Before Analysis, Song
The European cantus tradition, and what it claims about human beings
Spanish version. Antes del análisis, el canto. Translation by the author.
I. The name of a thing that had no name
No one in the Middle Ages, the Renaissance, or eighteenth century Naples ever said they belonged to the European cantus tradition. The phrase is ours, not theirs, and it is worth admitting that plainly from the start. It names a continuity its own participants never had to name, the way a fish has no word for water.
A retrospective name is not for that reason a fiction. Historians constantly name things their subjects could not have named, the Renaissance among them, or the ancien régime, or late antiquity. What matters is not whether contemporaries used the word but whether it isolates something real, a pattern that would otherwise stay scattered across a dozen specialist literatures that rarely speak to one another. Chant scholars, Josquin scholars, partimento scholars, and conservatoire historians have each described a piece of the same animal, without much agreement on what animal it is. The word cantus is a proposal about that.
Here is the proposal. For roughly a thousand years, the fundamental unit of European musical thought was neither the chord nor the form nor the timbre nor the work. It was the line, the cantus, the sung melody, a single strand of pitch moving through time under the control of a breath. Everything else in the system grew out of that unit or was invented to serve it. Theory came after singing, not before it. Notation came after memory. Harmony, once it became an object of thought in its own right, arrived as the residue of lines sounding together, not as their cause.
The Latin word itself carries no special weight. Cantus means song, singing, a melody, a tune, the plain noun of canere, to sing. Its ordinariness is the point. The word that stood at the centre of a thousand years of European musical training was the word an ordinary person would reach for, not a technical term at all, and not one reserved for scholars.
What follows tries to trace that continuity from plainchant to the conservatory, to be honest about where it frays, and to say what survives of it. What survives, I will argue, is not a historical claim but an anthropological one.
II. Sounds perish unless memory holds them
Isidore of Seville, writing his Etymologiae in the early seventh century, left a sentence that anyone thinking about this subject should keep close at hand. Nisi enim ab homine memoria teneantur soni, pereunt, quia scribi non possunt. Unless sounds are held in a man’s memory, they perish, because they cannot be written down.
A modern reader tends to see this as a description of a technological gap, a problem later solved. Read it instead as a description of an entire way of knowing. In Isidore’s world, music was a thing that existed in bodies and was then remembered, not a thing that existed on a surface and was then performed. The carrier of the repertory was a population of trained throats, not parchment.
The scale of what those throats carried is easy to underestimate. The Roman Frankish chant repertory that consolidated in the eighth and ninth centuries runs to thousands of items, introits, graduals, alleluias, offertories, communions, antiphons, responsories, each attached to a specific day in a liturgical year that never repeated itself the same way twice. A cantor was expected to produce the right item, in the right mode, at the right moment, from memory, having been formed in it since boyhood. Notker of St. Gall and the Carolingian cantors were not reading. They were recalling, within a system built to make recall possible.
That system is what modality should mean, and it looks little like the tidy diagram of eight church modes that survives in textbooks. The modes were not scales in the modern sense, abstract collections of pitches from which melodies get assembled. They were families of melodic behaviour, characteristic openings, recurring intonation formulas, cadential turns, reciting tones, permissible leaps, typical ways of getting from here to there. To know the second mode was to have internalised a way of moving, not a set of legal notes. Modal knowledge was procedural before it was propositional, a knowing how rather than a knowing that.
This is why so much medieval chant can be called formulaic without insult. The graduals and the tones for psalms and readings are built from stock gestures adapted to particular texts, in the same way an oral epic singer builds a line from formulas fitted to a metrical slot. Milman Parry and Albert Lord found this structure in South Slavic epic. Medievalists have found essentially the same structure in chant. Formula is not the enemy of art here. Formula is the technology of memory, the thing that lets an enormous repertory live inside human beings rather than inside a library.
And the training that produced such a person was not analytical. A boy in a schola or a monastic choir school learned by singing beside people who already knew, day after day, through the entire cycle of the year, absorbing modal behaviour the way one absorbs the prosody of a mother tongue, without rules, from immersion, by imitation, corrected by ear. The pedagogy was the practice. There was no separate subject called theory that a student passed through on the way to music. There was music, and there were people who had it in them, and there were people who did not yet.
