Paper 041 - The Dependency Model

Kyle Benford, Department of Jazz

Status of this revision

This paper replaces the first circulated version of Paper 041. That version did not survive checking. It carried a wage figure that was wrong, tuition figures that were out of date, a biographical claim about Thelonious Monk that could not be sourced, a statement that jazz was built "without institutions" that the historical record contradicts, an unsourced claim about what clubs pay, and an originality claim that the scholarly literature refutes. Apparatus A lists every correction, line by line.

The argument that remains is narrower than the first version's. The Department believes it is also harder to answer, because it no longer depends on anything a skeptical reader can disprove in one search. It is written to the standard the Department asks of others: every number carries a source, every source is labelled by how well it was checked, and every claim is stated at the strength the evidence supports and no stronger.

Abstract

Jazz education in the United States is usually defended by pointing to its teachers and attacked by pointing to its costs. Both arguments miss the structural question. This paper argues that the jazz curriculum, as it developed after 1945, transmits the work of the tradition (repertoire, harmony, rhythm, history) while removing the mechanism that originally disciplined that work: repeated, paid, public exposure to an audience that is free to leave. We call the resulting arrangement the dependency model. In it, the student's feedback, status and eventual employment are supplied by the institution that credentials them, rather than discovered in a market that prices what they actually do.

The paper does not claim that degrees cause mediocrity. No dataset measures mediocrity, and the first version of this paper was wrong to imply otherwise. It makes five narrower claims. First, jazz pedagogy has always been institutional, but its earliest institutions answered to touring and paid performance. Second, the contemporary pipeline answers largely to credentials and academic employment. Third, the vocabulary the pipeline teaches is unpriced by design, because copyright protects fixed expression and not idiom. Fourth, the available indicators (live jazz attendance falling from 10.8 percent of U.S. adults in 2002 to 6.3 percent in 2022, a federal occupational profile that lists no postsecondary degree as typical for entry, and a labor force in which roughly half of musicians are self-employed) are consistent with a field whose audience market has shrunk while its credentialing infrastructure has persisted, though they do not establish cause. Fifth, each claim generates a falsifiable prediction, and Section XI states what would refute the paper.

The paper also records what it got wrong. Apparatus A lists fifteen corrections to the first circulated version.

How to read the evidence

Every factual claim in this paper is labelled in Apparatus B with one of four tiers.

• A primary or near-primary source was read for this revision.
• Widely documented in standard biographies and histories. Not re-pulled for this revision. Cite the book, not this paper.
• Taken from the Department's research file, which compiled it from institutional web pages. Confirm on the current page before quoting.
• The author's argument or experience. Offered as a claim to be tested, not as a fact.

Where a sentence carries a number, its tier appears in parentheses the first time the number is used.

I. The claim, stated precisely

The thesis in one sentence: a jazz student can master the language of the tradition inside an institution whose feedback never tests whether anyone will pay to hear it spoken, and the institution has little incentive to find out.

Four terms carry the argument, and each is used in a fixed sense.

Any setting in which a performer plays for an audience that pays, can respond in real time, and can leave. A club, a wedding, a street corner, a session. The defining property is not hardship. It is exit: the listener's freedom to walk out is the discipline.

A certificate issued by an institution attesting that the holder completed a course of study. A credential is a signal. Whether it also certifies capability is an empirical question that differs by labor market.

A measurable stand-in for something harder to measure. A grade is a proxy for mastery. A jury is a proxy for readiness. A rebooking is a proxy for an audience's willingness to pay again. Proxies are indispensable, and every proxy drifts from what it stands for once it becomes the target.

An arrangement in which a student's three principal needs (feedback on quality, status within the field, and eventual employment) are all supplied by the same institution. The dependency is structural. The paper makes no claim about the character, ambition or talent of anyone who passes through such an arrangement. Most of them, as the author can attest from the inside, are extraordinarily serious people.

The contrast is with the bandstand model, in which feedback comes from audiences, status from peers on the stand, and employment from bandleaders who can be wrong about you and will be corrected by the room.

