The System Was Built For This

I. The Balance Wheel

Horace Mann is the patron saint of American public education and the most misread figure in its history. The mythology casts him as a democratic visionary who believed universal schooling would produce free citizens. The primary sources — his own twelve Annual Reports to the Massachusetts Board of Education, 1837 to 1848 — show something closer to a labor-discipline project dressed in Unitarian moral reform.

In 1843, at his own expense, Mann sailed to Europe to study how other countries ran their schools. He visited Britain, Ireland, and Scotland. The country that mattered was Prussia. His Seventh Annual Report, published on his return, was essentially an advertisement for the Prussian model. Prussia, he wrote, had “long enjoyed the most distinguished reputation for the excellence of its schools.”

That distinction was real. Prussia had, by 1843, the most politically instrumentalized school system in the Western world. Its origins trace to Frederick the Great’s 1763 Generallandschulreglement, which made primary education compulsory, tax-funded, and state-administered for children ages five to fourteen. The official purpose, in the historical record of the regulations themselves, was to instill loyalty to the Crown and to train young men for the military and the bureaucracy. Frederick wanted reliable soldiers, obedient bureaucrats, and disciplined subjects. The Prussian state got them. Mann bought the entire architecture — age-graded classrooms, teacher certification, compulsory attendance, a centralized state board — and brought it home to Massachusetts.

His American critics saw the move clearly and immediately. The Jacksonian Democrat Orestes Brownson, writing in the 1840s, warned that Mann was importing an authoritarian apparatus into a republic, that state-administered schooling under a central board would produce centralized political conformity rather than democratic citizens. Brownson was ignored. Massachusetts adopted the first compulsory attendance law in American history in 1852. By the end of the nineteenth century, thirty-four states had followed the Massachusetts template. The architecture was now national.

What the “balance wheel” sentence actually says

The canonical quotation from Mann’s Twelfth Annual Report (1848) is the “balance wheel” passage. Almost everyone who teaches it crops the sentence at the comma. The full line is more damning than the fragment:

Education then, beyond all other devices of human origin, is a great equalizer of the conditions of men, — the balance wheel of the social machinery.

Note the metaphor. Not a ladder. Not a lamp. A balance wheel — the mechanical governor on a steam engine, the device whose purpose is to prevent runaway acceleration and to keep the machine running at a stable rate. Mann’s vision of education was explicitly mechanical, explicitly stabilizing, and explicitly concerned with preventing the kind of social unrest that had produced the European revolutions of 1848, which were erupting the same year he wrote the report. The image he reached for to describe his life’s work was an industrial component. It was a part inside an engine. He did not pick it accidentally.

Mann elaborated. From the same report:

Are we not in danger of naturalizing and domesticating among ourselves those hideous evils which are always engendered between Capital and Labor, when all the capital is in the hands of one class, and all the labor is thrown upon another?

The common school was his answer. Not a democratic project — a prophylactic against class warfare. Mann was not hiding this. He was advertising it to a legislature full of textile-mill owners and railroad investors whose capital he needed to protect. His Fifth Annual Report (1842) made the economic case for the common school explicit: literate workers earned higher wages, exhibited greater reliability, and — the operative clause — were less likely to strike. Mann gathered testimonials from Lowell mill owners attesting that educated workers were more “docile” and “easily managed.” The reports were addressed to a Massachusetts Board of Education whose appropriations depended on the votes of the men who owned the mills.

This is not conspiracy theory. This is Mann’s own correspondence, in print, with the employers who funded his work. The common school movement was sold to capital as an investment in labor discipline. It was sold to workers as a path to self-improvement. Both pitches were true. That is the whole point of a balance wheel: it produces stability by counteracting force in two directions at once.