III. Guido, or the tools of the ear
Guido of Arezzo, active in the first third of the eleventh century, a Benedictine monk, teacher, and one of the most consequential practical minds in the history of Western music, is usually presented as the man who invented music notation. He did not, quite. What he did was more interesting, and more relevant here.
Guido’s achievements, as they appear in the Micrologus, the Prologus in antiphonarium, the Regulae rhythmicae, and the Epistola de ignoto cantu, cluster around a single practical problem. Teaching the chant repertory by pure imitation took, by his own account, something like ten years. Guido claimed he could do it in one or two. The staff with its lines and coloured indicators, the systematic use of letter clefs, the hexachord with its six syllables drawn from the hymn Ut queant laxis, the mutation procedure for moving between overlapping hexachords, the mnemonic mapping of pitches onto the joints of the left hand that later tradition attached to his name, every one of these is a device aimed at the same target.
The target was the speed of teaching it, not the preservation of music. That distinction is the whole matter. Guido’s staff is a diagram of intervallic relationships that lets a singer find a melody he has never heard, not a recording medium in the modern sense. The Epistola de ignoto cantu, the letter concerning an unknown chant, is titled with precision. Its subject is the astonishing new possibility that a trained boy could sight sing something no one had taught him. Solmization syllables, in the same way, are not analytical labels but handles. They encode the position of the semitone, mi fa, always and only mi fa, so the ear knows where it is standing, the way a climber knows a handhold by feel. The Guidonian hand, which post dates Guido himself and was elaborated over the following centuries (a small but honest correction to the usual story), serves the same purpose, a place to rest your eyes and your finger while your voice does the actual work.
Guido’s own account of why he did it is unsentimental. He wanted boys to be useful. He was tired of the waste. And in explaining his method he said something that could be inscribed over any conservatory door, that the notation was to be used so the singer might learn to do without it. The written sign was scaffolding, and scaffolding is a thing you eventually take down.
So the first great technological revolution in European musical literacy was designed by a man whose explicit purpose was to serve the ear. That pattern recurs for the next eight hundred years. The page enters the tradition as the servant of the voice. What happens later, when the relation reverses, is the subject of this essay’s second half.
IV. The quarrel inside the tradition, musicus and cantor
Any honest account of the cantus tradition has to face the fact that its own great teacher wrote one of the most contemptuous lines ever aimed at singers. Guido, in the Regulae rhythmicae, put it this way. Musicorum et cantorum magna est distantia, isti dicunt, illi sciunt quae componit musica. Nam qui facit quod non sapit diffinitur bestia. Between musicians and singers there is a great distance, the latter merely utter, the former know what music is made of. He who does what he does not understand is defined as a beast.
The line is no stray insult. It states the doctrinal position of the entire medieval learned tradition, inherited from Boethius, whose De institutione musica placed the musicus, the one who possesses rational understanding of proportion and number, decisively above the cantor, the practitioner, likened to a performing animal. Music in the quadrivium stood beside arithmetic, geometry, and astronomy. It was a science of ratio. The audible was almost an embarrassment, musica instrumentalis, the lowest of Boethius’s three kinds, beneath the harmony of the soul and the harmony of the cosmos.
A naïve version of the cantus thesis, that Europe always valued singing over theory until modernity ruined it, is simply false and should be abandoned. The theoretical, numerical, anti sensory strain runs right through the medieval material, not as an intruder but as a founding member.
But look closely at what the tension actually consists of, because the correction sharpens the thesis rather than breaking it. The Boethian musicus is a philosopher of music, not a producer of it. Nobody learned to make music out of De institutione musica. It was entirely possible to be a distinguished musicus and produce nothing. Meanwhile the actual transmission of the repertory, the daily, institutional, thousand year business of getting music from one generation into the next, ran entirely through singers. Prestige lived in the treatise. Competence lived in the throat.
