What survives at Nalanda is a row of identical courtyards. Each is a hollow square of small rooms, perhaps three metres by two, opening onto a verandah that runs around a brick-paved court; some have a well. There is no lecture theatre in the modern sense, no faculty building, no hall that could hold a tenth of the student body at once. Walk the site and the building type announces itself before any inscription does: this is not a university. It is a hostel with a curriculum attached.
That is the useful way in, because a hostel is an accounting problem before it is an architectural one. If every admitted student is housed and fed at the institution's expense, then the institution's size is not set by its reputation or its teachers. It is set by how many people its endowment can feed, and the whole of Nalanda's eight-hundred-year run is the working-out of that constraint.
What paid for it
The endowments are the institution. Xuanzang, who studied here in the 630s, records that the emperor Harsha assigned the revenue of a hundred villages, with two hundred households in each village charged to deliver rice, butter and milk. A later copper plate found at the site records something more revealing still: in 860 CE a king of Sumatra, Balaputradeva, built a monastery here and asked the Pala ruler Devapala for five villages to maintain it, and the grant was made.
Read that again as an administrative fact rather than as a monument to international prestige. A ruler in Sumatra could endow a foundation in Bihar, and the endowment was not money but revenue — the right to collect from named villages, in grain, clarified butter and milk, delivered. That is a food supply with a legal deed. It explains a great deal that otherwise looks like piety: no fees are recorded, because fees were not the model; and the monastery could survive a bad patron or a change of dynasty, because it did not depend on one treasury but on a portfolio of villages whose obligation ran with the land.
It also explains what the place actually consumed. Thousands of young men, no farms of their own, a kitchen running daily, butter for the lamps as well as for the food, and a building stock of mud, brick and timber that needed constant repair. An endowment in grain is an institution designed around spoilage: it must be eaten this year, which means the population it feeds cannot fluctuate much, which means admissions cannot fluctuate much either. The rule that every student is housed and fed is not generosity. It is a capacity plan.
The gate does the selecting the fees would have done
A free institution has one structural weakness: everybody wants in. Nalanda's answer was a doorkeeper with a syllabus. The office is recorded as the dvara pandita, the scholar at the gate, and the examination was oral, on Sanskrit and on the foundational texts. Yijing adds the other half of the filter: candidates came with a recommendation, and were tested on what they claimed to know.
This is the mechanism that makes the endowment work. Because entry was hard, the institution could promise housing to everyone it admitted without promising housing to everyone who applied; the gate converted an open-ended obligation into a fixed one. It also produced the thing Nalanda was famous for abroad — not degrees, which did not exist, but competence that could be tested in public argument. A monk who had passed the gate had been examined by people with a reputation to protect, in a place where losing an argument in front of the assembly was a career event.
What was actually taught, and to whom
The core was Mahayana doctrine and the eighteen schools' texts, but the curriculum was wider than the label suggests: logic and the theory of valid knowledge, grammar and philology, medicine, Samkhya philosophy, and enough of the Vedic corpus to argue with people who held it. Tibetan sources list four doctrinal positions taught here, which is a way of saying that students were expected to be able to state an opponent's case before refuting it.
The intake was international in a specific and commercial sense. Xuanzang went home in 645 with 657 Sanskrit texts in 520 cases on twenty horses; Yijing carried back about four hundred. Korean and Tibetan monks studied here; a Sumatran king endowed a monastery here. For most of its life, Nalanda's most valuable export was not trained people who stayed — it was corrected, copied, portable text, plus the ordination lineages that let a monk from outside India be recognised as legitimately ordained when he got home.
The library as a copying factory
The library is called Dharmaganja, the treasury of the doctrine, and the Tibetan accounts give it three buildings: Ratnasagara, Ratnaranjaka and Ratnodadhi, the last nine storeys high and reserved for the most protected texts. No part of it has been excavated, so treat the architecture as reported rather than measured.
The function, though, is not in doubt, because the function explains the traffic. A manuscript library in a world without printing is a copying institution: a text survives in the number of copies that leave the building. Pilgrims came with horses and cases because the library would let them copy, and the copies they carried to China and Tibet are, in several cases, the reason the text exists at all today. A library that only stored would have burned once and been finished. This one had already distributed itself.
What ended it, and what did not
Around 1200 the site was sacked, and there is a layer of ash across several buildings to prove that something burned. The passage usually quoted as proof of Nalanda's destruction, however, is Minhaj-i-Siraj on the burning of a great monastery at Bihar — and the monastery he names is Odantapura, a different foundation a few kilometres away. This is the kind of detail that ruins a good story, and it is worth keeping: the fire is documented; the identification is an inference.
What is documented is the aftermath, and it is more interesting than a fire. In the 1230s a Tibetan pilgrim, Dharmasvamin, spent a monsoon season here and found the place wrecked, largely deserted, and still teaching — a very old teacher, a small class, a lay patron keeping it alive. A decade or two earlier the last throne-holder had left for Tibet. So the institution did not stop on a single night. It thinned out.
The reason is the endowment. An institution whose buildings are maintained out of village revenue, and whose students are fed out of the same revenue, has no way to shrink gracefully: the moment the villages stop paying, the hostel becomes a liability rather than an asset, and the rule that made it famous — everyone admitted is housed and fed — becomes the reason it cannot continue. What survived was what could be carried. The books went to Tibet, and the teaching continued there in a foundation deliberately named Nalanda, built north of Lhasa in the fifteenth century.
This pillar keeps asking what had to be built, paid for and staffed before anything could be believed in public. Nalanda is the clearest answer in the set: a dormitory, a kitchen, a hundred villages' revenue, a doorkeeper with a syllabus — and, for eight hundred years, an intake of young men who arrived with nothing and were fed.
Sources and checks
- Xuanzang Si-yu-ki: Buddhist Records of the Western World, trans. Samuel Beal. London, 1884 — the account of the monastery, its admissions and its revenues
- Yijing A Record of the Buddhist Religion as Practised in India and the Malay Archipelago, trans. J. Takakusu. Oxford, 1896
- Roerich, George N. (trans.) Biography of Dharmasvamin (Chag lo tsa-ba Chos-rje-dpal). Patna: K. P. Jayaswal Research Institute, 1959 — the 1235 visit
- Sankalia, H. D. The University of Nalanda. Madras, 1934 — still the clearest assembly of the literary evidence
- Dutt, Sukumar Buddhist Monks and Monasteries of India. London: George Allen & Unwin, 1962 — on the endowment and the daily rule
- UNESCO World Heritage Centre Archaeological Site of Nalanda Mahavihara at Nalanda, Bihar — nomination dossier and inscribed property 1502, 2016
Last checked 2 November 2026 Yellowsea Huang Corrections