Sources & methodology
The three records
- Faxian, A Record of Buddhistic Kingdoms (c. 414 CE), translated by James Legge, Oxford, 1886. The earliest and shortest account: an overland journey out (399–404) and a sea voyage home via Ceylon and Java (409–414).
- Xuanzang, The Great Tang Dynasty Record of the Western Regions (646 CE), translated by Li Rongxi, BDK English Tripiṭaka, 1996. The most detailed: some 110 “countries” in twelve fascicles, of which he visited most in person between 629 and 645.
- Yijing, A Record of the Buddhist Religion as Practised in India and the Malay Archipelago (c. 691 CE), translated by Junjirō Takakusu, Oxford, 1896. Not a travelogue but a treatise on monastic practice — and the best surviving witness to which schools flourished where.
- Yijing (supplementary), Biography of Eminent Monks Who Went to the Western World in Search of the Law During the Great T'ang Dynasty (Kao Seng Chuan, c. 691 CE), translated by Latika Lahiri, Delhi, 1986. Yijing’s memoirs of fifty-six fellow pilgrims, with his own detailed description of Nālandā — used to corroborate and refine a handful of his entries (each such fact carries its own quotation under “Also in the pilgrim’s writings”).
How the data was made
Every place on the map is one pilgrim’s recorded visit, extracted from the translations above with a mandatory verbatim quotation as provenance. Visits to the same real place were then reconciled, so that a single pin can carry up to three witnesses across three centuries. Nothing is invented: where the pilgrim does not name a school or count the monks, the data says so.
A note on “Hīnayāna”
The Victorian translations use “Hīnayāna” (“lesser vehicle”) where the pilgrims described monasteries of the early Buddhist schools. The term is pejorative and is not used on this site: we say Śrāvakayāna (“the vehicle of the hearers”) throughout. Likewise Yijing never wrote “Theravāda”; where the sources say Sthavira, the site shows “Sthavira (Theravāda)” as a scholarly bridge, not a quotation.
Locations and honesty
Coordinates come from the translators’ notes and established scholarly identifications (Legge’s footnotes, Cunningham’s Ancient Geography of India, and successors). Each pin carries a confidence level — confident, probable, or approximate — drawn on the map as a halo: the wider and blurrier the halo, the less certain the identification. Hollow markers are places the pilgrim only describes from report; the route lines never pass through them. The soft colour fields showing yāna dominance are deliberately borderless — sparse medieval testimony cannot justify hard edges.
Sarvāstivāda and Mūlasarvāstivāda
Xuanzang’s translator writes plain “Sarvāstivāda” throughout, but Yijing — who lived ten years under the rule he describes — distinguishes his Mūlasarvāstivāda from the older Daśādhyāya Sarvāstivāda, and warns (in effect) against relabelling wholesale. The map therefore adds the Mūlasarvāstivāda site by site, only where Yijing’s regional testimony supports it: the Sarvāstivāda monasteries of Magadha and Northern India, and Nālandā itself on his direct evidence. The Tarim oasis monasteries, which followed the older tradition, are deliberately left as the pilgrims recorded them. The pilgrims’ own words in each place’s detail panel are never altered.
Vajrayāna: a hidden lineage, read backwards
No pilgrim writes “Vajrayāna”. In their era it was an esoteric lineage transmitted orally inside Mahāyāna communities, and it became public only in the century after Yijing sailed home. But the records carry early traces of its mantra practice, and — exactly as with the Mūlasarvāstivāda above — the map reads the later evidence backwards site by site, at two places only, each shown as a flagged “(derived)” value alongside the witnessed yāna and never replacing it. On the map it is drawn the same way: the bubble keeps the witnessed colour, and the derived Vajrayāna appears as a purple ring around it.