And when a theorist actually wanted to teach, as Guido did, he was forced immediately into the singer’s world, syllables, hands, memory, formulas, the ear. Guido despises singers who do not understand, and then devotes his life to inventing better tools for singing. That is the structure of the tradition, not a contradiction. Speculative theory sat above practice in the hierarchy of esteem and below it in the order of operations. A scholar could theorise music he could not sing, but no one could make music he could not sing.
Keeping both facts in view is what separates a historical claim from a nostalgia. The cantus tradition was never uncontested. It was one pole of a permanent European argument, the argument between the ear and the number, cantus and ratio, the hand and the page, and its distinctive claim is not that the other pole did not exist, but that for most of the tradition’s life, the other pole lacked the authority to overrule it in the training of musicians.
V. Note against note
The arrival of polyphony is usually narrated as the beginning of harmony, and therefore as the moment European music began to think vertically. Read the sources and this looks like a mistake of retrospect.
Consider what the vocabulary says. Organum is described in the ninth century Musica enchiriadis as a vox organalis added to a vox principalis, an added voice against a principal one, both of them melodies, one derived from the other by a rule of motion. Discantus is dis cantus, singing apart, a second song. Contrapunctus is punctus contra punctum, point against point, note against note, a term that names an operation performed between lines, not a sonority. The tenor of a thirteenth or fifteenth century polyphonic work takes its name from tenere, to hold. It is the voice that holds the preexisting chant, the cantus firmus, the fixed song around which the others are constructed. In sixteenth century partbooks the voices are labelled cantus, altus, tenor, bassus, and the topmost voice is called, simply, the song.
Every one of these words describes a horizontal object and a horizontal operation. Not one of them names a chord. The vocabulary of a discipline is a fossil record of what its practitioners were actually attending to, and this fossil record is unambiguous.
The compositional procedures point the same way. Josquin des Prez (1450–1521), the figure Renaissance writers themselves treated as the summit, typically began from a line, a chant, a chanson tenor, a soggetto, a solmization motto (la sol fa re mi), a hexachordal subject. From that line he generated others by canon, by imitation, by paraphrase, by inversion. The famous Missa Hercules Dux Ferrariae derives its cantus firmus by mapping the vowels of a patron’s name onto solmization syllables, a singing procedure applied to a line, generating the structural spine of an entire mass. The organising intelligence at work is melodic and combinatorial. The vertical result is governed by rules of consonance and dissonance treatment, strict, sophisticated, and entirely negative in form. They tell a composer what a line may not do against another line at a given moment. They do not tell him what chord to build.
None of this means fifteenth and sixteenth century musicians were deaf to sonority. Plainly they were not. The sound of full triadic writing in the late fifteenth century is cultivated, deliberate, and new. But they arrived at that sonority from the side, as the consequence of well behaved lines, and they conceived of it that way. The chord as an independent object with an identity, a root, and a life of its own, the chord as a thing you could name and manipulate before there were any lines at all, is a later invention, and we can more or less date it. Hold that fact in mind for Section IX.
VI. The composer who sang
By the age of Giovanni Pierluigi da Palestrina, roughly 1525 to 1594, the fusion of singing and composing had reached a state of such completeness that it is nearly invisible. It is the water, again.
Consider the shape of his career. Palestrina was a choirboy at Santa Maria Maggiore. He became an organist, then a maestro di cappella, then briefly a singer in the papal choir itself, then maestro at St. John Lateran, at Santa Maria Maggiore, at the Cappella Giulia. Lassus was a choirboy famously abducted for the quality of his voice. Victoria sang. Byrd sang. Almost without exception, the great composers of the sixteenth century were trained singers who had spent their formative years inside the sound of polyphony, producing one of its lines with their own bodies, hearing the other lines arrive around them.
This has consequences no amount of paper study reproduces. A singer inside a five voice texture does not perceive harmony as a sequence of vertical objects. He perceives his own line as a physical trajectory, and the other four as a moving environment he must locate himself within. He knows suspension not as a figure but as the sensation of holding while the ground shifts beneath him. He knows a cadence as an arrival his breath was already preparing four bars earlier. He knows the difference between an interval that is comfortable to tune and one that is not, because he has to find it with his own throat. This is knowledge that lives in the body and resists full translation into propositions, which is exactly why it had to be transmitted by apprenticeship rather than by text.
