As the storm around William Deresiewicz’s recent New Republic cover story illustrates, American higher education is currently coming in for more fierce discussion and criticism than at any time since the 1960’s. Is it a breeding ground for “excellent sheep,” as Deresiewicz would have it? A set of rigid and antiquated structures unsuited for the digital age, as Mark Taylor has argued? An ivy-covered embodiment of growing economic inequality, as Andrew Rossi’s much-praised film “Ivory Tower” charges? Or perhaps, still others suggest, all these questions are moot, because the entire system is on the verge of collapse, thanks to the internet, and the spread of free, online courses. The frenzy around these courses, called MOOCs, has cooled somewhat over the past couple of years, but the question is still being asked: “Are universities going the way of record labels?”

In fact, the answer to this last question is clearly “no.” The profusion of online courses is already changing the way the universities operate. But it is not going to destroy them. And the reasons why it will not do so point to a basic fact about higher education that has too often been neglected in the current debates. Universities do not just function as providers of contents and services. They are not just a sector of the economy. They are social institutions in the fullest sense of the word, deeply embedded in the American social structure.

If we examine the record label analogy closely, it shows why universities, far from being on the brink of a radical transformation, will remain, for better and for worse, remarkably difficult to change. To start with, serious learning requires a great deal more effort than listening to music. If mastering complex and difficult subjects were simply a matter of access to “content,” as many of the universities’ doomsayers seem to assume, then it would not require universities at all, online or in person. Books would suffice. To be sure, online courses can mimic something of the structure and routine a physical course provides. But the very flexibility that online learning offers works against them in this respect. It is shocking, I know, but students who don’t have hard and fixed deadlines for assignments tend to procrastinate, often indefinitely. Studies have shown that completion rates for MOOCs can fall as low as 7 percent. A university with a 7 percent graduation rate wouldn’t stay in business for very long.

The same flexibility hurts online courses in another way. Yes, with a MOOC, you can watch a lecture, in your pajamas, at 2:30 in the morning. But what if you don’t understand one of the lecturer’s points? At 2:30 in the morning there is unlikely to be anyone around, even by chat, to answer your question. Defenders of brick-and-mortar campuses often tend to extol the ineffable value of free-floating seminar discussions. But just as important, in truth, is the simple ability to have an expert in the subject explain a difficult point at the moment confusion arises. Similarly, you may be able to listen to a chemistry or psychology lecture in your pajamas at 2:30 in the morning, but doing the accompanying lab will be a bit more difficult. Getting the pronunciation of a foreign language right is a lot trickier if no one hears the mistakes you are making, and corrects them in real time.

In my own experience, by far the most successful online courses are the ones that mimic as closely as possible the structure of an in-person course. They have fixed meeting times for questions and discussions, and fixed deadlines for assignments. Evaluation involves more than simply the computer grading of multiple-choice questions. But, not surprisingly, courses of this sort, which require making an instructor available at particular times to give personalized attention to students, end up costing an institution almost as much as an in-person course—and this in turn requires serious tuition payments. In short, even in an all-digital future, higher education would still require institutions recognizable as universities.

Despite elite universities’ well-deserved reputation as hotbeds of liberalism, they have not come close to eliminating one of their most blatantly inegalitarian practices: preferences for alumni children in admissions.