This post is the text of a keynote address I gave on July 14 (Bastille Day!) at the Building Blackboard Together 2026 conference in Dallas. This is an annual event hosted by Blackboard — the learning management systems (LMS) company — which brings together the Blackboard staff and the system’s users at colleges and universities to talk about new developments in the company’s LMS. The Blackboard system is used at 980 institutions, primarily US colleges and universities. 

Here’s a summary of the text from Copilot, which ain’t bad:

The document argues that anyone attempting to reform or build technology for American education must first understand its deep structural complexity. The author warns that reformers often “fix one piece of the system without considering how this would play out in the other parts,” leading to harmful unintended consequences. The U.S. education system is described as an emergent institution—never centrally designed, but evolved through decentralized trial and error. Its core functions are inherently contradictory: democratic equality, social efficiency, and social mobility. These competing aims create tensions that shape schooling and make coherent reform nearly impossible.

A major consequence is credentialism: students pursue grades, credits, and degrees rather than learning, especially as families compete in a zero‑sum mobility race. Technology, including AI, amplifies this dynamic by enabling students to bypass learning entirely. The system is also loosely coupled, meaning change rarely spreads effectively, and institutional autonomy—especially in higher education—drives both excellence and resistance to reform. Attempts to impose efficiency often undermine genuine learning, which requires time, struggle, and deep engagement. The author concludes that educational technology should aim to support real learning rather than “move fast and break things.”

See what you think.

Know It Before Youth Mess with It:

How the American System of Education Works

I’m a sociologist not a philosopher, so I’m not going to explore what education should be. Instead I’m going to examine how the US system of education actually works. Such an understanding is particularly important for people who are working in the field of educational technology. After teaching about the history of school reform for many years, I learned that you need to figure out the fiendishly complex education system before you try to intervene in it. Unintended consequences are the norm in the school reform business – and most of those consequences caused educational harm.

My aim here is to explore the nature of the American system of education. The focus on the interior logic, core tensions, and major implications of this system. My students at the Stanford Graduate School of Education were mostly made up of former teachers who were returning to the university for advanced degrees. After years in the classroom, they wanted to build on their experience to address what they saw as major problems with the structure of schooling. They were teachers who wanted to become school reformers.

My message to them was to step back from this mission, take a breath, and spend some serious time exploring how this system they wanted to change actually functions. Past reform efforts usually made things worse because they were focused on fixing one piece of the system without considering how this would play out in the other parts of the system. Best to make a close study of the larger structure of schooling before you start tinkering. And it would be prudent to start small before going to scale and to introduce innovations that could be rolled back quickly if they turned out to be doing more harm than good. This would be my advice to ed tech entrepreneurs as well.

Of course, no serious reformer would even think of following this advice. New mayors and governors and superintendents and college presidents know full well that they only have a few years to make an impact on the system. As a result, they’re eager to think big, work fast, scale up, and make sure the changes will be hard for successors to reverse. This is why Dick Elmore and Milbrey McLaughlin called school reform “steady work.”[1] Reformers make work for the next wave of reform to fix what they have messed up.

Impulsive action doesn’t pay off in the education world. And a central reason for this is that anyone seeking to have an impact on education needs to develop a certain respect for this institution. The American system of education is an amazing institutional success story. It started in the early 19^th century as a modest and marginal enterprise, confined largely to young children and operated at private expense. But it rapidly rose to become today’s educational behemoth, which consumes a third of every state’s budget, the bulk of local property taxes, and a giant share of the lives of citizens. Think of how many years you spent as a student. In my case, it was an astonishing 27 years. That’s 2 years preschool, 12 in school, 5 in college, and 8 in graduate school; more than a third of my life. As I used to ask my grad students, most of whom were in their late 20s, “When are you going to grow up?” I never did. I stayed on as a professor.