It is also why the Renaissance treatises read the way they do. Zarlino’s Le istitutioni harmoniche of 1558, the most systematic musical treatise of the century, contains a great deal of number and proportion in the Boethian manner, and then, when it turns to the actual making of counterpoint, becomes a manual of what a line may do. Its rules are framed as advice to a singer imagining a second singer. And the pedagogical practice that stood beside such books, contrappunto alla mente, improvised counterpoint invented in real time above a chant by singers who had internalised the conventions, is the clearest evidence available of what kind of competence was actually being cultivated. Men who could not write a note could improvise four voice polyphony over a plainsong. The skill was not literary. It was closer to what a jazz musician has, a grammar so thoroughly absorbed that invention within it becomes immediate.
Composition, in this world, is performance slowed down and made permanent, not a separate profession from performance. The sources often use the verb cantare even when they mean comporre. A musician learned music by doing music, and what he did was sing.
VII. Naples, the last complete system
The most remarkable thing about the cantus tradition is how late it survived intact, and where. The answer is eighteenth century Naples, in four charitable institutions originally founded to house orphans and poor boys, Santa Maria di Loreto, Sant’Onofrio a Porta Capuana, the Poveri di Gesù Cristo, and the Pietà dei Turchini. The word conservatorio originally meant a place where children were conserved, not a place where music was. That it came to mean the second is a fact about how astonishingly well these places worked.
Their output is difficult to overstate. Out of this system came Pergolesi, Porpora, Vinci, Leo, Durante, Piccinni, Jommelli, Traetta, Paisiello, Cimarosa, and, through its long afterlife, Zingarelli’s pupil Bellini. For a century, Europe’s opera houses from Lisbon to St Petersburg were staffed by men trained in a few square kilometres of one city. Something in the method was working, and thanks to a generation of scholarship, Robert Gjerdingen, Giorgio Sanguinetti, Nicholas Baragwanath and others working directly from the surviving manuscript exercises, we now know in some detail what it was.
It was not a theory course. There were no lectures on harmony, no textbook of chords, no analysis. There were, essentially, two intertwined disciplines, and both of them are cantus disciplines.
The first was solfeggio, not solfège in the modern sense of a sight singing drill, but a years long daily practice in which a student sang melodic exercises, often to a written bass realised at the keyboard by the master, using the old hexachordal syllables with mutation, the direct descendant of Guido’s system, still alive in Italy in the age of Mozart. The solfeggi of Leo, Durante, Porpora and others survive in quantity. They are exercises in idiom, not exercises in reading. Sung across years, they install in the student’s ear and throat the entire melodic vocabulary of the style, its cadential formulas, its sequences, its ornamental habits, its typical ways of leaving a phrase and returning. The student was not being taught to decode notation. He was being taught a language in the only way languages are ever really learned, by producing it, thousands of times, until it becomes automatic.
The second was partimento. A partimento is a bass line, sometimes figured and often not, written on a single staff, from which the student improvises a complete composition at the keyboard. It looks like almost nothing on the page. In practice it is a compressed score. The student must supply the upper voices, the imitations, the sequences, the modulations, the texture. Fenaroli’s regole, Fedele Fenaroli (1730–1818) is the great codifier, and his rules were still being reprinted a century after his death, begin with the regola dell’ottava, the octave rule that assigns a standard harmonic behaviour to each degree of the scale ascending and descending, then move through the moti del basso, the standard bass motions, the cadences, the suspensions, and the modelling of imitative textures.
What is being installed by all this is best described by Gjerdingen’s term, schemata. Stock contrapuntal patterns, the Romanesca, the Prinner, the Fonte, the Monte, the Ponte, the various cadential formulas, each consisting of a characteristic pair of lines, a bass and a melody moving together in a fixed relationship, with everything else filled in as needed. These are not chord progressions, although they can be described that way after the fact. They are voice leading gestures. A student who has absorbed two hundred of them can generate music the way a fluent speaker generates sentences, not by consulting rules but by deploying patterns he no longer has to think about, varied and combined and elided in the moment.
