Xixia Language Studies and the Lotus Sutra (II) Tatsuo Nishida 1. The Birth and Death of the Xixia Language—Its Life IMMEDIATELY after he had established the Xixia state in 1032, its first emperor Li Yuanhao ordered the creation of a new written language and a writing system for properly recording it, one that would replace the Tibetan and Chinese scripts that had been used till then. He intended, internationally, to enhance the prestige of his empire and, domestically, to strengthen the unity of its constituent tribes and to pro- mote vernacular culture.1 The language and its ideographs are called the Xixia language and Xixia characters, respectively. The project was quickly and superbly executed and in 1036 the fruits of the endeavor— the language and the characters—were promulgated throughout the country. As a result, a new culture area was formed in the northwestern corner of the Chinese civilization sphere, as Xixia characters replaced the use of Chinese characters. The creation of the Xixia characters also represented the birth of a unique type of ideograph. Most of the tribes which composed the Xixia state, such as the Mi and the Minyak, spoke their own dialects, derivative of a common par- ent language (i.e., they were dialects of Dangxiang yu ). Though these dialects are considered to have shared common features,2 they must have been fairly different from each other in terms of phonology, grammar and vocabulary formation. For example, these differences between the expressions of their own designations in the Mi and Minyak dialects can be noted: Mi ( ) Minyak ( ) mi(cid:4) (rising 10) m(cid:5) (rising 28) n˘(cid:7)a(cid:4) (rising 18) n˘(cid:7)a(cid:4) (level 21) “black” Mi “black” Regarding the self-names in Xixia, the Mi tribe pronounced it as rising tone 10th rhym mi(cid:4), and the Minyak tribe as rising tone 28th rhym m(cid:5). Regarding the word for “black,” both had it as n˘(cid:7)a(cid:4), the Mi pronounced it with level tone, and the Minyak with rising tone. 222 XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) 223 It is very probable that the second syllable, mi(cid:4) (level 11), of ŋww- mi(cid:4) (level 11) meaning “imperial family” was one of the corresponding cases of the Mi. Its meaning might have been the Mi of Wei .3 From the time of its promulgation in 1036, the Xixia language improved and developed in its power of expression through the process of translation of Buddhist canons, official documents and Chinese clas- sics, and through the creation of the nation’s own vernacular literary works. The language reached the apex of its development during the reigns of Hui Zong (1069–1086), Chong Zong (1086–1139) and Ren Zong (1139–1194).4 Then, as the state declined, the Xixia language was brought to the verge of extinction. It appeared to have ended its life after the destruction of the Xixia state by Mongolian troops, who invaded from the north in 1227. However, the Xixia lan- guage received a new lease of life thanks to the Yuan (Mongolian) dynasty’s preferential treatment of the language. This invigoration came about through the publication of The Yuan Version of the Xixia Tripitaka . The homeland of Xixia (Tangut) was, like Tibet, under the rule of the Office of Administrative Control (later, the Office of Benevolent Administration ), and the Tangut were given preferential treatment as the semu ren (persons with special status). Shi Zu (Kublai Khan), who saw in the promotion of Buddhism a means of exerting con- trol over the populace, ordered the production of a set of printing blocks of the Xixia Tripitaka in Dadu , the capital of the Yuan empire. The project was suspended in 1294 when Cheng Zong ascended the throne, but it was soon resumed and continued at Wanshousi in Hangzhou. It seems that by 1302, more than 3,620 volumes of Xixia scriptures, called the Hexi Tripitaka , were completed and donated to temples in the Tangut homeland. By this time, more than 70 years had passed since the fall of the Xixia state. The completion of the project is an almost