What the paper does not claim

• It does not claim that jazz degrees cause mediocrity. No dataset measures mediocrity, and the word is used in this paper only in a defined sense (Section V).
• It does not claim that jazz education is a cause of any broad social problem. The first version's circulated summaries suggested this. There is no evidence for it, and the claim is withdrawn.
• It does not allege that any institution or individual has acted unlawfully or in bad faith. Section VI uses an analogy to a historical scandal to describe a structure. It is not an accusation.
• It does not evaluate any living musician, teacher or student.
• It does not claim the academy has produced no excellent players. It has produced many, and Section III says so.

The claim that remains is smaller, sharper and testable. It is that institutions which supply a student's feedback, status and employment together will, over time, optimize the student for the institution's proxies, and that this matters most in an art form whose original discipline was the audience's freedom to leave.

II. Before the academy

The first version of this paper said that the founders of jazz "did not have institutions." That sentence is false, and it is false in a way a historian can disprove from memory. Jazz was taught in institutions from the beginning. What matters is what those institutions answered to.

The standard record offers several examples. Louis Armstrong learned the cornet at the Colored Waifs' Home for Boys in New Orleans, where Peter Davis led a band that played in public. The Jenkins Orphanage in Charleston, South Carolina, ran bands that toured, and its alumni include the trumpeters Jabbo Smith and Cat Anderson. Duke Ellington took piano lessons as a child in Washington, D.C., from a teacher named Marietta Clinkscales, and he never took a conservatory degree. Charlie Parker, according to a review of Stanley Crouch's biography in the New York Review of Books, studied with a teacher named Alonzo Davis, and the review characterizes that teacher as an heir to a tradition of conservatory-trained Black teachers. The "self-taught Parker" is, on that account, a myth. John Coltrane studied at Granoff Studios in Philadelphia on the G.I. Bill and took private lessons from the guitarist Dennis Sandole.

A reader of the record finds pedagogy everywhere. The correct claim is therefore not that jazz had no schools. The correct claim concerns three features of the schools it had.

Davis, Sandole and the bandleaders of the orphanage bands were not credentialed to teach by an academic body. They were credentialed by the fact that they were playing, for money, in public, at the time they taught.

The Waifs' Home band marched in New Orleans. The Jenkins bands toured. Parker's apprenticeship at Kansas City sessions is the most famous case. In 1936, at the Reno Club, the teenage Parker tried to sit in and was cut off by the drummer Jo Jones, in a story usually told with a cymbal thrown across the room. The story exists in several versions, and the cymbal may be embellishment. In every version, the room corrected him in public, and he went away and practiced. That is price discovery performed with a drum kit.

The audience could leave. The bandleader could cut a tune. The next gig did not come. The feedback was crude and it was honest, in the specific sense that nobody receiving it could be paid to soften it.

The market was not kind

Nothing in this paper romanticizes the market that disciplined these musicians. It was segregated. Club owners and agents took large cuts. From the 1940s until 1967, New York City required performers in establishments that served alcohol to hold a police-issued cabaret card, and a conviction, including a drug conviction, could cost a musician the right to work in the city's clubs. Billie Holiday and Charlie Parker both lost theirs. Parker died in March 1955 at the age of 34. Bud Powell died in August 1966 at 41.

The claim is that the market functioned as a feedback mechanism, not that it was a moral order. The Department's program (Section X) is an attempt to keep the feedback and refuse the predation. A paper that confused those two things would be a nostalgia piece, and the tradition has had enough of those.

III. The academy arrives

The Schillinger House opened in Boston in 1945 and later became Berklee. North Texas, which claims the first jazz-studies degree in the country, began offering jazz degrees in the late 1940s. The G.I. Bill of 1944 put tuition money behind a generation of returning servicemen who wanted to study music, among them Coltrane. By the 1960s and 1970s, jazz education had spread through universities and high schools across the country.

The scholarly literature on this transition is substantial, and the first version of this paper wrongly claimed that its argument was original. Paul Lopes, in (2002), traces how jazz moved from popular entertainment toward art in part through universities, nonprofits and critics. Scott DeVeaux's 1991 essay "Constructing the Jazz Tradition" shows how a canon was assembled. Kenneth Prouty's (2012) examines jazz pedagogy and its effects on community and canon. This paper stands on that work. Its contribution, if it has one, is to put the question in economic terms: not whether institutionalization legitimated jazz (it did), but what feedback the institutions removed in doing so.