Now consider the higher ed component of the system. The broader history of the university in Western Europe is as deep as it gets. Clark Kerr pointed out that

About eighty-five institutions in the Western world established by 1520 still exist in recognizable forms, with similar functions and with unbroken histories, including the Catholic church, the Parliaments of the Isle of Man, of Iceland, and of Great Britain, several Swiss cantons, and seventy universities. Kings that rule, feudal lords with vassals, and guilds with monopolies are all gone. These seventy universities, however, are still in the same locations with some of the same buildings, with professors and students doing much the same things, and with governance carried on in much the same ways.”^[2]

Set against this long history of the university, the story of the US higher ed system is particularly stunning. From the perspective of 19^th century visitors to the US, the system of higher education was a joke. It wasn’t even a system, just a random assortment of institutions claiming to be colleges that were scattered around the countryside. Overbuilt, underfunded, academically weak, located in small towns along the frontier, and lacking in compelling social function, the system seemed destined for deserved obscurity. But by the second half of the 20th century, the system had assumed a dominant position in the world market for higher education. Compared with peer institutions in other countries, it came to accumulate greater wealth, produce more scholarship, win more Nobel prizes, attract a larger proportion of talented students and faculty, and dominate global rankings. A system like this has earned our respect. Mess with it at your peril.

An Emergent System

First, American education is a system without a plan. No one designed it. No single agency governs it. Control of the K-12 system is dispersed across 50 states and 13,000 individual school districts. And higher education is even more disaggregated, with 4,000 degree-granting institutions, both public and private, operating with a striking degree of institutional autonomy.

Instead of being created from above, the system emerged from below, within the fertile political and social ecology of the early American republic. The education system evolved in much the same way as a biological system, through trial and error in a competitive environment until it developed an institutional model that proved sustainable in the long run. The higher ed system was even more chaotic, since colleges first developed as a series of private enterprises, with the state entering the scene only in the mid-19^th century. Accordingly, this was a messy process littered with failures. The result is a system that is both incredibly inefficient and remarkably effective, a theme I will return to later, which has profound implications for producers and users of educational technology.

Emergent systems, whether biological or social, share certain characteristics. They’re resilient and adaptable, open to continuing evolution but resistant to radical change. However, social systems have an advantage over biological systems that work through genetic inheritance. Lots of biological creatures developed mechanisms for seeing and hearing on their own through independent evolutionary processes. They may look similar, because they serve parallel functions, but each had to develop on its own. Social systems, however, can imitate each other, so innovations can spread laterally with remarkable speed. As a result, we see the paradox of an American education system that has no centralized control but shows a remarkable degree of similarity across widely dispersed institutions. Walk into a school or college anywhere in the country and you will find self-contained settings for teaching and learning that simply scream “classroom.”

Emergent systems don’t arise from ideals but in response to needs. Biological systems come into existence in order to enhance the survival of the species. For social systems the need is the survival of culture and society. That means the education system needs to provide people with the capacities and skills and the values and norms that they need to sustain the social order, while also giving individuals the wherewithal to survive and thrive within that society.

A System Built on Competing Goals

So, what are the key functions that we are asking school to carry out? I suggest three broad social functions that have driven the evolution of the American education system: democratic equality, social efficiency, and social mobility.  

Democratic equality was the rationale for constructing the common school system in the early days of the American republic. The history of republics had shown that they tend to collapse into tyrannies unless that can maintain civic virtue in their citizens – a commitment to the common good. In the early 19^th century US, a burgeoning market economy threatened to split the newly formed republican community into competing interest groups fighting over personal advantage. Founders of the common school saw the answer in bringing all the young people in town together in an educational setting where they could develop shared values and undergo a common set of learning experiences.

A second goal for schooling – social efficiency – emerged at the end of the century, when the rise of white-collar work meant that education became a useful mechanism for providing the cognitive skills needed by the new corporate, professional, and government bureaucracies. Schooling for social efficiency is what today we call the production of human capital, and it has become the rationale for schooling that is most prominent in the rhetoric of educational policymakers.

Then, in the 20^th century, the system developed a third function, social mobility. Employers increasingly based hiring decisions on the quantity and quality of education attained by applicants. This meant that families needed to focus their energies on getting the best education possible for their children in order to enable them to qualify for the best paying jobs. Whether the aim was to move up the social ladder or to preserve a privileged rung on the ladder, the mandate was the same: to get a good job you needed a good education.