Two features of this training are worth underlining, because they define the whole tradition. The ear and the hand were trained together and never separated. Solfeggio was sung, partimento was played, and both were done daily, in the same room, by boys who also sang in church and played in the orchestra. There was no stage at which a student learned about music before doing it.
And more radically, the direction of causation ran from pattern to rule, not rule to pattern. A Neapolitan student did not learn the principle of harmonic progression and then apply it. He learned two hundred concrete musical objects and could, if pressed, describe what they had in common. Competence preceded articulation. This is exactly how a child acquires a native language, and close to the opposite of how music theory has been taught for most of the last hundred and fifty years.
Naples is therefore the crucial case, because it is close enough to us to be documented in detail and yet still fully inside the old order. It shows that a system built on singing, imitation, memory, and improvisation was a highly evolved, extraordinarily productive technology of instruction that produced fluent composers in numbers we have not matched since, not a primitive stage Europe grew out of once it developed proper theory.
VIII. Paris, the tradition becomes an examination
The Paris Conservatoire was founded in 1795, an institution of the Revolution, and it inherited the Italian system in a peculiar way. It took the material and changed the frame.
The material really was Italian. The Conservatoire’s own Solfèges du Conservatoire, assembled around 1801 by Cherubini, Méhul, Gossec and others, drew directly on the Italian solfeggio repertory. The curriculum’s assumptions were still recognisably those of the older world. A musician was expected to sight sing, take dictation, harmonise a given part, transpose at sight, realise a figured bass, accompany, and write counterpoint and fugue. These are practical, generative, ear based competences. Nothing on that list is a matter of analysis. A student who could do all of them possessed, functionally, most of what a Neapolitan student possessed.
But the frame was new, and the frame did the damage. A Neapolitan conservatory was a household. Boys lived there, sang the daily offices, played in each other’s performances, and were apprenticed to a master whose judgement was personal, continuous, and unwritten. Advancement meant becoming useful. The Paris Conservatoire was a modern bureaucratic institution, fixed classes, fixed years, syllabi, printed method books produced for the purpose, competitive examinations, public concours, prizes, medals, and eventually the Prix de Rome. It was the Napoleonic state applied to musical training, and it worked the way such things work, enormously extending access and standardising quality, while quietly transforming what the training was for.
An examination has to be gradable. Grading requires that the thing assessed be discrete, comparable, and defensible. Under that pressure, competences that were once continuous and embodied turn into exercises with correct answers. The most visible casualty was improvisation, hard to mark and easy to disagree about. It survived longest in the organ tradition, where it remains alive to this day, and withered nearly everywhere else. Harmony, once the byproduct of realising a bass at a keyboard, became a written exercise in four parts to be corrected in red ink. Solfège, once a years long immersion in style, contracted into the reading of pitches in time. The dictée musicale, musical dictation, that quintessentially French institution, became a graded, systematised discipline with its own published literature.
It is exactly here that Paul Rougnon (1846–1934) belongs, a long serving Paris Conservatoire professor and prolific author of graded solfège, theory, and dictation manuals that trained several generations of French musicians. Books like his are the tradition in its administered phase, and it is worth saying plainly that this is not a criticism of the books themselves. Rougnon’s dictation volumes assume that the ear is the organ of musical intelligence and that it can be systematically developed. That assumption is pure cantus tradition. What changed is that the ear is now being developed toward an examination rather than toward the ability to make music on the spot. The competence survived. The purpose drifted.
This is the characteristic fate of living traditions inside modern institutions. They get conserved, in the museum sense of the word. The Neapolitan conservatorio conserved children. The nineteenth century conservatoire began to conserve music.
IX. Rameau, Rousseau, and the inversion
Behind the institutional story lies an intellectual one, and it has a date and a book. Jean Philippe Rameau, Traité de l’harmonie réduite à ses principes naturels, Paris, 1722.
