Graph omission and abbreviation in Tangut script

Introduction

Tangut (Xixia) script was created and issued under the guidance of Emperor Yuanhao 元昊 in the year 1036. From then on, there were not any official additions appearing in use except for some handwritten variants. However, in a few Tangut manuscripts the copyists might occasionally omit some graphs of the character or even create new characters by means of graph omission and abbreviation. A few newly created characters were spread merely across a limited sphere and were not included by any authoritative dictionaries.

Graph omission in a manuscript of Tangut dictionary

In the Khara-khoto collection of the Institute of Oriental Manuscripts, Russian Academy of Sciences (RAS), there are some fragmentary handwritten copies of an unidentified Tangut dictionary. In one of those manuscripts, Inv. No. 4152 (Figure 1), ¹ brief explanations on the character structure can be found. Each entry uses several incomplete graphs that are not recorded elsewhere – that is, , 修, 眩, 裀, 衈, 萨 and – as if the copyist merely wrote down the radicals and deleted other elements of the characters. Jia (2007) pointed out that they were ‘omitted graphs’ being used to indicate a certain part of standard characters borrowed to make new ones, meaning ‘top’, ‘bottom’, ‘left’ or ‘right’, respectively. This conclusion is possibly further proved by collating the Tangut authoritative rhyme dictionary Wenhai (Sea of Characters), which includes parallel entries (Table 1).

According to Table 1, the correspondence between omitted graphs and their original characters may be shown in the following:

= 緵 ɣu, ² head, top

修 = 崇 tśhjɨj, bottom

眩 = 纳 pha, left

裀 = 緈 bjịj, right

衈 = 蛃 njij, middle

萨 = 饲 iọ, rim

= 笒 ŋowr, whole

Since the ‘unidentified dictionary’ was copied in a casual manner, it must be a booklet more for copyist’s private reference than for popular reading; thus, it is understandable that the reason why the copyist deleted part of the graphs that frequently occurred in the same speech environment was to save his writing efforts. In any case, these incomplete graphs are caused by personal inspiration and cannot be regarded as real innovations.

OPEN IN VIEWER

Figure 1.

Inv. No. 4152, Institute of Oriental Manuscripts, RAS.

Graph abbreviation in Tantric manuscripts

Merely omitting elements of a character, as seen in the previous section, will not bring any new ones into the writing system, but a few Tangut characters prove to be created by combining two or three omitted graphs and present a special pronunciation, which might be regarded here as ‘abbreviation characters’. The most interesting examples come from the Tangut phonetic transcription of the Tibetan word badzra, which is the native reading of the Sanskrit word vajra, meaning ‘diamond’ (Chinese jingang 金刚). ³

The word badzra, repeatedly appearing in Buddhist texts, especially in dhāraṇīs and gāthās, is usually transcribed into Tangut 傅滴蚐 ba-dzjɨ-rjar or 菤滴蚐 bja-dzjɨ-rjar. ⁴ Because the consonant cluster dzr- (Sanskrit jr-) is impossible in their language, Tanguts had to insert a vowel ɨ between the two consonants and transcribe it into dzjɨr- (滴蚐 dzjɨ-rjar). Sanskrit consonant clusters troubled Tangut transcribers a lot, so they tried to find some way to present them as correctly as possible. One of the most commonly used methods was to write the second character slightly smaller to indicate that it should be read rapidly in combination with the previous one (Sun, 2010: 138) – for example, 滴蚐 for dzra; however, there were other ways to treat this syllable.

In the Or.12380-638 of the Stein collection in the British Library (Xibei Dier Minzu Xueyuan et al., 2005: 232 (Figure 2), ⁵ on lines 2–4 there appears twice a peculiar character, , which was not included by any dictionaries in Xixia or modern times. Here is the relevant stanza in an unidentified gāthā:

OPEN IN VIEWER

Figure 2.

Or.12380-638, British Library.

OPEN IN VIEWER

Figure 3.

Or.12380-596, British Library.