What the academy got right

The academy gave access to students who had no mentor and lived in cities with no clubs. It preserved the tradition through transcription, recording, archives and history. It built theory that makes some things learnable in months that once took years. It offered a room safe enough to fail in. And it gave working musicians a livelihood: teaching is real work, and many of the field's most important players have paid their rent with it. Anyone who argues against jazz education without admitting all of that is not arguing in good faith.

In September 2026 the MacArthur Foundation named a jazz musician among its fellows. The Department treats that as good news for the whole field. It is also, for this paper's purposes, a reminder that the argument does not depend on denying the academy's excellence. The claim concerns incentives and feedback, not talent.

IV. What the numbers say, and what they do not

Reading the table honestly

The first version cited a median annual wage of about $32,000. That number was not supported. BLS reports a median of $42.45 per hour for musicians and singers who are wage-and-salary workers. The figure excludes the self-employed, who are roughly half the occupation, and it cannot be annualized, because the hours are irregular. The honest summary is that the hourly rate is not the problem. The problem is the number of paid hours, and no federal series captures it well. This is why Section XI proposes a direct survey.

The same NEA survey shows declines across most live arts forms between 2017 and 2022, a period that includes the pandemic. The jazz decline therefore cannot be pinned on jazz education. What the figure does show is that the audience, the market in which the bandstand model operates, shrank by roughly 40 percent as a share of adults over two decades, while the infrastructure for credentialing players continued to exist and, in many cases, to grow. That is a correlation, and the paper offers it as one.

The first version spoke of "credential overproduction" and "pollution of the scene." The count of jazz-studies completions in 2023 was 589 nationwide. That is a niche, not a flood. The count undercounts, because many students who play jazz hold broader music-performance degrees under other classification codes, but no number in hand supports the word "overproduction." The claim is withdrawn as a quantitative statement and retained only as a hypothesis (H2, Section XI).

The sticker price overstates what most students pay, because aid is widespread and some programs are free. The point of the arithmetic below is scale, not any individual's debt. A four-year sticker price of about $222,000 at Juilliard's 2025-26 rate, or about $209,760 at Berklee's, divided by the BLS hourly median of $42.45, equals roughly 5,230 and 4,940 paid hours respectively. At twenty paid hours a week, that is about five years of full-rate gigs with no living expenses, no taxes and no gaps. No one lives that life. The exercise only shows that the price of the credential is measured in a currency, paid hours, that the credential does not itself supply.

None of these figures measures artistic quality, the value of the education a given student received, or what any alumnus earns. Those are the questions that would test the thesis, and Section XI describes how to ask them.

V. The pricing problem

A jazz musician who wants to work faces two labor markets. They have different buyers, different prices and different rules about what counts as qualification.

The buyers are audiences, clubs, festivals, private clients and bandleaders. They pay for a night. Nobody in this market asks to see a transcript. A bandleader hiring a drummer wants to know whether the drummer can hold the time at 2 a.m. when the room is tired. The audience wants to know whether the night was worth the cover charge. The credential carries no information that the playing does not already carry, and the playing is observable the same evening.

The buyers are universities, conservatories and schools. They pay for a semester. In this market the credential is the entry ticket: teaching positions in higher education typically require a degree and often a graduate degree. The credential has real value here, and its value is that of a signal, in exactly the sense Bryan Caplan's (2018) describes: it certifies traits such as persistence and conformity to a buyer who cannot observe the student directly.

The structural observation is that these two markets are not symmetrical in size or in control. The performance market is large, diffuse and, as Section IV shows, shrinking as a share of adult attendance. The academic market is smaller, but the institutions that issue the credential are also the institutions that hire the credentialed. We call this the closed loop. A student who completes a jazz degree can, if all goes well, be hired by the kind of institution that issued it, to credential the next student. The loop is not corrupt. Many of the people inside it are exactly the working musicians the tradition would want teaching. But a loop is a loop, and whether it is the main destination for graduates is an empirical question (H2, Section XI).

Inside the loop, the measurable things are grades, juries, auditions, placements and recitals. These are proxies. Paper 040 named the general failure: a system that keeps optimizing a proxy after the connection to the source has frayed. Goodhart's law (1975) states it plainly: when a measure becomes a target, it ceases to be a good measure. A jury is a decent proxy for readiness at the start of a degree. When it becomes the thing a student optimizes for over four years, the product can be a player who is excellent at being evaluated by a jury. Nothing about this requires bad faith from anyone. It follows from the structure.