Lawrence Lessig’s framework for understanding power as a function of four modalities — law, norms, markets, and architecture — maps onto the nineteenth-century common school movement with almost embarrassing precision. The law was the compulsory attendance statute Massachusetts passed in 1852. The norms were the Protestant moral-reform vocabulary Mann himself supplied. The markets were the labor-discipline arguments addressed to mill owners. The architecture was the age-graded Quincy classroom that began the same year. Lessig’s typology is supposed to describe a mature system whose modalities have to be analytically separated to be seen. The common school movement is the historical case in which all four were deployed simultaneously, by the same people, in coordinated sequence, on a single decade’s timeline. It is the textbook example of coordinated modality deployment because Mann was running it as if Lessig had handed him the manual.

The wage discount that built the system

The common school system was built on cheap female labor. Catharine Beecher, beginning in the 1830s, campaigned to recruit women into teaching on explicitly economic grounds — arguing to school boards (not, crucially, to women) that female teachers could be paid roughly one-third of male wages while exercising the nurturing skills they would later use as wives and mothers. Before her campaign, elementary teaching was predominantly male. By 1870 the profession was majority female. By 1920 it was overwhelmingly so. The feminization of teaching made mass public education financially feasible and simultaneously locked the profession into a pay structure that has never recovered.

The gendered wage discount in American teaching is a design feature of the common school system, not an accident of the culture around it. Beecher said so. The school boards said so. The accountants who balanced the budgets said so. A profession that pays its workers one-third of comparable male wages is a profession that has been engineered to do so, and the engineering does not unbuild itself when the rationale becomes embarrassing. Teacher salaries in 2026, adjusted for inflation and education level, still sit roughly 26 percent below the salaries of comparably credentialed workers in other fields. The ratio is not new. It is the original architecture, depreciated.

Jürgen Habermas would recognize the move immediately. What Habermas calls the “colonization of the lifeworld” is the process by which spheres of life that were once organized through communicative deliberation — family, education, care — get reattached to the instrumental logic of money and bureaucratic power. The care and instruction of children is the paradigmatic lifeworld activity. Beecher’s campaign decoupled it from the household economy and reattached it to a wage-labor economy at a structurally discounted price. Pedagogy stopped being something a community deliberated about and started being something an institution executed. The colonization was so thorough and so early that by the time Habermas was writing in the 1980s, the original arrangement was already two generations in the rearview. Beecher had already won.

II. The Morrill Land

The Morrill Act of 1862 is routinely celebrated as the founding charter of democratic higher education in America. Its official title is more revealing than its reputation: An Act Donating Public Lands to the Several States and Territories which may provide Colleges for the Benefit of Agriculture and the Mechanic Arts.

Read the operative clause of the statute itself. The funds from the land sales were to endow at least one college in each state where, in the act’s own language, “the leading object shall be, without excluding other scientific and classical studies, and including military tactics, to teach such branches of learning as are related to agriculture and the mechanic arts, in such manner as the legislatures of the States may respectively prescribe, in order to promote the liberal and practical education of the industrial classes in the several pursuits and professions in life.”

Three things in that clause matter, and almost no one reads them out loud anymore. First: agriculture and the mechanic arts is the leading object. Classical studies are permitted but not required. The federal government was funding a specific kind of higher education — vocational, applied, and aimed at the labor needs of an industrializing nation. Second: including military tactics. Every land-grant college was required, by federal law, to teach military drill. This is why land-grant universities still host ROTC programs more than 160 years later. The Morrill Act was also a military manpower act, signed in the middle of the Civil War, by a Union government desperate for trained officers. Third: the industrial classes. Morrill said repeatedly in his 1858 speech defending the bill that the colleges were meant to be “accessible to all, but especially to the sons of toil” — not, crucially, daughters, and not, as the record would soon make clear, Black sons either.

Mark Blyth is useful here, briefly, as a way of seeing what the language is doing. Blyth’s contribution to political economy is the observation that ideas function as institutional weapons. The phrase “the industrial classes” is not a neutral description of an economic stratum. It is a political weapon. It tells the recipients of the federal land grant who they are educating and who they are not educating. It legitimates the resource transfer by attaching it to a category that does the political work of defining the deserving. The Latinist sons of Boston Brahmins did not need the Morrill Act. Harvard already existed. The act’s entire rhetorical purpose was to mark its beneficiaries as a different kind of citizen than the kind Harvard was already producing — a kind whose education would not threaten the existing prestige economy because the kind of education on offer was structurally different. The phrase did what it was supposed to do. We still use it.