Democratic equality and social efficiency both see education as a public good. We all benefit by having fellow citizens who are law abiding and civically engaged, and we all also benefit from having fellow workers who are productive at their jobs and thus promote economic growth. But social mobility construes education as a private good. The aim is for me to get a better education than you, so I get the good job instead of you. Education as a public good is a win-win game in which we all benefit from the education of each other’s children; but education as a private good is a zero-sum game. My win is your loss.

All three of these goals are still present in the educational system today. We see democratic equality in the shared experience of students, especially in elementary school. We see social efficiency and social mobility in the increasingly stratified educational system that students encounter progressively in middle school, high school, college, and graduate school – in which they prepare for radically different places in the hierarchical job structure.

One value in looking at this mix of goals for education is that it reveals the dynamic contradictions that are built into our system of schooling. We want schools to promote both equality and inequality, getting ahead and staying ahead, commonality and stratification. These ongoing tensions are at the core of the system, a feature not a bug.

Over the course of the 20^th century, especially in the postwar period, the social mobility goal has grown increasingly dominant within the system. As the system gradually expanded upward – from commonality in the elementary level, to curriculum tracking in the secondary level, to extreme stratification in the tertiary level – the differential benefits provided by gaining more and better education than your peers began to displace the concern for providing a good education for everyone. The incentive for students was to get a distinctly uncommon education. The prime directive? To gain what Thorstein Veblen called “invidious distinction.”

These competing goals for education are obviously at odds with each other. That is a core reason for the immense complexity of the system. There is no unified goal for education, and the tensions inherent in the system are unresolvable. In some ways, we all want education to accomplish all of these goals. We want other people’s children to get a good education but just not at the expense of our own children. We want to be surrounded by good citizens and productive workers; but when push comes to shove, we want more for me and mine than for you and yours.

Any effort to intervene in education, whether through technological innovation or some other reform effort, has to take into account how the new intervention is going to play out in the ongoing struggle among the system’s competing goals. Good luck trying to make such a prediction.

A System that Promotes Credentialing over Learning

A key impact that social mobility has had on the education system is the way it shapes motivation. The fervor with which students and families approach the education system is a logical outcome of the structure of the modern world. In premodern times, you inherited your job from your parents. Parents passed the farm or the shop or the craft on to their kids. Some social mobility was possible through the military or the church, but that was the exception. In the modern world, parents don’t – indeed can’t – pass their positions on to their children. Instead, individuals need to demonstrate their individual merit through performance in school, and jobs are allocated according to educational credentials.

The good news is that this potentially opens up mobility to everyone. But the modern structure of status attainment also means that parents in an advantaged position can’t simply pass that advantage along to the next generation. Doctors and lawyers cannot make their children doctors and lawyers. Their kids need to attain those positions the same way their parents did, by doing well in school. Thus the panic that upper-middle-class families experience in trying to usher their children successfully through the educational system. They need to buy a house in the right school district, get their kids into the right school, into the gifted and talented program in elementary school, and into the AP track in high school.

Then comes the big magilla: negotiating the complex hierarchy of higher education. Families feel it’s crucial to gain admission for their children to the most selective college that will admit them, and then the most selective graduate school. They have lots of advantages in pursuing this route – economic capital, cultural capital, and social capital. But there’s still a chance they their children will lose out in the zero-sum competition of college admissions to some smart poor kid. This leads to behaviors that are frantic but not crazy. Parents know that downward mobility is a real possibility – just not on their watch.

Think about how this pressure shapes the behavior of students as they work their way through the educational system. After all, if school is all about getting ahead, then you need to focus on accumulating the merit badges that will allow you to accomplish this goal. In the system, these badges consist of grades, credits, and degrees. Your aim is to get the best grades, accumulate the most course credits, and snag the highest and most prestigious degrees.