certain fact, judging from colophons attached to Da Zong di xuanwen benlun in the collection of Zenpukuji temple in Japan, as well as preface texts for Xixia scriptures obtained by Paul Pelliot, and fragments in the col- lection of the Tenri Central Library .5 In 1970, I found extant portions of the original Xixia triptaka at the Swedish National Museum in Stockholm, as confirmed by the “Bai- shang Daguo Xinyi Sancang shengjiao xu” (Pref- ace to the Principal, Great and Superior [i.e., Xixia] State’s New Trans- lation of the Three-Basket Sacred Teaching) that was attached to these 224 XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) scriptures. Six sutras remained: the Foshuo Yueguang Pusa jing (Buddha-delivered Moonlight Bodhisattva Sutra), the Foshuo Liaoyi Bore boluomi jing (Buddha-delivered, Meaning-clearly-determined Perfection of Wisdom Sutra), an unidenti- fied sutra—only the latter part of which remains, without the beginning and ending portions—along with three other fragments (A, B and C).6 As the preface refers to itself as “New Translation” , the scriptures must belong to what I have termed the second period of the Xixia Tripi- taka translation. Though it is not possible to determine how many of the volumes of the tripitaka (sutra, vinaya and abhidharma) were actually included, one may assume that it was a compilation of newly-translated Buddhist scriptures and a reprint of the tripitaka completed during the reign of Ren Zong (1139–94). The fact that the translation project was completed suggests that the Xixia language was still widely used—and that knowledge of its charac- ters had been retained—among people in the middle of the 14th century. On the inner wall of Guojieta (Crossing Street Tower) at the Juyong-guan (Common Dwelling Pass), which is on the way from Beijing to the Western Regions, there are inscriptions of dharani texts and the account of building the tower in six languages and scripts. One of these inscriptions is in Xixia language and characters. The inscriptions were made in 1345, thirty-three years after the completion of the Hexi Tripitaka. It is possible that the Xixia script was included because the Xixia people remained an influential ethnic group even 118 years after the fall of their state. Although the Xixia texts in these inscriptions has lost the B-stems of verbs, and the use of view sentences is not as frequent as in earlier times, the script still appears to fully pre- serve the basic grammatical structure of the Xixia language, to the extent that it uses both of the two series of prefixes.7 Regarding Xixia texts after the Juyong-guan inscription, there is a paper that introduces the extant text of “Gao Wang Guanshiyin jing Gan De xu” (Preface by Gan De to the Gaowang Bod- hisattva Perceiver of the World’s Sounds Sutra) (29 lines) dated the fifth year of Hongwu (1372) of the Ming dynasty. This text is preserved at the Palace Museum in Beijing. The text in the preface still frequently uses the prefix series B.8 The approaching end of the Xixia language is evident in the text on an octagonal stone pillar, Xiangsheng chuang , unearthed in Hanzhuang , Baoding , 130 kilometers southwest of Beijing.9 It was probably built by descendants of the Xixia people who had been relocated to Dadu. As the date of the pillar’s completion is considered to XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) 225 be the 15th year of Hongzhi (1502), by that time 275 years had passed since the fall of the Xixia state. In the Xixia text on this pillar, the prefix series A appears, but the text does not include the B-stems of verbs. In this regard, no view sentences appear at all. The Xixia charac- ters are used to enumerate proper nouns such the names of offices and people. One gets the impression that the characters are now little more than the empty hulls of the language. It is also possible that the charac- ters had been handed down while the language had fallen out of use and been almost forgotten by then. In any case, it appears that the complete demise of the Xixia language was then imminent. The written language is considered to have lived for a little more than 450 years following its birth in 1036.10 I believe the extinction of a written language and that of a natural, spoken language should be treated as different matters. 