A State-Sponsored Script
The Western Xia kingdom (Xī Xià, 西夏, 1038–1227) used writing as part of its state administration and cultural policy. In 1036 CE, Emperor Jingzong commissioned the polymath Yeli Renrong (野利仁榮) to design a national writing system for the Tangut language. The resulting system contained roughly 6,000 characters. They are dense, square and multi-component like Chinese characters, but the characters themselves were newly constructed for Tangut.
The visual resemblance was deliberate. Located between Song China and the Liao Khitan kingdom, the Tangut state used its own script to support administration and to assert cultural parity with its neighbours. The components nevertheless carry Tangut-specific phonetic and semantic values, so knowledge of Chinese characters does not make a Tangut text readable.
Structure Without Chinese Meaning
Like Chinese, Tangut characters are built from recurring graphic elements — roughly 300 basic components. A character typically combines two such elements to produce a new meaning, analogous to Chinese phono-semantic compounds. But the analogy is surface-level: the components themselves, their combination rules, and their readings are all native Tangut inventions. For example, the Tangut word for “person” uses a component that superficially resembles the Chinese radical 人 (rén) but has a completely different origin and sound.
The language encoded is Tibeto-Burman (distantly related to Tibetan and Burmese), which has no genetic connection to Sinitic Chinese. The script therefore applies a Chinese-influenced visual model to a Tibeto-Burman language with a different structure.
The Bilingual Glossary 番漢合時掌中珠
Our best window into Tangut vocabulary is a 12th-century bilingual glossary, the Fānhàn héshí zhǎngzhōng zhū (番漢合時掌中珠, “Pearl in the Palm, Tangut and Chinese for the Times”). Compiled in 1190 by a Tangut scholar named Gule Maocai (骨勒茂才), it arranges Tangut words alongside Chinese equivalents across thematic categories — body parts, tools, kinship terms — making it a central source for Tangut studies. The preface states that neither a Tangut who cannot read Chinese nor a Chinese who cannot read Tangut is truly educated.
Later Use and Modern Decipherment
When the Mongols destroyed the Western Xia capital Zhongxing (modern Yinchuan) in 1227, the Tangut state was erased, but the script was not. Tangut communities went on writing and printing it under the Yuan and into the Ming; the last known example is a pair of Tangut dharani pillars erected at Baoding, in present-day Hebei, in 1502. The script then ceased to be used for roughly four centuries. In 1908–09, the Russian explorer Pyotr Kozlov excavated the ruined city of Khara-Khoto in the Gobi Desert and unearthed a cache of thousands of Tangut manuscripts — Buddhist sutras, legal codes, dictionaries, poems — now held at the Institute of Oriental Manuscripts in St. Petersburg.
Systematic decipherment began in earnest only in the 20th century. The first comprehensive frameworks for reading the language came from the Japanese scholar Nishida Tatsuo (西田龍雄), whose Seikago no kenkyū (西夏語の研究, 1964–66) paired a reconstruction of Tangut phonology with a decipherment of the script, and from the Russian scholar Mikhail Sofronov, whose Grammatika tangutskogo yazyka (1968) set out a reconstruction alongside a full grammar. Gong Hwang-cherng (龔煌城) later refined the phonological system in a series of studies; his reconstruction was adopted for Li Fanwen's Tangut–Chinese dictionary (Xià-Hàn zìdiǎn, 夏漢字典, 1997) and summarised for English readers in his ‘Tangut’ chapter of The Sino-Tibetan Languages (2003). Together with ongoing projects at Kyoto, Beijing, and St. Petersburg, this work allows scholars to read an estimated 70–80% of attested Tangut characters today.
Digital Encoding and Historical Significance
Tangut is a documented case of large-scale script creation under state sponsorship. Its visual relationship to Chinese writing also reflects the political setting in which it was designed. Unicode finally encoded Tangut in 2016 (U+17000–U+187FF), giving the script its first digital home nearly 800 years after the kingdom that created it was destroyed.