The system sees these badges as signals of merit for individuals who attain particular degrees of learning. But from the student’s perspective, it’s the badges that are the target not the learning. The aim is to get as many badges as possible and to do so as efficiently as possible. You’re better off if you can accomplish this with a minimum of effort – which means with as little learning as you can manage. The pursuit of education thus turns into the pursuit of credentials. Any way you can game the system is worth giving a shot. Example: 40% of Stanford undergrads have sought and received certification for a disability in order to get more time on tests, deadline extensions, or a private dorm room. The point is to get an edge on the competition.

Thus the education system finds itself confronting the cold reality posed by Goodhart’s Law. This law states that when a measure becomes a target it is no longer a valid measure. When grades, credits, and degrees become the target for students, they are no longer valid measures of academic learning. The key skill required for doing school is to game the system.

And technology obviously plays a key role in this process of pursuing credentials on the cheap. Why read the book when AI can summarize it for you? Why write the paper when AI can write it for you? If you can buy your education at a discount, you’d be a sucker to pay full price. And AI provides the ideal tool for gaming the credentialing system. Why not have an AI agent sign into your Blackboard or Canvas account for a class, do all the assigned reading, watch the lectures, and then submit your papers?

A Loosely-Coupled System

 Another saliant trait of American system education is that it’s loosely coupled. What does that mean? As Karl Weick has pointed out, organizational systems can be categorized on a spectrum from tightly coupled to loosely coupled.[3] In a tightly coupled system, all parts of the organization are so closely connected that a change in one part of the system can rapidly spread to other parts of the system. A good example is a nuclear power plant. Charles Perrow did a classic organizational study of the Three-Mile Island nuclear disaster in Harrisburg, PA. He noted that when a problem arose at one place in the plant, say overheating, a safety mechanism shut down that process, which then triggered a reaction elsewhere, which set off another safety mechanism. Ironically, the safety systems almost brought about a meltdown of the plant. The title of his book is telling: Normal Accidents.[4] His point is that in a tightly coupled system like this, an accident of the sort that developed was to be expected at some point. You wouldn’t know when and where it would happen, but tightly coupled systems are susceptible to catastrophic failure.

By contrast, in a loosely-coupled system the elements are partially segregated from each other. Consider the US school system. It has six levels nested in each other: federal, state, district, school, classroom, and student. Each is partially autonomous, buffered from the level above and also buffered from other units at the same level – physically separated and operating independently. As result, it’s not easy to control it from above, and it’s also not easy to disseminate change across units at the same level.

Higher education is even more loosely-coupled than schools. The 50 states have separate systems of public colleges and universities, and individual institutions retain considerable autonomy compared to universities elsewhere in the world. In addition, of course, the huge array of private colleges and universities operate largely independent of the state and of each other.

The result is that the loosely-coupled educational system is both hard to improve and hard to ruin. It’s quite effective at resisting the spread of both good and bad innovations.

A System Grounded in Institutional Autonomy

Institutional autonomy is a core characteristic of the American system of higher education. In part this derives from the prominent role that private institutions play in the system. One of the odd aspects of the history of American higher ed is that private colleges preceded public. Another, which in part follows from the first, is that private colleges are also often more prestigious. Nearly everywhere else in the world, state-supported and governed universities occupy the pinnacle of the national system, while private institutions play a small and subordinate role, supplying degrees of less distinction and serving students of less ability. But in the U.S., the top private universities produce more research, gain more academic citations, attract stronger faculty and students, and graduate more leaders of industry, government, and the professions. According to the 2025 Shanghai rankings, 8 of the top 10 universities in the world are American, and 7 of these are private.[5]  

When states began opening public universities later in the 19^th century, the new institutions modeled themselves after the existing private institutions. In part this was because it was a proven model for how to succeed in the American context, and in part because states were not able to fund them adequately. State funds came only through occasional state appropriations; regular annual funding didn’t start happening until the early 20^th century. So, like their private peers, they needed to rely on attracting and retaining tuition-paying students and soliciting donations.

After the Civil War, American colleges began to develop a variety of mechanisms for bonding students and alumni to the alma mater, through fraternities and sororities and also through intercollegiate athletics, especially football. Thus emerged the advantaged position that American public and private colleges enjoy today, in which college is not just the place you go to school, it’s who you are. Foreign visitors are stunned to see how many people in the US, on campus and off, wear clothing bearing the college insignia. Bonded students become loyal alumni who then become generous donors. This provides a strong basis for institutional autonomy.