2. Analysis of the Character System and the Restoration of the Linguistic System Since I began studying the Xixia language and Xixia characters in the 1950s, I have always taken heed of treating the two tasks of (1) analyz- ing the character system and (2) restoring the linguistic system on dif- ferent levels, and giving consideration to a balance between these two concerns. It is a cardinal rule that level of development of the characters and that of the language itself should not be confused when one carries out linguistic studies. It is, however, true that development in the under- standing of one area accelerates development in the other. Since one can explore the Xixia language only through Xixia characters, the two, in particular, are closely connected and limitlessly intertwined. A differ- ence in one character reflects a certain morphological difference in the language.11 The concepts of twin characters and double twin characters that I con- ceived has had a substantial effect in this field of discipline.12 Pairs of twin characters with similar forms represent various correspondence and contrast on the morphological level. I do not however, discuss this fur- ther herein. Difference in character form reflects a difference of meaning unit, i.e., phonological difference. To put it simply, when one reads a character form, there arises the question of which of the meaning units the charac- ter form refers to. In this regard, the research on the difference of char- acter form is closely connected with the restoration of the phonological system of the Xixia language and, consequently, it is deeply concerned 226 XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) with the genealogical study of this language. At an early stage of my study in the 1960s, I attempted to restore the phonological system mainly based on the system from the lexicon Tongyin « »13. In the 1980s I published a revised version of my ideas after a major reexamination. For this I made reference to systems formu- lated in the rhyme table Wuyin qieyun « », and the rhyme lexi- cons Wenhai(Baoyun) « ( )» and Wenhai zalei« ».14 Since then, I have continued to revise this version for my personal use, but there remain a fair number of matters that require further exami- nation. The main points among them are: (1) To consider the correspon- dence between the subgroup of level tones and the subgroup of rising tones ; this is concerned with the restoration of ris- ing-tone initial consonants . (2) To reconsider initial conso- nants and rhymes from the perspective of morphology; e.g., the reconstruction of vowels based on contrasts between (a) the voiced sounds of intransitive verbs and the voiceless unaspirated sounds of transitive verbs, and (b) between relaxed vowels of non-causative verbs and tightened vowels of causative verbs.15 I hope to extensively discuss these problems in articles in the near future. Here I will consider some basic points concerning the meanings of each particular characters, making reference to the Lotus Sutra and related texts. Books like Sancai zazi « », Tongyi yilei « » and Fanhanyu yaoji « », which are works the Xixia people at the time left as the products of their studies,16are, needless to say, very important material for analyzing the meanings of Xixia characters. However, I put these works out of my consideration here. 3. The Analysis of Meanings Lexicons, vocabulary books and phonological works all help provide important evidence in verifying the meanings of Xixia words.17 Howev- er, the final determination is made through a correspondence between the Xixia word and its counterpart from a known language, such as Chi- nese or Tibetan. Let us take the basic example of verifying the meaning of the word “Buddha” . The meaning