The Department's Jazz Ascensionism paper states that "Mediocrity is elitist," and observes that conservatories "teach repertoire and technique but rarely teach philosophy or markets." Here the word is used in one defined sense: mediocrity is the state of a system that rewards conformity to its proxies over the discovery of value an audience would actually pay for. It is a property of systems, not of people. A system can be mediocre while every person inside it is brilliant. That is the painful part, and it is the part worth saying out loud.

Picture a twenty-two-year-old who can play any tune in the canon in any key, who knows the history, who has transcribed a hundred solos. She has been told for four years that she is excellent, and she is. She has never had to fill a room. Nobody has told her what a Tuesday night is worth, or how to find out. She has never booked her own gig, because the institution booked it for her. This is an illustration, not a data point, and it appears here because the Department's claim is that this is a typical shape of the gap. The gap is not her fault. The place where she was supposed to learn the price of her playing was never put in the curriculum.

On a bandstand, price discovery happens nightly and in public. The door count, the tips, the second set that is half the size of the first, the leader who rebooks or does not: these are market signals, and they arrive within hours. The Department's thesis is that performance is the asset and the recording is the archive. In jazz the live night is the original work, and a recording is a copy of it. The recorded-music industry has spent a century paying for copies. A curriculum that teaches the archive without the original has taught the wrong asset.

Some programs require paid public performance, teach business, or book their students into real venues. This paper's claim concerns the typical program, and it is testable (H1). If programs with those features show no difference in alumni outcomes, the paper's mechanism is wrong.

VI. A structure, not an accusation

Alan Freed was the disc jockey who helped carry Black rhythm and blues to white teenage audiences in the 1950s. He accepted payments from record labels for airplay, and he took a co-writing credit on Chuck Berry's "Maybellene" (1955). The practice, called payola, produced congressional hearings in 1960. Freed was fired in 1959, pleaded guilty to commercial bribery in 1962, and died in 1965 at the age of 43.

An earlier Department summary compressed the argument to "jazz education is payola." That was a slogan, and slogans overreach. Payola was concealed payment for access. Tuition is disclosed. Nobody at a conservatory is alleged to have broken any law, and the Department makes no such allegation. What the analogy captures is a structure with three parts:

1. A gatekeeper stands between a source and an audience.
2. Money flows to the gatekeeper for transmission, not to the source.
3. The source is credited as heritage while the gatekeeper is compensated.

The analogy fails in several places. Berry was a living artist whose records generated income for him, credit disputes aside. Payola was secret, and tuition is public. And the gatekeepers in education are mostly working musicians who need the work. The paper therefore keeps the structure and drops the word.

Why the vocabulary is unpriced

Copyright protects fixed expression: a particular recording, a particular composition. It does not protect an idiom, a harmonic language or a rhythmic approach. Nobody owns the bebop vocabulary. This is, on the whole, a good rule, because a world in which a harmonic device could be owned would be a world in which no one could play.

But the rule has a consequence the first version of this paper mishandled. It claimed that the lineage "has never been paid." That overstates the case. Estates and labels that hold rights in recordings and compositions do receive royalties, and the paper has no evidence about any particular estate. What is unpriced is narrower and more interesting: the transmission of the idiom itself. When a conservatory teaches a student to play in the language that Parker and Powell developed, the idiom is free, the tuition is not, and nothing in law or in the market routes any of the value of that transmission back to the source.

The Department does not propose a legal remedy. It proposes institutions that price what can be priced (the night, the commission, the fellowship) and that treat the living lineage as stakeholders rather than as heritage.

VII. What the unpriced source cost

A question has been circulating in tech-philosophy circles: where is America's Virgil? In an interview published on September 5, 2025, the writer Johnathan Bi and Brendan McCord of the Cosmos Institute discussed autonomy and AI. Google's summary of the transcript attributes the question to Bi, as a way of pointing at an absence of foundational artistic genius in the United States relative to its power. The transcript is behind a paywall. The exact wording and argument should be checked against it before anyone quotes it, and the Department engages the question as it circulates, not as any individual's settled position.

The Department's earlier paper, Jazz Ascensionism, answered the question directly: we made Coltrane. This paper adds that the answer costs the country something to say, because of what the country did with the people who made the music.