And autonomy is the secret sauce that makes American higher education so successful. The U.S. enjoys a glut of higher education institutions, which means they have to compete in a crowded market for students, faculty, grants, donations, and forms of distinction. This is not just an issue for the private sector. Public institutions remain heavily dependent on nonpublic sources of income. On average, public colleges and universities receive only 40% of their income from the state.[6] At public research universities, this proportion can fall as low as 5%. With funding comes control.

The bad news is that American institutions suffer from less government funding than their counterparts around the world. The good news is that this leaves them free from the high degree of state control that foreign universities experience. This relative autonomy makes them nimbler in pursuing new academic opportunities, developing new programs, and competing effectively in the evolving higher education marketplace. Not being as compelled to work through many layers of bureaucracy or to please political masters, administrators and professors have more freedom to do what they want to advance both the interests of the institution and their own careers.

A study of universities in Europe and the U.S. by Caroline Hoxby and associates shows that this mix of institutional autonomy and competition is strongly associated with higher rankings in the world ecology of higher education. They found that a one percent increase in the proportion of the university budget coming from state funding brings about a decline of three ranks in the Shanghai Index.^[7] In other words, autonomy pays high intellectual and organizational dividends for universities. Less state funding grants more autonomy and greater institutional success.

And the loosely-coupled nature of individual US universities means this autonomy pervades at all levels within each institution – where colleges, schools, departments, centers, and institutes operate as semi-independent enterprises, pursuing their own interests in the competition for funding and academic distinction. It even reaches the level of individual faculty members, who can conceive of themselves as independent entrepreneurs, using the university as the launching pad for their own careers.

A System in Which Efficiency Undermines Effectiveness

One characteristic of emergent systems and loosely-coupled systems, such as the American system of education, is that they are inefficient. This inefficiency drives policymakers and reformers and business executives crazy. Why can’t schools and colleges be operated like a business? Why tolerate all of its internal contradictions, autonomous units, and duplication? As MBA students in my classes at Stanford used to ask, “Why doesn’t the superintendent of schools or the college president just tell people what to do?” The answer, I want to suggest here, is that in education, efficiency is the enemy of effectiveness. Consider what’s been happening to K-12 education in the last 50 years.

Since the 1983, when the federal report A Nation at Risk^[8] crash-landed into American classrooms, US educational policy had been driven by the accountability movement. The idea is that schools should be held accountable for meeting curriculum standards by means of statewide multiple-choice tests to determine the level of student achievement in individual school subjects. The aim of accountability is to wring the chronic inefficiencies out of a structure of schooling that arose from its radical decentralization. This means trying to turn the longstanding loosely-coupled system of education into a more tightly-coupled system. Reformers understand how often past efforts at school reform bounced off the classroom door, buffering the process of teaching and learning within from outside influence.

However, this local autonomy, which makes education annoyingly inefficient in the eyes of policymakers, is essential in the effort to make education effective. Teaching is not a delivery system for academic content but a fiendishly complex form of professional practice that seeks to induce students to learn in the absence of any efficient mechanism for insuring that they will do so. Students only learn when and what they choose to learn. The core skill of a good teacher is the ability to lure students into making the choice the teacher is aiming for. And this means that teachers need to have the flexibility to adapt their teaching approaches to the peculiarities of the group of students they find before them and also to the differences in individual students in the class. 

And now we see that the efficiency police are coming after American higher education as well, thus threatening to undermine the most effective system of education that the world has ever seen. Of course, they are seeking to enlist educational technology in this effort.

The charges against higher ed are, to say the least, capacious. College students aren’t learning. They aren’t reading. They’re not being prepared for the workforce. They accumulate too much debt. They’re majoring in subjects that don’t pay off in the marketplace. Colleges inflate their costs with administrative overhead. Professors spend too little time teaching. And on and on and on.