The land was not unclaimed

The “land” in Land-Grant Act was not vacant terrain. It was Indigenous homeland. The 2020 High Country News investigation “Land-Grab Universities,” by Robert Lee and Tristan Ahtone, mapped the full scope of the transfer. The Morrill Act distributed 10.7 million acres taken from approximately 250 Tribal nations through more than 160 treaties and land cessions, many signed under coercion or outright military threat. These acres were sold to finance the endowment of fifty-two institutions, including Cornell, MIT, the University of California, and the University of Wisconsin. Cornell alone acquired title to nearly 990,000 acres across fifteen states, and remained for many decades the single largest beneficiary of the act in dollar terms. At least sixteen land-grant universities still earn revenue from the expropriated acres they retained — an estimated $8.7 million in fiscal year 2019 alone. The transfer is not a historical curiosity that closed in 1890. It is an active line item.

The democratic-flourishing mythology has to work very hard to avoid noticing this. The flagship state universities are taught to their alumni as monuments to the public spirit of the Republic. The financial infrastructure underneath them is a 160-year-old land seizure. Both can be true at once. They are. The mythology persists because keeping the two facts apart is easier than holding them in the same sentence.

The Second Morrill Act and the structural underfunding

Twenty-eight years after the first Morrill Act, the same senator returned to Congress to address what the first act had quietly permitted: the systematic exclusion of Black students from land-grant colleges throughout the South. The Second Morrill Act of 1890 required states to either admit Black students to existing land-grant colleges or establish separate institutions for them. Every former Confederate state chose the second option. Nineteen historically Black land-grant institutions — Alabama A&M, Prairie View A&M, Tuskegee’s federal designation, Langston, and others — trace their origin to this compromise.

The Second Morrill Act is where honesty matters. It is simultaneously the federal government’s explicit concession to segregation and the single largest expansion of Black higher education opportunity in the nineteenth century. Both are true. The 1890 HBCUs exist because the 1862 act was a whites-only program in everything but letter, and because Morrill and his allies were unwilling to force integration and chose instead to fund parallel institutions at a systematically lower level. The underfunding persists. A 2023 Century Foundation report documented that the 1890 institutions have been underfunded by their own states, relative to their 1862 counterparts, by roughly $12.6 billion cumulatively since 1987. The architecture that produced HBCUs also produced their structural poverty, and the structural poverty is not a residue of historical neglect. It is active state and federal policy that has never been corrected.

The Second Morrill Act is the legal template for “separate but unequal” in federally funded higher education, six years before Plessy v. Ferguson. Article 2 will trace what Plessy did to the K-12 system. The pipeline starts here, in the federal land-grant statute, with two of the same names on it.

III. The Factory

The physical and temporal architecture of American public education was not inherited from time immemorial. It was designed, on paper, by specific administrators with specific industrial analogies in mind. The age-graded classroom is the load-bearing example. It was invented in Boston in 1848 — the same year Mann published his Twelfth Annual Report — when the Quincy Grammar School opened under principal John D. Philbrick with twelve identical classrooms, each built to seat fifty-six students and one teacher, with students sorted by age into “grades.” The model spread to every large American city within two decades. Mann praised it explicitly because, in his words, it allowed for “the proper classification of the scholars.”

The source of the idea was the Prussian Volksschule, which had used age-graded classrooms since the early nineteenth century. Boston imported the floor plan whole. Before 1848, American schooling was done in one-room schoolhouses that mixed children of all ages, where older students helped teach younger ones. The Quincy model destroyed this. The children were sorted into batches by birth year, each batch assigned to a single adult who delivered standardized content for a standardized amount of time. If a child fell behind, they failed the grade and were held back. If they raced ahead, they waited.