Rameau’s achievement is genuinely enormous, and it should not be caricatured. He proposed that the bewildering variety of chords in use could be reduced to a small number of fundamental entities, triads and seventh chords, which retain their identity under inversion. Each such entity has a basse fondamentale, a fundamental bass that may not be the sounding bass, and harmonic progression can be understood as the motion of these fundamental basses. Later he tried to ground the whole system in nature through the corps sonore, the overtone series. This is the origin of essentially all subsequent harmonic thinking, roman numerals, chord inversion, functional harmony, the entire apparatus of the harmony class.
And it inverted the order of the tradition, explicitly and on purpose. Rameau argued that melody is derived from harmony, that a singable line is intelligible because it expresses an underlying harmonic succession, which is the prior and generative fact. The line becomes an epiphenomenon of the chord.
Rousseau saw immediately what was at stake and attacked it for the rest of his life. In the Essai sur l’origine des langues and elsewhere, he argued that melody is prior, that song and speech share an origin in the accented, passionate cry, that music’s power over us is a power of intonation and therefore of imitation of the human voice, and that harmony is a northern, mathematical, corrupting refinement that had begun to smother the melodic and expressive core. Much of Rousseau’s specific ethnography and history is simply wrong. But the philosophical position is exactly the cantus thesis stated in Enlightenment terms, and the Rameau Rousseau quarrel is therefore the founding modern statement of the argument this essay is about, not a squabble about taste. Is the line the origin, or the consequence.
Rameau won, institutionally and completely. Not immediately, Naples continued undisturbed for decades, and Italian practice kept its own vocabulary well into the nineteenth century, but decisively. D’Alembert’s popularisation of 1752 put Rameau’s system into general circulation. The Conservatoire’s harmony teaching absorbed it. The nineteenth century systematised it further through Fétis’s theory of tonalité and, in Germany, through Riemann’s functional categories. By 1900 a European musician’s basic conceptual furniture consisted of chords, keys, functions, and forms. By 1950 the training of a composer began with analysis.
The inversion can be stated plainly. The old order ran from song to ear to memory to improvisation to composition to theory. The order that gradually replaced it ran from theory to analysis to notation to interpretation.
In the first, theory is the last thing a musician acquires, a description of what he can already do. In the second, theory is the first thing he acquires, and performance is the terminal step, in which an interpreter renders a text that was itself generated from concepts. In the first, the musician is a producer of music. In the second, he is, in the last analysis, a reader.
X. What the twentieth century was resisting
Certain twentieth century musicians are unintelligible unless you see them as reacting against this inversion. Their pronouncements sound eccentric, mystical, or merely cranky if taken as claims about aesthetics. They are not that. They are claims about pedagogy and about the order of knowledge.
Chopin is the earliest and clearest case, and he comes from within the century that made the change. His documented teaching is almost entirely about singing. He told students to listen to the great Italian singers, held up Bellini’s line as the model of phrasing, said that a badly formed phrase reveals a person who has never heard good singing, and grounded his technique on the idea that the hand should produce a legato imitating the voice. His own sketch of a piano method opens with sound, not mechanism.
Alfred Cortot built an entire interpretative pedagogy on the same premise, that the pianist’s first task is to identify and project the singing line and that everything technical exists to serve it. His editions and his cours d’interprétation return again and again to the same word. Chantez. He was fighting the pianist who plays chords.
George Enescu was a legendary case of the older kind of musicianship surviving into the age of recordings, a musician of nearly total aural memory, who taught by singing phrases at his students and by playing whole scores from memory at the keyboard. Menuhin’s accounts of him describe not a pedagogue explaining structures but a musician transmitting a shape directly, ear to ear, in the manner of a master craftsman.
Sergiu Celibidache made the argument explicit and gave it a philosophical frame. His hostility to recordings, his refusal to permit them, his insistence that a record is not music but a photograph of a meal, amounted to a claim that music exists only as a lived temporal act between people in a room, and that anything else is a document about music. His insistence that tempo cannot be fixed in advance but arises from the acoustical and human conditions of the moment is the same claim applied to time. His whole position is a refusal of the primacy of the text.