as “Buddha” is one of several characters known at the very beginning stage of early 20th century semantic research on the ideographs, when M. G. Morisse made public his studies.18 It wouldn’t be accurate to say that Morisse deciphered the meaning; the Chinese character had been written by an anonymous Chinese beside the original text in the Xixia Lotus Sutra that Morisse obtained. XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) 227 At the present time, I can say the following about this character : It belongs to the 17th rhym (yun ) in Wenhai (Baoyun) and is pro- nounced thɑ(cid:4). It was originally used as the two-syllable word fw(cid:5)(cid:4) thɑ(cid:4). Later thɑ(cid:4) by itself came to represent “Buddha.” It is not an original Xixia word but a loanword. The word fw(cid:5)(cid:4) thɑ(cid:4) is a Xixia transliteration of the Chinese word Fotuo , which, in turn, is a Chi- nese transliteration of the Sanskrit “Buddha.” I assume that, while the first syllable of the Chinese Fo (copying “bud”) came to represent the meaning of “Buddha” in Chinese, the second syllable dhɑ (copying “dha”) was used to represent “Buddha” in Xixia.19 Translators of the Xixia Buddhist scriptures were able to rely on the Buddhist terms established in Chinese and Tibetan. They did not trans- late directly from Sanskrit but created many transliterated Xixia words from the equivalents in Chinese. Some terms in the original were accepted as established, commonly-used Buddhist terms. That is to say, they were foreign words phonetically copied with vernacular characters. Besides the examples of transliteration such as thon lɔ(cid:4)(cid:2)i and thɑ(cid:4) mɔ(cid:4) , there are many words which were creat- ed through semantic translation.20 The correspondence of names of things, in particular, can be made with relative ease. Among familiar examples are the Seven Treasures ( ). In chapter 18 (The Benefits of Responding with Joy), juan 6: 9 k(cid:3)e˘ ŋɔ(cid:4) naw (cid:4)y(cid:5)(cid:5) ləw n˘e(cid:4) n(cid:25)(cid:3)ik(cid:25)(cid:5)n ˘s(cid:27)kh(cid:25)(cid:27) l(cid:5)(cid:4) t˘s˘(cid:7)(cid:25)əw (L66)(R42) (R20)(L69) (R38)(L39) (R60)(L32) (L28)(L28) (L29)(R40) In chapter 11 (The Emergence of the Treasure Tower), juan 4, (pearl) and (carnelian) appear in place of (coral) and (amber), respectively. In chapter 1 (Introduction), the transliterated words (mani jewels) and (crystal) are used follows: n(cid:25)(cid:5)i (cid:4)y˘ε l˘υ(cid:4) nɔ(cid:4) mɔ(cid:4) (cid:2)i phɔ(cid:4) li (R60)(L34) (R6)(R42) (R42)(L10) (L49)(R9) The Lotus Sutra, of course, contains a wealth of Buddhist terms. Most of them are semantic translations from Chinese words. The following is the example of the six påramitås( ): 228 XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) (cid:6)i(cid:4) m(cid:4) ke˘ (cid:4)ie(cid:4) p˘(cid:7)ɔ(cid:4) (cid:6)˘(cid:7)ew khu d(cid:6)ε(cid:4) (cid:7)˘(cid:7)an nd˘(cid:7)en sε(cid:3) ? (L11)(L14) (L9)(L39) (L51)(R41) (L1)(L35) (L26)(R37) (R54) On the other hand, the word påramitå itself takes a fixed form in Xixia, pɔ(cid:4) lɔ(cid:4) m(cid:4), a transliteration of the Chinese term. Rulai (Thus Come One, tathågata) is always rendered m˘(cid:7)or (L90) "ε(cid:3) (L61), which may be literally translated into Chinese as shilai (Truly Come One). The Xixia word is thought to have been translated not from Chinese but from Tibetan yang-dag-par gshegs-pa, which is assumed to be an older translated term. Shizun (World-Honored One, bhagavat) is rendered in various ways including with the compound of r˘(cid:7)ur (L76) p˘(cid:7)u(cid:5) (L59), meaning “world” and “respect”; ( ), meaning “[one] tri- umph[ed over] possessing and passed [over],” a replacement of Tibetan bcom-ldan-˙das; and phɔ(cid:4) khɑ(cid:4) x(cid:25)an and mbɑ(cid:4) (cid:4)ɑ(cid:4) mbɑ(cid:4), replacements of bojiafan , a Chinese transliteration of San- skrit bhagavat. The latter transliteration, mbɑ(cid:4) (cid:4)ɑ(cid:4) mbɑ(cid:4), which is found in Molitianmu zongchi , is assumed to be a new term closer to the Sanskrit original. The term