The logic is industrial. It is batch processing. And it was adopted at precisely the moment American cities were becoming industrial batch-processors of cotton, steel, and immigrants. The architecture was not chosen because the educators of the 1850s discovered some pedagogical truth about age and learning. It was chosen because batch processing was the industrial logic of the surrounding economy, and the surrounding economy was who the educators worked for and whose children they were processing. Once the architecture was in place, the pedagogical justifications followed it. They have been written and rewritten ever since. The architecture is the binding constraint. The justifications are downstream of it.

It is worth dwelling for a moment on what the Quincy floor plan removed. The one-room schoolhouse was, by any modern measure, pedagogically primitive. It was also, structurally, a place where a fourteen-year-old who had grasped fractions could spend ten minutes teaching them to a nine-year-old, and where the same nine-year-old could overhear, on any given afternoon, the older students working through a passage of Cicero. The arrangement was not designed by anyone. It emerged from rural population density and the cost of building schoolhouses. It contained, by accident, several features that contemporary cognitive science has spent the last forty years rediscovering: peer instruction, mixed-age modeling, exposure to material above one’s assigned level, and the “productive struggle” that comes from watching someone slightly older than you work through a problem you cannot yet solve. Philbrick’s twelve-classroom design eliminated all of it. The age cohort became the unit of instruction. The teacher became the only adult in the room. The student became, as Cubberley would later put it, the raw product. None of this was a pedagogical advance. It was an administrative convenience that made batch processing possible at urban scale, and the educational research community has spent six generations trying to claw back, inside the architecture, the features the architecture removed.

Cubberley says the quiet part out loud

The most famous articulation of the industrial analogy is Ellwood P. Cubberley’s Public School Administration, published in 1916. Cubberley, dean of the Stanford School of Education, was the man who wrote the textbook that trained American school superintendents for the next two generations. The book was reprinted in 1922 and again in 1929. It was on the syllabus of every major school of education in the country. What it says about the purpose of schools is the operating manual of American public education in the period when the architecture calcified. Here is the sentence:

Our schools are, in a sense, factories, in which the raw products (children) are to be shaped and fashioned into products to meet the various demands of life. The specifications for manufacturing come from the demands of twentieth-century civilization, and it is the business of the school to build its pupils according to the specifications laid down.

This is not a hostile critic’s characterization of the system. This is the dean of Stanford’s education school, in his own textbook, praising the factory model as a goal to aspire to. The raw products are the children. The specifications come from the labor market. The business of the school is to build its pupils according to the specifications. The verb is “to build.” The metaphor is not metaphor. It is statement of intent.

Cubberley was working in the intellectual shadow of Frederick Winslow Taylor, whose Principles of Scientific Management (1911) had become the dominant framework for industrial administration. Taylor’s argument was that productive efficiency could be radically increased by breaking work into measurable tasks, timing the tasks, eliminating wasted motion, and rewarding workers for hitting standardized benchmarks. The textile mills loved it. The steel mills loved it. Cubberley’s project was to apply Taylorism to education: the standardized test as efficiency metric, the superintendent as plant manager, the teacher as line worker, the student as product. This is the period when the bell schedule, the fifty-minute period, the standardized test, and the age-graded cohort calcified into the architecture they still occupy a century later. None of it was inevitable. All of it was chosen.

There is a scholarly counterargument that deserves engagement. Beginning around 2015, a cluster of historians of education — Audrey Watters and Jennifer Borgioli Binis among them — began arguing that “the factory model of education” is a relatively recent phrase, traceable to Alvin Toffler’s Future Shock (1970), and that the historical parallel between schools and factories is overdrawn. Watters is right that the specific phrase “factory model” post-dates Toffler. She is wrong to conclude that this disposes of the analogy. Cubberley’s Public School Administration — the sentence quoted above, published in 1916 — uses the word factories explicitly. The analogy was not invented in 1970. It was recovered in 1970. It had been in the primary sources for fifty years.