But ed tech can come to the rescue. Western Governors University and University of Maine, Presque Isle, have instituted online degree programs that are based on competency tests. As soon as you can pass the test for a class, you collect the college credit and move on to the next class. The Washington Post reported that one North Carolina woman completed a BA degree at Presque Isle in 3 months, while working full time, and then went on to win an MA in just five weeks. Total cost for two degrees: $4,000. Now that’s efficiency.

On the one hand, this competency-based approach seems a natural answer to the problem outlined earlier, namely the way credentialing has displaced learning as the prime function of education. Students here have to demonstrate what they actually learned in order to gain credit for a class and move toward a degree.

But this is a remarkably narrow vision of learning, construing it as the accumulation of knowledge about a subject. Knowledge is important. However, the key is not what you know but what you can do with this knowledge. You need to be able to integrate it with what else you know, play it off against alternative conceptions, critique it, and deploy it in a variety of situations. This requires deep engagement with the subject and the development of your understanding by working out your thoughts in writing.

This takes time. An efficient education, like the Presque Isle approach, is fast food, but a real education is slow food. You can’t fast-forward through the process in three months and claim that you’re educated. Learning stuff is hard. If you try to make it easy, you’re shortcutting the cognitive work that real learning entails.

Take writing for example. Writing is not how you put ideas down on paper. It’s how you develop ideas that are too complex to work out in your head. If writing feels easy, it means you’re being glib rather than thoughtful. And take reading as another example. College is a place where we ask students to struggle with difficult texts that explore complex ideas, tasks that they would be unlikely to do on their own. Otherwise we wouldn’t need colleges at all. You could just educate yourself at the library.

However, in practice today at American universities, students are able to use educational technology to bypass the process of learning entirely: They can have AI do the reading and writing for them.

Conclusion

So where does that leave us? My point is that educational technology – like any other mechanism for tinkering with the American system of education – needs to develop a rich understanding of how this system works before messing with it. This is an enormously complex, highly resilient, and remarkably successful institution. As I’ve pointed out, it is a system that is emergent, contradictory to the core, caught up in credentialism, loosely coupled, grounded in institutional autonomy, and necessarily inefficient.

All of these components of the system increase its complexity but should be understood as strengths. This is the system you are working with. You should be trying – unlike so many reformers before you – at least to do no harm. You need to work with its complex dynamics, so you can enhance its promise rather than undermining its effectiveness.

What students need, I suggest, is a different form of educational technology than the current AI model. One that both supports them and challenges them as they work through the difficult process of becoming an educated person. This is not a setting where it makes sense to follow the longstanding Silicon Valley motto: “Move fast and break things.”

[1] Elmore and McLaughlin (1988).

[2] Kerr (2001), p. 115.

[3] Weick, 1976.

[4] Perrow (1986).

[5] Institute of Higher Education (2025).

[6] National Center for Education Statistics (2023).

[7] Aghion et al. (2010), p. 12.

[8] National Commission on Excellence in Education, 1983.

References

Aghion, Philippe, Mathias Dewatripont, Caroline Hoxby, Andreu Mas-Colell, & Andru Sapir. (2010). The Governance and Performance of Universities: Evidence from Europe and the US. Economic Policy 25: 61, pp. 7–59.

Elmore, Richard F., & McLaughlin, Milbrey W. (1988). Steady work. Santa Monica, CA: Rand.

Institute of Higher Education, Shanghai Jiao Tong University. (2025). Academic ranking of world universities – 2025. https://www.shanghairanking.com/rankings/arwu/2025.

Kerr, Clark. (2001). The uses of the university, 5^th ed. Cambridge: Harvard University Press.

Labaree, David F. (1997). How to Succeed in School Without Really Learning: The Credentials Race in American Education. New Haven: Yale University Press.

National Center for Education Statistics. (2023). Postsecondary Institute Revenues. https://nces.ed.gov/programs/coe/indicator/cud/postsecondary-institution-revenue.

Perrow, Charles. (1986). Normal accidents: Living with high-risk technologies. New York: Basic Books.

Weick, Karl. (1976). Educational organizations as loosely-coupled systems. Administrative Science Quarterly, 21, 1-19.