannot give the child that love directly, however clearly the reasons are laid out and however often the singing is repeated. What can actually be transmitted, beyond the shared activity itself, is not the resonance but the conditions, the exposure, the scaffolded practice, the repeated co-listening, under which a child might eventually go on to acquire a resonance of their own, on their own timetable and according to their own singular history. Cultural reproduction, on this account, reproduces instruments, recordings, opportunities, and shared activity. It does not reproduce a taste itself, and the disappearance of the Magic Flute attachment shows this more clearly than its emergence did.
XXI. The Extinction of Taste
For the better part of a year, The Magic Flute had enormous recursive relevance inside one family's mesocosm. Then that relevance declined, and it has not returned. The musical object remained exactly where it was throughout. The recordings did not deteriorate. Mozart did not become less good. The household's access, encouragement, and available time did not disappear. Yet the recursive resonance that had been reproducing itself, night after night, for the better part of a year, simply stopped reproducing itself.
This gives the argument a phenomenon it has not yet had room for: the extinction of taste. Taste is not only acquired and then layered indefinitely. Some resonances weaken, close, or become dormant, and their disappearance deserves as much attention as their emergence, since a theory that can explain how an attachment forms but not how it recedes has only told half the story. It would be too strong to say that a lapsed attachment disappears completely. Hearing The Magic Flute again twenty years from now might reopen an astonishing amount of sedimented experience in either child, since a great deal of what was so thoroughly rehearsed rarely vanishes without trace. What did disappear was the active demand for repeated encounter, the nightly request, the spontaneous singing, the argument over which production was better.
This connects taste directly to the more general rhythm of sedimentation, muting, and reopening that governs recursive life well beyond music. A musical attachment can become highly active, sediment deeply through repetition, lose its active recursive relevance once whatever sustained it changes, whether that is a developmental stage, a sibling's shifting interests, or simply the passage of time, and remain available, in principle, for a later reopening that may or may not ever come. Nothing in the conditions that produced the Magic Flute year required the attachment to continue past it, and nothing in the theory developed here would have allowed anyone to predict, in advance, exactly when or why it would end.
XXII. Siblings: Convergence and Divergence
The usual sibling puzzle in a sociology of taste is that children raised in the same household, with similar cultural resources, can develop different musical tastes from each other, and a straightforwardly Bourdieusian account tends to file this under ordinary statistical variation around a shared probabilistic tendency. The Magic Flute case poses a considerably sharper version of the puzzle. It is not merely that these two siblings diverged in taste, which they eventually did, but that they first converged, with unusual intensity, on a single extended piece of music, sustained that convergence for the better part of a year, and then diverged again, neither one carrying the attachment forward into any lasting taste for opera or classical music.
The very same family environment produced both the convergence and the divergence, which means neither outcome can simply be read off family background, since family background stayed constant while the outcome changed twice. Same household conceals, beneath a single portable label, two children who did not inhabit strictly identical mesocosms even during the year of their closest convergence. Birth order, individual temperament, and the ordinary asymmetries of a two-year age gap all shaped how each child approached the opera even while they approached it together, and those same small asymmetries plausibly shaped how and when each child's attention eventually moved on to something else. The sibling case, read this way, becomes a small, well-controlled demonstration of the relationship between social patterning, mesocosmic specificity, and mixed futurity all at once: a shared social and material environment can produce a shared attachment that neither sibling could have been predicted to form in advance, sustain it through active interrecursive engagement between two people, and then release both of them from it again, all without the household itself changing in any way that would explain either the convergence or the divergence on its own.
XXIII. The Limits of Prediction
Three separate questions have to be kept apart. Why was a person's taste possible at all, given the conditions available to them. How did that particular taste actually develop. Could it have been predicted in advance, on the basis of everything known about that person beforehand. A theory of taste can aim for rich answers to the first two questions without owing anyone a confident answer to the third. Mixed futurity implies that a recursive being's future cannot simply be deduced from an inventory of its prior conditions, however exhaustive that inventory might be.
The successful result of an argument like this one is never a claim that the theory explains why a particular person likes a particular composer, as though the outcome were written into their biography in advance and merely waiting to be read off. It is the more modest claim that, given a person's singular history, it becomes possible to understand how a particular piece of music acquired the possibility of resonating for them, without pretending that history made the outcome necessary. No parent, however attentive, could have predicted in September that by the following spring two small children would have built an entire shared world around a single Mozart opera, and no theory of taste should claim it could have predicted this either. Sociology explains why such histories are patterned across a population. A biography explains how one particular history unfolded. Neither one abolishes the contingency of the actual encounter, and that contingency is why music can still surprise people after everything reasonable has been explained about them.