(1924-1966) was one of the architects of modern jazz piano. From the mid-1940s he was repeatedly institutionalized. According to Peter Pullman's biography, which draws on hospital records Pullman obtained through a court order, Powell received about thirty electroconvulsive treatments at Creedmoor in 1948 and thirty-three more in 1952. A police beating in 1945 is also widely blamed for his troubles. Accounts differ, and the paper takes no position on it. In a letter written days before he died at Kings County Hospital in 1966, he wrote: "I'm a writer and composer, and these treatments are destroying my brain." The line is reported in Wikipedia's entry on Powell, and the primary text should be checked before it is reproduced.

was playing Birdland in the winter of 1951, according to Alfred Appel's , with Igor Stravinsky in the audience. Parker slipped a quotation from into his solo as though it had always belonged there, and Stravinsky was reported to be delighted. Four years later Parker was dead at 34, in a suite belonging to the Baroness Pannonica de Koenigswarter at the Stanhope Hotel. (Paper 039 called it her "apartment." That is corrected in Apparatus A.)

On electroshock, a version of Parker's story says he was subjected to it too. The documentary record does not support that. The biography excerpt read for this revision says that doctors at Camarillo State Hospital prescribed electroconvulsive therapy before his release in early 1947, and that people close to him intervened to stop it. One of them was Ross Russell, whose reliability as a source is contested. There is no record that Parker received it. A draft summary of this section said he did. The claim is excluded.

None of this is evidence about conservatories. Psychiatric hospitals held Powell, not music schools. The paper tells these stories to establish what "unpriced" meant in human terms for the people who generated the idiom: no institution existed whose business was to answer to them. The institutions that arrived later answered to students, and students paid. The people who made the vocabulary could not enroll.

VIII. The mechanism

The first and second rows carry most of the weight. An audience's power is that it can leave. A grade has no such discipline, because the evaluator's income does not depend on the student's eventual success in a market, and the institution's income depends on enrollment and retention, which a harsh and accurate verdict might reduce. Nobody has to act in bad faith for this to bend the signal over time. Incentives are enough.

The third row can be made concrete. A musician working three nights a week for forty-eight weeks receives about 144 market signals a year, or about 576 over four years. A student receives perhaps eight grades and a handful of juries over the same period. That is more than an order of magnitude fewer signals. The classroom signal can be richer per event, since a master's comment on a solo may carry more information than a night's tips. The claim is only that it is rarer and less coupled to price.

A rough way to state the difference, illustrative and not estimated:

learning rate ~ (signals per year) x (information per signal) x (fidelity to what an audience will pay for)

bandstand:  many   x  modest  x  high
classroom:  few    x  high    x  uncertain

The product is not obviously larger on either side. The Department's claim is that the third factor, fidelity to what an audience will pay for, is the one the academy has the least ability to supply, and that its absence is invisible from inside, because the proxies keep returning good news.

IX. The strongest objections

It does. The paper does not deny it, and Section III says so. But the most talented students are also the ones most likely to attend selective programs, so excellence among alumni does not separate what the program added from what the student brought. H1 is designed to get at the difference.

Agreed. The Department does not propose closing schools. It proposes adding the feedback that schools cannot supply by themselves: paid public performance, price discovery and ownership of what a night produces.

True, and it is a survivorship problem in the canon: the history of jazz we revere is the history of those who survived the market. The claim is about feedback mechanisms and not about outcomes. The market was also cruel (Section II), which is why the Department's program tries to keep its signal and not its predation.

It does: streaming, the pandemic, ticket prices, demographics, the decline of venues. The paper offers the correlation as a correlation. H3 proposes a test that could weaken the thesis.

Many now teach business, require public performance, and offer scholarships. That is good news, and it is testable (H1). If it holds, the paper's description applies to a shrinking share of programs.

It does, and Section XII discloses it. The remedy is the claim register in Apparatus B: any reader can check what the argument rests on.

The author studied at Juilliard. The critique is offered from inside, by someone who has seen the proxies work, and who has also seen what they leave out. If the argument is wrong, the data in Section XI will say so.

X. What the Department builds

The Department's four lines of work map onto the mechanism in Section VIII. Each one supplies a factor the classroom cannot.

is a filmed live series that prices presence over playback. It supplies the missing signal: a paying audience, a door, a night that happens once. gives artists and scientists time and money to study swing, improvisation and the economics of excellence without the dependency the paper describes: the fellow is answerable to the work and the audience, not to a grade. , including BassGPT and JazzGPT, are built as practice partners and never as replacements for the player. They raise the number of feedback events per hour of practice, which is the first factor in the learning-rate expression of Section VIII. They do not substitute for the audience, which supplies the third factor. teaches teams to decide under uncertainty, which is the transferable skill the bandstand has always taught and the classroom rarely tests.