Binis adds a useful nuance: the factory-model critique as commonly deployed focuses on what white middle-class children experienced and ignores the very different architectures imposed on Black, Indigenous, and Chinese children — Indian boarding schools, segregated Southern schools, Americanization campaigns aimed at Mexican children in California. Her critique sharpens rather than softens the main argument. The school system had multiple tracks from the start. The factory was for white workers; the boarding school was for Indigenous erasure; the Jim Crow school was for Black subordination. The machine had more than one setting. Article 2 will document the racial settings in detail. The point here is that the existence of additional racialized architectures does not vacate the factory analogy. It tells us the factory was one floor of a larger building.

Niklas Luhmann’s framework of binary codes is the diagnostic tool that lets us see what Cubberley’s factory was actually producing. Luhmann argues that mature social systems develop binary codes that allow them to process complex inputs at speed by reducing every decision to a yes/no operation. The legal system runs on legal/illegal. The economic system runs on payment/non-payment. The factory school began generating its own binary code somewhere between 1848 and 1918, and the code became formal doctrine in 1918, which is where we are headed next.

IV. The Track

The single most important period in the architectural history of American public education is 1910 to 1940, the years Claudia Goldin and Lawrence Katz have called the “human capital century” and the “high school movement.” During these three decades, the American high school graduation rate rose from roughly 9 percent to 51 percent. The United States built an entire tier of universal secondary education decades before any European nation attempted it. This is the core of the steelman case for the system, and any honest argument has to engage it. We will, in section VI. The point for now is that during the same decades the high school movement was creating universal access to secondary schooling, the architecture of sorting was being formalized inside it as doctrine.

The Cardinal Principles report

The critical document is the Cardinal Principles of Secondary Education, issued in 1918 by the Commission on the Reorganization of Secondary Education of the National Education Association. The commission chair was Clarence Kingsley. The report ran to thirty-two pages. It is, for its genre, remarkably explicit about its purposes. The report identified seven “main objectives” of secondary education. They are listed in the report’s table of contents in this order: Health; Command of fundamental processes; Worthy home membership; Vocation; Citizenship; Worthy use of leisure; and Ethical character.

Notice what is and is not on the list. “Command of fundamental processes” — the closest item on the list to academic learning — is one objective out of seven. It is listed second, after “Health.” “Vocation” is its own separate aim. So is “Worthy home membership,” which in the report’s own text explicitly meant training girls in household management. The 1918 Cardinal Principles are not a curriculum. They are a sorting schema. The report recommends that secondary schools offer “differentiated” curricula keyed to students’ future occupations. Some students would take the academic track. Others would take vocational, commercial, industrial, or home-economics tracks. The assignment to tracks was to be made by the school, on the basis of assessed “ability” and “probable destination.” The report does not contemplate the possibility that a student in the home-economics track might later want to study law.

This is the document that locked American secondary education into its tracked architecture. Herbert Kliebard’s definitive The Struggle for the American Curriculum, 1893–1958 frames the Cardinal Principles as the victory of the “social efficiency” faction over three rival camps: the humanists, who wanted classical academic curriculum for all; the developmentalists, who pursued child-centered pedagogy in the Pestalozzi-Dewey lineage; and the social meliorists, who wanted schools to teach critique of existing social arrangements. Kliebard’s judgment is that no faction won unconditionally, but the social efficiency victory in 1918 was the closest thing to a knockout blow in the entire history of American curriculum.

Luhmann’s binary codes apply to this moment with unusual clarity. The 1918 report formalized a code — academic / vocational — that operated with exactly the structural blindness Luhmann predicts of binary codes. A student coded “vocational” at age fourteen became, statistically, invisible to the academic track’s selection mechanisms. The code was self-reinforcing. The vocational track did not require Latin, so the vocational student did not learn Latin, so the vocational student could not pass the Latin examination that gated entry to college, so the vocational student remained, by definition, vocational. The code could not see itself producing the very outcome it claimed to be measuring. Luhmann’s point is that this is what binary codes do; they generate the future they claim to be merely describing. The Cardinal Principles is the historical case where the code was written down, signed, dated, and printed by the federal government.

