You're in your supervisor's office, second year, and you've just cited a theorist whose name clearly does not belong in this room. The frown that crosses his face isn't hostile. It's almost paternal, the frown of a man who has watched students make this particular mistake before and has decided, generously, to save you from yourself. It says: that's not really how we do things here. And you, wanting to finish, wanting the approval that signs off your chapters, quietly adjust.

That moment is the whole story, compressed.

The Apprenticeship Model Was Never Neutral

Doctoral training is structurally an apprenticeship. One senior scholar, credentialed and tenured, takes on one junior scholar and transmits not just methods but taste, priorities, and a mental map of what questions are worth asking. The model is ancient and, in many disciplines, genuinely effective at producing rigorous specialists. It is also, almost by design, a replication machine.

Consider what the supervisor controls: reading list construction, feedback on draft arguments, introductions to journal editors and conference organizers, and the single most consequential document in a young academic's career (the reference letter). That concentration of influence in one relationship is extraordinary by the standards of almost any other professional training. A medical resident rotates through departments. A barrister is trained by the bar, not by one silk. The doctoral student has, in many programs, essentially one intellectual parent for four to seven years.

The result is predictable. A study of faculty hiring in North American universities found that a small number of elite programs produced a disproportionate share of tenure-track hires, and within those programs, a handful of supervisors produced a disproportionate share of the hires from those programs. Intellectual lineage, in other words, compounds like interest. The grandchildren of a few prestigious advisors end up running departments their grandparents shaped.

How the Hierarchy Gets Transmitted, Practically

The mechanism isn't conspiracy. It's something more mundane and therefore harder to resist: the normal operation of academic taste.

Two students, call them Priya and Marcus, both begin doctoral programs in political theory at comparable universities in the same year. Priya's supervisor trained under a scholar whose intellectual formation was mid-century Anglo-American analytic philosophy. Marcus's supervisor came up through a European tradition of critical theory. Both students are talented. Both work hard. But Priya will spend four years learning to treat certain continental thinkers as decorative where they might otherwise be foundational, and Marcus will spend four years learning to treat analytic precision as a kind of intellectual poverty. Each will write a dissertation that reflects those priorities. Each will cite the scholars their supervisor's network respects. Each will, at job talks, deploy the theoretical vocabulary that signals belonging to their tribe.

Neither is wrong, exactly. But both are narrower than they might have been. Neither narrowing was chosen consciously.

The oral examination compounds this. A viva or dissertation defense is assessed by scholars who are themselves products of a lineage. They are not grading the work against some Platonic standard of good scholarship; they are grading it against the implicit norms of their own formation. Work that challenges those norms is not automatically failed, but it faces a steeper climb. The student who writes a dissertation that fits neatly into the existing literature is easier to pass, easier to publish, and easier to hire. Originality, in practice, is rewarded only within certain corridors.

What People Get Wrong About Meritocracy in Academia

The standard defense of the current system goes something like this: doctoral training is demanding and selective, therefore those who succeed have earned their place on merit alone. This is the argument's weakest point. It deserves more pressure than it usually gets.

Merit is always assessed by someone. The criteria for what counts as a good thesis, a sharp argument, or an important research question are not discovered in nature; they are inherited and enforced by the people doing the assessing. When those people were themselves trained in a particular tradition, by particular scholars, with particular blind spots, the assessment reflects all of that. The test is real. The grading is genuine. But the syllabus was written by the previous generation, and it wasn't written to be neutral.

This is not a small distortion. Think of it like a compass placed near a strong magnet: the needle still moves, still points somewhere, still functions as a compass, but it no longer points north. Academic judgment still functions as judgment. It's just oriented toward something other than pure intellectual merit.

Here's the question worth sitting with: can you name the three or four scholars whose approval implicitly structures every important hiring decision in your department? If you can, you're looking at the magnet.

Why Reform Is Slower Than Anyone Admits

The people who benefit most from the current structure are also the people with the most power to change it. Not a cynical observation. A structural one, and it applies to every guild-like profession. But academia has a particular feature that makes reform unusually slow: its prestige economy runs on scarcity.

A journal's value comes partly from its rejection rate. A department's reputation comes partly from how few people it hires. A supervisor's status comes partly from how exclusive their recommendation is. Widen any of these apertures and you dilute the currency the entire system runs on. The incentive to preserve the hierarchy is therefore not just personal vanity; it's rational self-interest operating within a system that rewards exclusion. That is worth saying plainly, because the literature on academic reform tends to treat resistance as inertia when it might more accurately be read as strategy.

There have been genuine attempts to interrupt the loop. Co-supervision models, where students work with two or three scholars from different traditions, have expanded in some European programs. Structured training curricula, common in STEM and increasingly adopted in social sciences, reduce the single-supervisor bottleneck. Blind review processes for doctoral funding have, in some cases, redistributed resources away from historically dominant lineages. These are real changes. They matter at the margin.

But the reference letter remains. The job talk remains. The invisible canon of what counts as serious work remains. And as long as the people writing those letters, sitting on those panels, and enforcing that canon were themselves shaped by the same apprenticeship model, the reproduction continues, a little slower, a little less total, but continuous.

The university likes to present itself as the place where received wisdom gets challenged. In the training of its own future members, it is more often than not the place where received wisdom gets a salary, a title, and a set of keys to pass on.