Udyāna (631 CE): Xuanzang found a Mahāyāna country whose people “take the recitation of spells as their profession” and whose monks are “specially adept in reciting incantations”. This is the Oḍḍiyāna that Vajrayāna tradition reveres as the homeland of Padmasambhava and a source of the tantras, and openly Vajrayāna rock sculpture appears in the Swat valley in the seventh–eighth centuries. Nālandā (670s): Yijing himself “regularly went to the altar” there and tried — unsuccessfully, he admits — to master the Vidyādhara-piṭaka, a corpus of mantras he says ran to 100,000 stanzas and could not be understood until expounded orally. Under the Pālas, a few decades later, Nālandā taught Vajrayāna openly, and its excavated bronzes include Vajrayāna deities.
The ring, not a recolouring, is deliberate. A monastic community that takes up the mantra path does not stop being what it was: the monks still keep their vinaya — and every ordination lineage is a Śrāvakayāna inheritance, for the Mahāyāna never produced a vinaya of its own — while studying Mahāyāna literature and practising Vajrayāna meditation. A “Vajrayāna monastery” is therefore really triyāna, the three vehicles layered rather than exclusive; only outside the monastery could the mantra path be practised on its own. For the same reason, “Mahāyāna” on this map describes what a community studied and revered, not a separate institution. Yijing says as much: “Which of the four schools should be grouped with the Mahāyāna or with the Hīnayāna is not determined”, for “both adopt one and the same discipline (Vinaya)” — the Mahāyānists are simply “those who worship the Bodhisattvas and read the Mahāyāna Sūtras”. The “mixed” places, where both literatures were studied side by side, are the same layering seen from the other end. On the definitional problem see Jonathan A. Silk, “What, If Anything, Is Mahāyāna Buddhism? Problems of Definitions and Classifications”, Numen 49/4 (2002), pp. 355–405.
The derivation is gated on the pilgrim’s own testimony: it applies only where the witnessed community is Mahāyāna, so Faxian’s Udyāna of 403 CE — a Śrāvakayāna country in his account — is left exactly as he recorded it. Dhānakaṭaka, despite Xuanzang’s story of Bhāvaviveka reciting the Vajrapāṇi-dhāraṇī there and the later Kālacakra tradition attached to the site, is deliberately excluded: the community he records is Mahāsāṃghika. References: Anna Filigenzi, Art and Landscape: Buddhist Rock Sculptures of Late Antique Swat/Uḍḍiyāna (Austrian Academy of Sciences, 2015); Giuseppe Tucci, “Travels of Tibetan Pilgrims in the Swat Valley” (1940), for the identification of Swat as Oḍḍiyāna; Chou Yi-liang, “Tantrism in China”, Harvard Journal of Asiatic Studies 8 (1945), on Yijing’s Vidyādhara-piṭaka notice; Ronald M. Davidson, Indian Esoteric Buddhism (Columbia, 2002), on the tradition’s emergence into the open.
What the data cannot say
About a third of recorded visits name no yāna at all — those places stay neutral on the map rather than being guessed. Xuanzang reports 26 countries at second hand. Three visits have no derivable year and are shown with estimated dates. Vajrayāna was only beginning to emerge in the pilgrims’ era and never appears under that name in their records — the two derived attributions above are the closest the sources allow.
The pilgrims often count in phrases rather than figures — “several thousand monks”, “a few hundred”. Marker sizes on the map use documented representative values for such phrases (several myriads ≈ 10,000; several thousand ≈ 3,000; several hundred ≈ 500; few ≈ 20), but the detail panels always show the original wording.
Two famous monasteries are conspicuously absent: Odantapurī and Vikramaśīla were founded under the Pāla dynasty in the eighth century — a generation or more after Yijing sailed home — so none of the three pilgrims could have seen them. Their absence from this map is itself a small piece of historical evidence.
Colophon
Basemap geometry from Natural Earth (public domain); terrain relief from the Mapzen/AWS open elevation tiles. Type set in EB Garamond, Gentium Book Plus and Noto Sans. Built with SvelteKit, MapLibre GL, D3 and GSAP.