Snedden, Dewey, and the debate that mattered

The intellectual foundation of the social efficiency victory was laid by David Snedden, a Massachusetts educator and Columbia Teachers College professor, and his student Charles Prosser. Snedden argued openly that most students derived little benefit from academic study and should be sorted into vocational training matched to their probable adult occupations. He accepted social stratification as inevitable and believed the school’s job was to make the stratification efficient. The phrase he used was “social efficiency.” What he meant by it was that schools should not waste resources educating students for positions they would never occupy.

John Dewey, writing in Democracy and Education (1916) and in a 1915 exchange with Snedden in the pages of The New Republic, named the problem precisely:

Any scheme of vocational education, which takes as its point of departure from the industrial regime that now exists, is likely to assume and perpetuate its divisions and weaknesses, and thus become an instrument in accomplishing the feudal dogma of social predestination.

The feudal dogma of social predestination. That is what Dewey called the Snedden program. It is hard to improve on the phrase. He saw it coming. He wrote it down. He addressed his arguments to the same readers Snedden was addressing, in the same magazines, on the same timeline, with more eloquence and a deeper philosophical apparatus. He lost.

Snedden won, and in the way that mattered most: through legislation. Charles Prosser, Snedden’s student and protege, drafted the Smith-Hughes Act of 1917, which established the federal-state funding structure for vocational education and — in the detail that mattered most — required that vocational programs be “administered and taught on a mass scale separately from general education.” The separation was not incidental. It was the whole point. A federal statute could have funded vocational training inside comprehensive high schools alongside academic instruction, with movement between tracks on the basis of student interest and aptitude over time. Prosser made sure it didn’t. The federal money was conditioned on physical and administrative separation. Once the federal funds were on the table, the schools that took them were structurally locked into the separation.

Smith-Hughes made the tracked high school a funded federal program. It is the act that put concrete and steel under the Cardinal Principles report. Dewey had already lost the philosophical argument by 1916. The political loss was sealed in 1917. By 1918, when the Cardinal Principles were issued, he was watching the architecture of his defeat being installed in real time, with federal subsidies, in the country whose democratic possibilities he had spent his life arguing for.

The sardonic register here is earned. Dewey saw it coming. He named it. He published the naming in the most influential book of educational philosophy in twentieth-century English, in the year the Smith-Hughes Act was being drafted in the basement of the Capitol building. The book is still read in every doctoral program in education. The architecture he warned against is still the architecture every American high school student moves through. The dignity of the loss is in the precision of the prediction. The horror of the loss is that the prediction was so precise that the people building the architecture could read it and keep building. It did not slow them down.

V. The People Who Believed It

One of the questions any honest reading of this history has to answer is: did the people running the system know what they were running? The answer, embarrassingly for the comfortable version of the story, is yes. Cubberley’s textbook was the training text for two generations of American school superintendents. The people who ran the machine were taught, as a foundational matter of professional identity, that their job was to sort children into life tracks matched to “the demands of twentieth-century civilization.” This is relevant to any later argument about whether the system was “captured” or “corrupted.” The system was not captured later. It was built this way. It was staffed by people who had been explicitly trained to operate it this way. The capture argument requires a baseline of innocent intent. The baseline is not there.

The ideology of social efficiency had a vocabulary. Its key terms were repeated, in slight variations, across the educational journals, the superintendent conferences, the teachers’ college curricula, and the federal commission reports of the 1900s through the 1930s. The terms were: scientific management, differentiation, probable destination, social utility, industrial classes, vocational guidance, ability grouping, standardized measurement. Each term did political work while sounding technical. Mark Blyth’s argument about ideas as institutional weapons applies here in its strongest form: this vocabulary was not invented to describe a pre-existing reality. It was invented to legitimate a sorting practice that needed legitimation, and the legitimation worked because the vocabulary sounded scientific in an era that had just decided to trust science for its industrial successes. The schools borrowed credibility from the steel mills.