XXIV. Beyond Music
The pattern uncovered through music reaches well beyond musical taste, correcting a much older assumption than any particular claim about taste. Living beings do not simply seek maximal smoothness in their coordination with the world. Play deliberately introduces obstacles into activities that could otherwise proceed smoothly without them. Humour builds an expectation for the sole purpose of violating it at exactly the right moment. Suspense makes uncertainty pleasurable rather than merely bearable, stretching out precisely the discomfort ordinary coordination would try to resolve as quickly as possible. Erotic anticipation prolongs an incomplete coordination rather than resolving it at the first opportunity. Curiosity seeks out encounters capable of disturbing an already settled understanding, and ritual, in many of its forms, holds a community inside a controlled and scheduled departure from ordinary conduct.
A roller coaster is engineered for almost no purpose other than manufacturing a bounded, survivable disturbance to the rider's vestibular sense, and it works as entertainment only because the track, the harness, and the rider's own accumulated trust in both are secure enough to make the drop thrilling rather than dangerous. A tickle fight between a parent and a small child depends on the same structure in miniature: the child's confidence that the parent will stop the moment it stops being fun is what converts what would otherwise be an unwanted physical intrusion into something sought out and laughed at.
The larger discovery these cases point to is that living beings sometimes construct unusually secure, well-differentiated coordination for the specific purpose of making a controlled disturbance possible within it. The security is not incidental to the disturbance, tolerated as a cost of pursuing it, but often built, deliberately and at some expense, so that the disturbance can happen safely, which corrects any model of recursive life in which coordination simply strives, whenever it can, toward an ever smoother, more frictionless fit with its surroundings. A well-loved genre's sedimented conventions, a trusted relationship's settled rhythms, and a religious community's carefully repeated calendar all exist, at least in part, to make a certain kind of valued disruption possible, rather than existing solely to prevent disruption from occurring.
XXV. Bourdieu at the Limit
The Magic Flute episode does not falsify Bourdieu. It sharpens exactly where his account stops working. The children's social position explains why an extended eighteenth-century opera was readily available to a four- and a six-year-old, in multiple high-quality recorded productions, inside a household willing to indulge months of repeated listening, singing, and improvised staging. That much sits comfortably within a Bourdieusian frame.
What Bourdieu's framework cannot supply is an account of why this particular opera caught both children at precisely this moment rather than at any other, why it caught them together rather than separately, why their attachment to it became interrecursive rather than remaining two parallel individual preferences, why it lasted roughly nine months rather than nine days or nine years, or why it disappeared afterwards without producing any lasting taste for opera or classical music in either child. Class position altered what was available and what a household was prepared to indulge. It did not, and structurally could not, reach down to the encounter itself, which is exactly where sociological probability meets its proper limit.
XXVI. The Year It Stopped
Return, at the end, to the two children and the opera. For a short period, a piece of music written more than two centuries earlier became extraordinarily recursively relevant to two small children. Its comparatively stable, nonrecursive form could be replayed as often as they wanted, and each replay operated within a short enough metabolic window to sustain intricate, moment-by-moment anticipation even after dozens of hearings. Repetition sedimented an increasingly differentiated field around this one piece of music with striking speed. Mozart's own constrained departures, the exact place a melody turns, the exact moment a scene resolves, generated recursive resonance within that field, positively valenced and precisely located rather than diffuse. The children did not simply resonate with the opera side by side. Their responses to it circulated between them, through singing, through staged performance, through recognition of each other's anticipation, until each child's attachment was feeding the other's.
Their family and their class position made the whole encounter possible. Neither determined its outcome. No amount of explanation, singing together, or repeated exposure transmitted the resonance itself from parent to child or from child to child, only the conditions under which each child's own resonance could form. And nothing in the original conditions required the attachment to continue once it had formed.
Then it stopped. Nobody decided it should. The music remained exactly where it had always been, note for note, production for production. The recordings did not change. What changed was the two children who had once been so completely caught by them. The most convincing evidence that taste is a recursive achievement in its own right rather than something programmed into a listener by their circumstances is not that an attachment can form so suddenly and so intensely between two small children over an eighteenth-century opera. It is that even an attachment as vivid and as thoroughly shared as that one carries no guarantee, written anywhere in its own conditions, of its own continuation.