Salvador Chuliá belongs to this lineage in its most ordinary and most convincing form, the working teacher who has not read a manifesto about any of this and does not need to, because the tradition is simply the way he was taught and the way he corrects. A student brings a harmony exercise. It follows every rule. The teacher does not examine the page. He sings it. And says, está perfecto, pero no canta.
We will come back to that sentence. What these figures share is not a style, not a period, not a nationality, but a conviction that the authority in music lies somewhere other than where their century had put it, not in the rule, not in the analysis, not in the page, but in something audible and singable and finally bodily, which a trained musician can hear directly and which no correct answer can substitute for.
XI. Objections
A thesis this large invites objections, and the honest ones improve it. The first is that this is nostalgia. The past gets described as a lost wholeness and the present as a fall, a shape of argument that should trigger suspicion on principle, since it is available for every subject and is almost always false. The correction is straightforward. What was gained by the shift is enormous and deserves to be said plainly. Notation centred training democratised access to music at a scale no apprenticeship system could approach. Analysis gave us the ability to understand structures too long and too complex to hold in the ear, nobody improvises their way to Wozzeck. The nineteenth century canon, the whole idea of a repertory of past works studied and performed, is a product of the literate order and one of the great achievements of European culture. The claim here is that it was a trade, and that the cost side of the trade has gone largely unexamined, because the winners write the curriculum, not that the inversion was a catastrophe.
The second is that the line never actually died. True, and importantly so. Schenker’s analytical project, whatever one makes of it, is a reassertion of linear primacy. The Urlinie is a melody, and his whole method reads harmony as the elaboration of voice leading. Ernst Kurth’s Grundlagen des linearen Kontrapunkts of 1917 argued for melodic energy as the generative force in Bach. Twentieth century pedagogy has repeatedly tried to restore the ear, Dalcroze through movement, Kodály through singing and a mobile do solmization that descends directly from Guido, Orff through elemental music making, Ward through chant. The Solesmes restoration of plainchant, and the Motu proprio of 1903 that endorsed it, was an institutional attempt to put chant back at the centre of Catholic musical life. The cantus pole of the argument has never lacked defenders. What changed is not that it vanished but that it lost the default position.
The third objection asks whether any of this is specifically European, and it is the sharpest one, because voice first, memory based, formula driven, orally transmitted musical training is the human norm, not a European peculiarity. Maqam, raga, Byzantine chant, Georgian polyphony, flamenco, Balkan epic, Irish sean nós, West African praise singing, everywhere one looks, music is transmitted by ear, learned by imitation, generated from internalised patterns, and carried in bodies. If the cantus thesis simply says music comes from the voice, it is true and unremarkable.
The correction is worth making carefully, because it locates what is genuinely European. Europe is unusual not for singing but for writing. No other tradition built so powerful a notation so early, or let it accumulate so much authority. The distinctive European situation, from Guido onward, is the coexistence and tension between an oral vocal practice and a written apparatus of extraordinary capability, a tension that made possible things no purely oral tradition can attempt, a forty part motet, a symphony, a work fixed for four centuries, and that carried, from the beginning, the risk that the apparatus would swallow the practice. The cantus tradition is the singing pole of a specifically European argument between voice and letter, not simply that Europe sang, an argument that ran for a thousand years and was, somewhere between Rameau and the modern conservatory, decided.
The fourth objection asks whether this conflates vocal music with melody. Partly, and deliberately. The claim is that vocality functioned as the criterion even where voices were absent, not that all European music is for voices, which is obviously false, that instrumental lines were judged by whether they could be sung, that phrasing was theorised on the model of breath, that the terms of praise, cantabile, singend, chantant, are vocal terms applied to instruments. The voice was not the medium. It was the standard.
XII. The anthropological core
Strip away the history and a claim remains that is not historical at all. It concerns how music enters a human being, and it is the reason any of this still matters to someone who will never sing a note of plainchant.
Music enters the human being through the voice. This is not a metaphor. An infant’s first musical experience is a human voice modulating in pitch and rhythm at close range, motherese, which is cross culturally universal, exaggeratedly melodic, and understood by the child long before any word is. Musical response is present before language. The instrument by which the child first participates is the same instrument by which the child first cries.