The performance event

The model that bridges the thesis and the market is the performance event. The performance is the event, not the background: a company's launch, a patron salon or a festival night becomes an occasion in which the music is the reason the audience came. Each performance happens once, so it can be owned once. The night is captured and offered as a single authenticated edition, with provenance recording what it is and who holds it. Ownership opens a relationship between the buyer and the artist: private performances, commissions, partnerships. The revenue split is published in advance. Every band member is paid for the night and signs a release covering the recording and the edition.

The edition is described as a collectible and a relationship, never as an investment. The paper makes no statement about resale value, price growth or returns, and none should be inferred. Editions use original compositions or tunes in the U.S. public domain unless licenses are cleared, since recordings of other writers' songs and the video that accompanies them require licenses. Counsel reviews the edition terms and the platform before any sale. The Department does not offer fractional ownership.

Why now

Three conditions have changed. Philanthropy is increasingly recognizing jazz as intellectual work, and the 2026 MacArthur recognition of a jazz musician is good news for the field. A federal earnings-based accountability rule is described in the Department's brief as scheduled to begin affecting some low-earning college programs from the 2028-29 year. It bans no major, and its exact application to music programs should be checked against the final rule text before anyone relies on it (tier C). And artificial intelligence has made fluent output abundant, which raises the relative value of what cannot be automated: a human taking a risk in front of an audience that can leave. That last point is the author's position (P).

This paper asks for nothing. Where the Department raises money for any of these lines, it does so through separate vehicles with separate documents, and financial-instrument and token architecture described elsewhere plays no role in the claims made here.

XI. Predictions and a research agenda

The Department publishes the conditions under which this paper would be wrong.

Among jazz and music-performance programs, those that require paid public performance or business coursework should show a higher share of alumni income from performance five to ten years after graduation. the Strategic National Arts Alumni Project (SNAAP) survey data, College Scorecard field-of-study earnings. no difference, or the reverse.

A large share of jazz-studies alumni who work in music should derive their primary income from teaching positions, and those positions should be disproportionately at the kind of institutions that credential. SNAAP, NCES completions and faculty data. a majority of working alumni earning primarily from performance and sessions.

After controlling for population and urbanicity, regions with more jazz-degree completions per capita should not show higher live jazz attendance. NEA Survey of Public Participation in the Arts, NCES. a positive and robust relationship, which would suggest the academy builds audiences.

Nominal per-night pay for working jazz musicians in New York and comparable cities should have grown more slowly than inflation across 2016-2026. a Department survey, benchmarked against union scale where published. pay rising with or above inflation. The first version of this paper asserted a stagnation without data. This hypothesis replaces the assertion.

Students or fellows placed in a structured program of paid public performance during their first year should show faster growth on audience-retention measures (repeat attendance, rebooking) than a matched comparison group. the Department's own pilot, with pre-registered measures. no difference.

The Department commits to publishing results whichever way they fall. A confirmed hypothesis becomes a benchmark. A refuted hypothesis becomes a correction, and corrections in this series are public.

XII. Limits, conflicts and conduct

The Department builds products and programs in the space this paper describes. It benefits if the paper's argument is accepted. The claim register in Apparatus B exists so that a reader can check the argument without trusting the author.

The author studied at Juilliard, plays drums professionally in New York, and founded the Department. The critique is offered from inside the system it describes. That is a source of knowledge and also a source of bias, and the reader should weigh both.

The paper names no living musician, teacher or student for praise or criticism, beyond the standard credits that appear in the historical record. Statements about the dead (Powell, Parker, Freed) are limited to what the record supports.

The paper does not allege wrongdoing by any institution, school or person.

The evidence concerns the United States. Jazz education elsewhere is organized differently, and nothing here should be read as a claim about it.

Section XI.

