Comparison

The Animal Mouths of Anavatapta: Why Traditions Differ

Sui–Tang Chinese, Angkorian Khmer, Thai manuscript and modern reception

There is no universal animal-to-direction code for Anavatapta. West–horse recurs in several well-known witnesses, while the other pairings change. The Pāli Jātaka corpus has no animal-mouth system at all. The disagreement is evidence of transmission and transformation, not an error to be corrected away.

Anavatapta animal mouthsfour animal heads Buddhist cosmologyAnotatta animalsfour rivers animals

1828 · Kyoto University Library, Muroga Collection RB00022886 Tenjiku yochizu Tenjiku yochizu (1828), Kyoto University Library, Muroga Collection RB00022886. Kyoto University Library free-license terms · source · The institutional credit and source link are retained. Any crop or alteration must be disclosed.

What this page does

One source family at a time

This guide links each major assertion to a named source and a boundary. It is designed to be useful for readers, artists, teachers, and researchers without pretending that later images can repair earlier silence.

The comparison that changes the question

Many popular images make the four animals seem like a universal Buddhist code. The controlled archive shows a different picture. The animals must be read with their witness: date, medium, language, source relationship, and exact location all matter.

Instead of asking “which animal is east?” in the abstract, this page asks “which witness says which animal is east?” That makes the tradition more interesting and makes the answer checkable.

A compact witness table

Sui Chinese T24/T25: east elephant, south bull, west horse, north lion. Tang Great Tang Records: east silver ox, south golden elephant, west horse, north lion. Neak Pean: north elephant, east human, south lion, west horse. Berlin Traiphum II 650: east lion, south bull, west horse, north elephant.

The shared west–horse is a pattern worth noticing, but it is not enough to make the rest identical. Each row belongs to a different textual, architectural, or manuscript setting.

What early Pāli evidence does not say

The absence matters as much as visual abundance. The Pāli Jātaka source family does not supply an animal-mouth system. The Samantapāsādikā likewise does not provide a connected four-animal directional map.

A later image can be an excellent object to study. It cannot silently fill an early textual gap. This is why the site uses witness tables rather than a universal decorative compass.

Evidence ledger

Claims you can trace

These are the precise points supporting the guide. Each includes a claim type, source, locator, and limit.

Direct sourceChinese Buddhist texts

Chinese rows

Sui and Tang records preserve different animal-direction arrangements.

Locator: T24/T25 and T2087 comparison ledger

Does not establish: Terms and directions must stay attached to individual texts.

Visual witnessNeak Pean

Khmer row

Neak Pean has north elephant, east human, south lion, and west horse outlets.

Locator: Architecture record and outlet survey

Does not establish: Architecture does not name the monument Anavatapta.

Direct sourceJātaka commentary

Pāli absence

The audited Pāli Jātaka family does not offer an animal-mouth system.

Locator: History 02 controlled extraction

Does not establish: The absence is bounded to the audited corpus.

What not to carry forward

Do not use a four-animal diagram without naming the witness it follows.
Do not supply a Pāli animal code that the audited Pāli evidence does not contain.

Sources

Read the evidence layer

Primary textWestern Jin–Tang, 266–646 CE

Chinese Buddhist Anavatapta source family

T23, T24, T25, T197, T199, T2087; History 04 source manifest

Translations and Xuanzang/Bianji’s record create witness-specific Kunlun, river, mountain, and animal-direction configurations.

Boundary: Source-reported geography is not a modern field survey and does not erase other variants.

Architecturelate twelfth century CE

Neak Pean architectural record

Jayavarman VII’s Angkor complex: central basin, island, and four satellite pools

A measurable sacred-water monument associated with Jayataṭāka and the island called Rājyaśrī in K. 908.

Boundary: Built form can be compared with cosmographic traditions; it does not prove a single textual source or label.

Manuscript objectcommission associated with 1776

Berlin Traiphum II 650 research record

Anotatta at fols. 35v–36v; six-metre public project section

An eighteenth-century Thonburi cosmogram with an object-specific animal-direction arrangement.

Boundary: The surviving manuscript’s relationship to the 1776 commission remains disputed; this site does not reproduce unlicensed imagery.

Open source ↗

Primary textreceived fifth-century commentary tradition

Jātaka-aṭṭhakathā narrative geography audit

Ja 514 (Chaddanta), Ja 536 (seven lakes), Ja 485 and Ja 504 (kinnara settings)

Narrative landscapes are held as their own family so individual stories cannot silently rewrite the Samantapāsādikā baseline.

Boundary: A landscape in one Jātaka is not automatically a universal Himavanta layout.

Research control record2026

Himavanta cross-history research synthesis

Cross_History_Thread_of_Truth.md; separately frozen history packages

The project comparison layer connects independently audited source families without merging their claims.

Boundary: A synthesis is not a primary witness and cannot fill a source’s silence.