The most consequential of these terms is probably standardized measurement. The intelligence test — in its early American forms, the Stanford-Binet adapted by Lewis Terman at Stanford in 1916, the Army Alpha and Beta tests administered to 1.7 million World War I draftees in 1917 and 1918, the National Intelligence Tests deployed in elementary schools through the 1920s — arrived in American education at exactly the moment the Cardinal Principles tracking schema needed an instrument for sorting children into tracks. The test was that instrument. It was promoted as objective. It was, in fact, deeply shaped by the cultural assumptions of its designers, who openly believed that intelligence varied by race and that the immigrant children flooding American cities were measurably less intelligent than the Anglo-Saxon stock. Terman wrote, in The Measurement of Intelligence (1916):

Their dullness seems to be racial, or at least inherent in the family stocks from which they come… Children of this group should be segregated into separate classes… They cannot master abstractions but they can often be made into efficient workers.

This is the founder of American academic intelligence testing, in the textbook that established the field, recommending that the children he had just defined as racially dull be sorted into separate classes and trained to be efficient workers. The test gave Cardinal Principles its scientific cover. The cover was political theology dressed in numbers. It worked because the numbers were unfamiliar enough to look authoritative and the political theology was familiar enough to feel obvious.

Michael Sandel’s argument about the moral corrosion of meritocracy starts here, even though Sandel himself dates the rise of the meritocratic ideal to the postwar Conant era. The deeper root is the moment when sorting was first presented as fair because it was “scientific.” That is the root of meritocratic moral corrosion. The sorting was unfair in 1918 in exactly the same structural way it would be unfair in 1968 and in 2026: it converted inherited and arbitrary distributions into the appearance of measured desert. Sandel’s contribution is the moral diagnosis that this conversion is corrosive even when the measurement is honest, because it tells the people at the bottom that they deserve to be there and tells the people at the top that they earned what was given. The 1918 report built the moral corrosion into the foundation. Conant put a more efficient measurement instrument on top of it in 1933. The hubris was structurally available from the day the architecture went up.

VI. The Strongest Defense

A series that refuses to state the counterargument fairly is not journalism. The strongest case for the system — the case that what was built in 1840 to 1920 was, on net, an extraordinary democratic achievement that the architecture critique fails to credit — deserves to be made in its strongest form before the article cuts through it.

Before Horace Mann, the overwhelming majority of American children received no formal schooling at all. In 1800, perhaps 10 percent of American children attended any school for any part of the year, and most of those schools were private, religious, or tuition-based, accessible only to families with money. By the time of Mann’s death in 1859, Massachusetts had achieved nearly universal primary school enrollment, and the common school idea had spread across the Northeast and Midwest. By 1900, the United States had the highest primary school enrollment rate in the world. That is not a small thing. It is a civilizational-scale expansion of literacy, and the architects who built it — Mann, Henry Barnard, Catharine Beecher, and the state legislatures who funded them — deserve historical credit for the expansion, whatever else they were also doing.

Lawrence Cremin, the most authoritative twentieth-century historian of American education, made this case in its fullest form. Cremin’s trilogy — American Education: The Colonial Experience (1970), The National Experience (1980), and The Metropolitan Experience (1988) — frames the common school movement as genuinely democratic in aspiration, even when compromised in execution. Cremin was not naive about Mann’s labor-discipline rhetoric. He read the same annual reports we just read. His argument was that Mann’s language of economic utility was tactical, necessary to win legislative appropriations from a textile-mill legislature, and that the deeper commitment was to universal literacy as a precondition of self-government. This is a defensible reading of the record. It requires holding two facts in tension: that the rhetorical architecture was labor discipline and that the practical effect was widespread literacy. Cremin held the tension and gave Mann the benefit of it.

The Goldin-Katz case for the high school movement is stronger still. In The Race Between Education and Technology (2008), Claudia Goldin and Lawrence Katz argue that the American invention of mass publicly funded secondary education between 1910 and 1940 was the single most consequential educational policy decision of the twentieth century. The economic returns to a high school diploma in 1914 were roughly 12 percent per year — an astonishing private rate of return. The high school movement produced a generation of white working-class Americans whose children went to college on the GI Bill and whose grandchildren entered the professional middle class. The compression of the American income distribution between 1945 and 1975 is in part an artifact of the high school movement. Goldin and Katz’s six “virtues” of the American system — decentralization, public funding, open access, forgiveness of early failure, gender equality, general rather than vocational emphasis — are real virtues. They describe a system that, for one specific cohort in one specific window, distributed the good of learning more broadly than any other country in the world.