XIII. Conclusion

Jazz was always taught. It was taught in orphanage bands, in private studios, in school ensembles, on the G.I. Bill, and above all on the bandstand, by working players to pupils who had to face an audience soon and often. The academy that arrived after 1945 widened access, preserved the tradition and gave serious musicians a livelihood. It also moved the principal source of a student's feedback from a room that could walk out to an institution that is paid to stay. Over time, a system built that way will tend to optimize students for its own proxies. That is a hypothesis about incentives, and it needs no villain.

The first version of this paper said more than it could show, and it was wrong in places. This version says less and can defend all of it. The indicators are consistent with the thesis and prove nothing. The mechanism is plausible and testable. The vocabulary of the tradition is free, and the transmission of it is not, and nothing routes the value of the second back to the source of the first.

The people who made the idiom included a man who wrote from a hospital bed that he was a composer, and a man who quoted Stravinsky on a bandstand so naturally that the composer was delighted. A country that asks where its Virgil is may find the answer on a shelf of records. The Department's position is that it is a better use of everyone's time to build the institution that answers to the lineage than to keep asking.

Performance is the asset. The recording is the archive. Price presence, not playback.

Apparatus A. Corrections to the first circulated version

Apparatus B. Claim register

Tiers: verified in this revision; standard record, not re-pulled; compiled from institutional pages, confirm before quoting; position. An asterisk marks a claim read this revision only through a secondary page that cites the primary source.

Apparatus C. Claims considered and excluded

These were in circulation around this paper and were deliberately left out. Listing them is part of the standard the Department holds itself to.

1. Prescribed but, by the available account, blocked. No record of receipt.
2. Published in in 2016 from a 1994 interview; his family disputes it, it rests on one journalist's account, and it describes Nixon-era policy (1968-1971), a generation after Powell's treatment. It belongs in a separate paper, if anywhere.
3. Harassment of Billie Holiday by narcotics agents is documented in outline. The motive claim is disputed by historians, including Lewis Porter.
4. No evidence. The claim is withdrawn, and mediocrity is defined narrowly in Section V.
5. The reported text is "these treatments."
6. Reported; accounts differ.
7. The transcript is paywalled.
8. Excluded by the Department's own claims rules.

Apparatus D. Notes and sources

1. Peter Pullman, . Hospital-record counts of ECT at Creedmoor.
2. Bud Powell, letter days before death (1966), as reported at https://en.wikipedia.org/wiki/Bud_Powell. Check the primary text.
3. Alfred Appel, (Knopf, 2002). Summary: https://openculture.com/2016/10/the-night-when-charlie-parker-played-for-igor-stravinsky-1951.html
4. Excerpt of (credited to Carl Woideck in the Department's search notes; confirm title and edition before citing); Ross Russell, (1973), whose reliability is contested.
5. Stanley Crouch, (2013), as reviewed in the , "Bird!", November 7, 2013: https://www.nybooks.com/articles/2013/11/07/bird/
6. National Endowment for the Arts, (August 2025): https://www.arts.gov/sites/default/files/SPPA_Comprehensive_Report_FINAL.pdf
7. U.S. Bureau of Labor Statistics, Occupational Outlook Handbook, Musicians and Singers: https://www.bls.gov/ooh/entertainment-and-sports/musicians-and-singers.htm
8. DataUSA, Jazz and Jazz Studies (NCES completions): https://datausa.io/profile/cip/jazz-jazz-studies
9. Johnathan Bi, interview with Brendan McCord on autonomy and AI, September 5, 2025: https://www.johnathanbi.com/p/transcript-for-interview-with-brendan-mccord-on-autonomy-and-ai
10. Department of Jazz, : https://departmentofjazz.substack.com/p/jazz-ascensionism-a-12-bar-blues
11. Department of Jazz, Papers 039 (Harmonic Intelligence) and 040 (Testimony).
12. MacArthur Foundation, Class of 2026: https://www.macfound.org/fellows/class-of-2026
13. Bryan Caplan, (Princeton University Press, 2018).
14. Paul Lopes, (Cambridge University Press, 2002); Scott DeVeaux, "Constructing the Jazz Tradition: Jazz Historiography," 25(3), 1991; Kenneth Prouty, (University Press of Mississippi, 2012).
15. C. A. E. Goodhart, "Problems of Monetary Management: The U.K. Experience" (Bank of England, 1975).
16. Institutional tuition and scholarship figures: the Department's research file, compiled from school web pages. Confirm on the current pages before quoting.

Department of Jazz - Paper 041 - The Dependency Model