Before there were essays, there was a person asking questions and a person answering them. The disputation of the medieval university, the doctoral viva voce, the clinical exam at a patient's bedside: for most of the history of education, the way you proved you knew something was to defend it out loud against someone who could push back.
Notice where the oral exam survived. PhD defenses. Medical licensing. Bar qualifications in many countries. Precisely the places where an institution cannot afford to be wrong about what a person knows. When the stakes are highest, education still reaches for a conversation. That is not nostalgia. It is a revealed preference.
When the stakes are highest, education still reaches for a conversation.
The evidence backs the instinct. In a study of biology students split into oral and written groups answering the same questions, both cohorts scored significantly higher under the oral condition, and, notably, no student group was systematically disadvantaged by the format. A meta-analysis of structured vivas in health-professions education found that when the format is well designed, with consistent prompts and defined rubrics, its reliability rivals or beats written formats. The recurring finding is that unreliability is a property of sloppy design, not of the modality. And students themselves report something interesting: oral exams push them to study for understanding, because they know reproduction will not survive a follow-up question.
Cognitive science suggests why. Two decades of work on retrieval practice show that pulling knowledge out of memory is not just a way to measure learning. It produces learning, often more durably than re-reading ever does. And dialogue does work that a monologue cannot: a follow-up question finds the edge of what you actually understand, and forces tacit reasoning into the open. An oral exam is retrieval under questioning. It measures and teaches in the same ten minutes.
So why did most classrooms abandon it? Arithmetic. One examiner, one student, fifteen minutes each: a three-hundred-student cohort is seventy-five hours of examiner time for a single assessment, before any preparation or moderation. Written exams won the nineteenth century because they could be administered to a full hall at once. The retreat of the oral exam was never a judgment that it measures poorly. It was a concession to scale.
That distinction matters enormously right now. The essay's claim to validity is weakening, as we argued in the first post of this series. The oral format's validity was never the problem; its cost was. Which raises a question that would have sounded like science fiction five years ago: what happens if the cost constraint moves?