Diane Ravitch’s The Troubled Crusade (1983) defends the post-1945 educational system in similar terms: whatever its failures, it delivered expansion of opportunity at a scale without historical precedent. The expansion was real. The Pell Grant generation was real. The community college network that opened in the 1960s was a public good of genuine and enduring value. None of this is in question.

The pattern that holds the steelman and the indictment together is that the mythology is not entirely false. The mythology is true for a specific cohort during a specific window, and is false as a general claim about what the architecture was built to do. The Goldin-Katz window is real. So is the architecture that preceded it, produced it, and defined who would be inside it and who would be outside it. Both facts have to coexist in the argument, and the honest move is the conjunction Sandel insists on: and. The high school movement was a real democratization, and it was built on top of an architecture whose explicit purpose was sorting. The democratization happened in spite of the sorting logic, not because of it, and it happened during an economic window — abundant industrial labor demand, constrained immigration, suppressed Black and women’s labor market access, Cold War federal investment — that no longer exists. When the window closed, the sorting logic remained. The mythology was generated inside the window and then assumed to be a description of the architecture, when it was actually a description of the window.

Article 3 in this series will trace what happened when the window closed. Article 2 will trace who was excluded from the window even when it was open. The job of Article 1 is to make legible that the architecture under the window was built for sorting, was built openly, and is still standing. The window is gone. The architecture is still here. That asymmetry is the entire reason this series exists.

VII. The Reader

One last thing has to be said before the sources. The reader who has gotten this far is reading these articles inside a curriculum platform that operates on a different theory of education than the one Cubberley built. The unit that contains this article is connected to other units across other courses by hyperlinks the reader has probably already noticed. The cross-curricular connections are the architecture of a different machine — one that does not assume the student should be sorted into a track at age fourteen, that does not treat “Command of fundamental processes” as one item among seven coequal aims, that does not borrow its scientific authority from the assembly line. None of these features scale through the factory architecture. None of them survive the tracked classroom. None of them happen on the bell schedule.

The reader will notice. The series does not need to belabor the contrast. What it needs to say, once, plainly, and only once: the architecture is not a fact of nature. It was a choice. It was made by named people with named arguments, and it has been remade, in smaller ways, by named people with named arguments at every subsequent moment. A different choice is structurally available. The platform from which these articles are published is one example of a different choice, operating at a small scale, in a regulatory and economic environment that is hostile to it. Other examples exist, in other places, on other scales. The mythology insists they don’t, because the mythology is what keeps the architecture in place.

What ends this article is not a policy proposal. It is the recovery of a sentence Dewey wrote in 1916, before he knew he had lost. The sentence is that any vocational scheme that takes its starting point from the existing industrial regime is likely to assume and perpetuate that regime’s divisions and weaknesses. He meant the schools. He could have meant any institution that mistakes its inherited architecture for the world. He named the dogma of social predestination at the moment it was being signed into federal law. He was right.

The architects of American public education were not villains. Mann was a genuine reformer who believed, sincerely, that universal literacy would be a public good. Morrill was an honest Vermont legislator who wanted his state’s farmers to have access to applied science. Cubberley was a man of his moment who borrowed the most prestigious analytical framework of his era — Taylorism — and applied it to the problem he had been trained to solve. Snedden believed, in good faith, that he was helping working-class children whom the academic curriculum was failing. None of these people set out to build the feudal dogma of social predestination. Some of them, asked plainly, would have rejected the description with horror. The horror is that the architecture they built does not require them to have intended it. The architecture works on its own terms. It produced, and continues to produce, the outcomes its design specifies. Intent does not matter at the structural level. The system was built for this. The sources are below. The receipts are public. The question is what we are willing to read.