The tangoute is old a Langue Tibéto-Burmese E. It belongs to the group qianguic, which is only in a remote way related with the Chinese, with the Tibetan or the Burmese . Among the qianguic languages, one finds in particular the RGyalrong and the qiang.

It is the old official language of the Semi-nyag kingdom (also known under the name Chinese of Xixia 西夏), which obtained its independence of the Song at the beginning of the 11th century and was destroyed by Gengis Khan in 1227. The writing tangoute, that Sofronov (1968) regards as most complex of all the history of humanity, was created by decree of the emperor Li Yuanhao 李元昊 in 1038. The design of the writing was entrusted to Yeli Renrong 野利仁荣, a scholar close to the imperial family. After the destruction of the kingdom, the writing did not completely disappear, and was used at least until the end of the 15th century.

The majority of the texts tangoutes were excavated with Kara-koto in 1906 by Kozlov forwarding, and these documents are preserved now at Saint-Pétersbourg. In addition to the Buddhist gun and traditional confucéens, a great number of made up original texts in tangoute arrived to us.

The bond of the writing with the pronunciation of the language tangoute is limited even more than that of the Chinese writing with the pronunciation of the modern dialects. Whereas in Chinese more than 90% of the characters include/understand a phonetic element, this proportion is limited to 10% according to Sofronov. The rebuilding of the pronunciation tangoute had to count on other sources.

The discovery of the Fanhan heshi zhangzhongzhu 番汉合时掌中珠, a bilingual glossary tangoute-Chinese, allowed Ivanov (1909) and Laufer (1916) to propose the first rebuildings and to begin the comparative study of the tangoute. This glossary indeed indicates the pronunciation of each character tangoute by one or more Chinese characters, and conversely of each Chinese nature by one or more characters tangoutes.

The second source is the corpus of the transcriptions Tibetans of the tangoute. These data were studied for the first time by Nevsky (1925).

However, according to Sofronov, these two sources are not reliable by themselves for a systematic rebuilding of the tangoute. Indeed, these transcriptions were not written in the idea to represent with precision the pronunciation of the tangoute, but simply to help the foreigners to pronounce and memorize the words of a language which they could at that time hear around them.

The third source, which constitutes the base of the modern rebuilding, includes/understands the monolingual dictionaries tangoutes: the Wenhai 文海, two editions of the Tongyin 同音, the Wenhai zalei 文海杂类 and a dictionary without title. The notation of the pronunciation in these dictionaries is based on the principle of the Fanqie 反切, borrowed from the Chinese lexicographical tradition. Even if these dictionaries differ on minor details (for example, the Tongyin class characters by syllabic attacks and rhymes without taking account of the tone), they adopt all the same system of 105 Rime S. a certain number of rhymes are in complementary Distribution compared to the Joint of initial, like rhymes 10 and 11 or rhymes 36 and 37, which shows that the scholars who composed these dictionaries carried out a very precise analysis of the phonological system of their language. With the differences of the transcriptions in foreign languages, the fanqie tangoutes mark the distinctions between the rhymes in a systematic and very precise way.

Thanks to the fanqie , we have a good comprehension of the phonological categories of the language. However, it is necessary to compare the phonological system of the dictionaries tangoutes with the other sources “to fill” the phonetic value of these categories.

Examples of writing tangoute:

The two characters in top are nga 1 “1sg” nja 2 “2sg” two suffixes verbal, and both of the lower part are dzji 1 “to eat” and wji 1 “to make”.