The word-by-word reading evidently shows that the Tangut character marked by ‘△’ is in parallel with the Tibetan syllable dzra. In reference to the standard transcription 傅滴蚐 ba-dzjɨ-rjar frequently used, it may be recognized that the peculiar character is a combination of the top and left bottom of 滴dzjɨ, and the left of 蚐rjar. That is to say, the Tangut character consists of two other omitted graphs and should be read as dzra or, according to its Sanskrit form, jra.

The second peculiar character is , appearing on the second line of Or.12380-596 in the British Library (Xibei Dier Minzu Xueyuan et al., 2005: 219) (Figure 3). Here are the relevant syllables in an unidentified fragmentary dhāraṇī:

On the sixth line of Inv. No. 4900 of the Kozlov collection in the Institute of Oriental Manuscripts, RAS. (Figure 4), there is also the same character appearing in the Sanskrit title of Tangut transcription:

OPEN IN VIEWER

Figure 4.

Inv. No. 4900, Institute of Oriental Manuscripts, RAS.

Obviously, the Tangut character marked by ‘△’ is in parallel with the Tibetan word badzra or Sanskrit word vajra. By reference to the standard transcription 傅滴蚐 ba-dzjɨ-rjar, it may be recognized that the peculiar character is a combination of the left of 傅 ba, the top and left bottom of 滴 dzjɨ and the left of 蚐 rjar. That is to say, the Tangut character consists of three other omitted graphs and should be read as badzra or, according to its Sanskrit form, vajra.

The third peculiar character is , appearing in the Tangut translation of a dhāraṇī in the Tibetan version Dpal kye-rdo-rje’i mngon-rtogs yid-bzhin nor-bu (hereafter Dkm): ⁸

Similar to badzra, the Tangut character badzra is undoubtedly an abbreviation merger of the standard transcription 菤滴蚐 bja-dzjɨ-rjar, and possibly the peculiar character is a combination of the left of 菤 bja, the top and left bottom of 滴 dzjɨ and the left of 蚐 rjar.

It is curious that the newly created dzra is suspected to be used directly to transcribe badzra without adding the left part of 菤 bja to present ba, as also in Dkm:

These characters are rarely seen in Tangut scriptures, so it is believable that their formation and usage were somehow discretional in private writing practice.

OPEN IN VIEWER

Figure 5.

Inv. No. 4373, Institute of Oriental Manuscripts, RAS.

Nature of the abbreviation characters

The most interesting fact is that the standard form 菤滴蚐 bja-dzjɨ-rjar and the two newly created characters dzra and badzra may occur adjacently in the same manuscript, as in Inv. No. 4373 of the Kozlov collection in the Institute of Oriental Manuscripts, RAS (Figure 5): ⁹

This fact indicates that Tangut abbreviations were created voluntarily, and that Tangut copyists randomly wrote any forms as they wished. It is noticeable that the abbreviations appeared only in the popular manuscripts of Tantric works in Tibetan style and cannot be found in any xylographs of Buddhist sūtras in regular Chinese tradition. This suggests that the spreading of abbreviation characters was limited to the Tantric realm and not accepted by Xixia official scholarship.

The idea of creating new characters by merging two or more abbreviated graphs might have been an ancient tradition established by Buddhists along the Silk Road, for in many manuscripts excavated in the Mogao Grottos at Dunhuang there is a Chinese character pusa as an abbreviation of 菩萨 pusa. ¹⁰ It is well known that all the characters in the Chinese writing system are usually monosyllabic, but still there were a few Chinese characters in the 20th century showing disyllabic words – for example, 浬 haili for 海里 haili (nautical mile), 瓩 qianwa for qianwa 千瓦 (kilowatt) – though they were banned by the Chinese government for character standardization.

Conclusion

Making new characters to record polysyllables or syllables with consonant clusters by means of graph abbreviation proves to be one of the popular ways to enrich the Chinese writing system, but such a method cannot be adopted officially because it only refers to several special terms occurring in a limited professional realm and will never be popularized in people’s daily lives.