From that fact follow four priorities, each of them a claim about order. Before there is harmony, there is intonation. To sing a pitch a person must find it, in his own body, with muscles he cannot see, guided only by hearing. Everything vertical presupposes this act. Nothing replaces it. Someone who cannot match a pitch does not have a theoretical deficiency.
Before there is form, there is gesture. A phrase is not a formal unit. It is something that departs and arrives, and it does so on a breath. Every large musical form in Europe, period, sentence, aria, sonata, symphony, is an elaboration of what a breath does, tension, release, going out, coming back. Form is expanded gesture, not the other way around.
Before there is notation, there is memory. Isidore’s sentence still stands. Notation is a mnemonic prosthesis of astonishing power, but a musician who can only reproduce what is set before him possesses the prosthesis and not the limb.
Before there is analysis, there is song. Analysis describes. Song produces. Both have their value, but only one of them generates anything, and confusing the order, teaching the description before the thing described, produces musicians who know a great deal and can make almost nothing.
The unifying idea is that music, on this view, is not primarily information, not a structure that happens to arrive in sound, but a bodily act with a temporal shape, performed by a creature that breathes, and its structures are the sediment left by that act. The score is a set of instructions for producing it. Theory is a description of regularities within it. Analysis is a map of it. Instructions, descriptions, and maps are all useful, and none of them is the territory.
This is why the tradition’s competences look, to a modern eye, strangely practical, sight singing, dictation, transposition, realising a bass, improvising a counterpoint. Each one is a way of connecting the ear to production without passing through concepts at all. They are not tests of musical knowledge. They are the musical knowledge.
XIII. Está perfecto, pero no canta
Return to the exercise on the desk. It is perfect. Every rule has been obeyed, no parallel fifths, no octaves, correct doubling, resolved sevenths, sound voice leading, an unimpeachable cadence. In the order that begins with theory, there is nothing more to say. The work has satisfied its criteria. It would receive full marks in any examination in Europe.
The teacher does not look at it. He sings it. And then he says the sentence that carries the entire tradition inside it. It’s perfect, but it doesn’t sing.
Consider what that judgement presupposes. It presupposes, first, that correctness and quality are different things, that the rules are a lower bound, a description of the avoidable failures, and not a definition of success. The rules were always derived from good practice. They were never its source. A student who has satisfied them has demonstrated only that he has done nothing wrong.
It presupposes, second, that a court of appeal exists above the rules, and that this court is audible. Not audible in the weak sense that music is heard, but in the strong sense that a competent musician can hear whether a line is alive, can hear it directly, and can be right about it in a way that admits no further justification. The judgement is not derived from anything. There is no rule the teacher could cite that the exercise violated. He is exercising a perception, formed over decades, which is the whole content of his training.
It presupposes, third, that the way to test music is to produce it. He did not read the exercise. He sang it, put it back into a body, gave it breath, made it move in time, because that is the only condition under which the property in question becomes detectable at all. Cantability is not visible on paper, not a feature of the notation, but a property of what happens when the notation is executed by something that breathes.
And it presupposes, fourth and last, that the student can be brought to hear it too. This is what makes the sentence a piece of teaching rather than an assertion of authority. The teacher is not overruling the student. He is pointing at something and saying, listen. Once you have heard it, you will not need to be told again. It is the same transaction as a Carolingian cantor correcting a boy’s intonation formula, as Zarlino’s imagined singer, as a Neapolitan master shaking his head at a partimento realisation, as Cortot saying chantez. It has gone on, in essentially unchanged form, for a very long time.
That is what the European cantus tradition finally is, not a school, not a period, not a style, but a millennium long agreement about where the authority in music resides. The tradition holds that it resides neither in the rule nor on the page, but in the living, singable line, a thing that exists only when someone produces it, that can be perceived only by someone trained to perceive it, and that no correct answer has ever replaced.
Está perfecto, pero no canta. Five words, and a thousand years standing behind them.
Edition and citation
How to cite. Josu De Solaun, “Before Analysis, Song,” Melosías, 31 August 2026, permanent link.