d(cid:6)an (R23) t(cid:7)h˘(cid:7)u (R2) is used for translating zhongsheng (living beings, sattva) throughout each volume of the Lotus Sutra. This form is equivalent to youqing , the term adopted by Xuanzang , meaning “[those] having sentience.” Some Buddhist scriptures adopt se(cid:4) (R33) nd˘(cid:7)u(cid:4) (L3) (qing-you , sen- tience-having). This is similar to the manner of translating pusa (bodhisattva) d(cid:6)an-tse(cid:4)(R33) “those perceiving sentience.” (Cf. WrT. sems-can, byang-chub sems-dpah.) These Buddhist terms were used in translating Chinese classics. For example, the passage from the second chapter, Yingxu , of Liutao « », “ ” (prohibiting evil and falsehood by moral stan- dards…) is rendered in Xixia as “ ” ( , breaking and subduing evil and falsehood by honest and upright intent). Zhishi , xiewei and cuifu are terms used in the Lotus Sutra: t(cid:7)h˘(cid:7)a (R16) tu(cid:5) (L58), dzəw (R38) l(cid:25)o(cid:5) (L94), k(cid:25)(cid:5)n (L32) (cid:6)˘(cid:7)ew (R41). In Zhenshi miao jing , zhenya is equivalent to cuifu . The terms in Liutao appear to have been devised in a fairly elaborate XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) 229 manner. Three different equivalents are allocated to the character shou in liushou , shoutu and shouguo , respectively: t(cid:7)˘(cid:7)ew(L46) (cid:4)˘(cid:7)e(cid:4)(L39) (six kinds of shou : ren , yi , zhong , xin , yong , mou ) lu(R1) (cid:4)˘(cid:7)ε(cid:3) (L34) (defense of the national territory) 2 hle(cid:4)(R?) (cid:6)on(R47) (maintenance of the state) Though one can understand that these three characters, (cid:4)˘(cid:7)e(cid:4), (cid:4)˘(cid:7)ε(cid:5) , 2 (cid:6)on, i.e., equivalents for shou , are interrelated, it is difficult to con- cretely determine the differences among them. In the Lotus Sutra appear examples such as and (cid:6)on(cid:6)ε(cid:5) (L61) . 4. The Systematic Classification of the Xixia Vocabulary The task of systematically classifying Xixia vocabulary has not yet been completed. There are several major issues to resolve. The Xixia poems, Yueyue leshi « » and Dashi « » contain two conspicuous lay- ers of vocabulary.21This reflects the heterogeneous composition of Xixia society—the royal family, the nobility, and the Mi and the Minyak eth- nic groups. At present, I classify the two layers as layers I and II. The vocabulary of layer I, which is assumed to be the layer of the elegant vocabulary, cannot be identified yet. Layer II is that vocabulary usually regarded as the Xixia language. This layer consists of Buddhist terminology as well as vocabulary in day-to-day use, as represented by the words from Fanhan heshi zhangzhong zhu « », a collection of Xixia-Chinese words and phrase examples. Loanwords (transliterated words) from Chinese and Tibetan are easy to distinguish. Numerals and pronouns are also distinguishable, as is the terminology of sets such as shi gan (the ten calendar signs), shi’er zhi (the twelve zodiac signs), shi’er zhi (twelve fortune signs), shi’er yuansheng (the twelve-linked chain of dependent arising), shi’er xinggong (the twelve constellation signs) and ershi ba xiu (the twenty-eight constellations). The identification of these sets of words requires meticulous compari- son between various documents. From an investigation of several Bud- dhist scriptures, including Shi’er yuansheng xiangrui jing « » (Twelve-linked Chain of Dependent Arising Good Omen Sutra), shi’er yuanshengappears as:22 230 XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) mb˘(cid:7)u me(cid:4) d(cid:6)(cid:4) se(cid:4) m˘(cid:7)e(cid:4) ts(cid:27)(cid:5) t(cid:7)h˘(cid:7)ew ndɔ(cid:4) ts˘(cid:7)u(cid:5) ? dzu (L2) (L36) (?) (R33) (R35) (R68) (L46) (R42) (L59) (L1) hl(cid:25)i(cid:4) nd˘(cid:7)u(cid:4) w˘e(cid:4) tɑr(?) s(cid:5) (L11)(L3) (L12) (R73) (L30) (This term appears in the Xixia Lotus Sutra, but is unfortunately missing in the St. Petersburg text.) I have been able to extract the set shi’erzhi (twelve fortune signs) from Xinyi Tongren zhenjiu dian « » (A New Translation of the Bronze Figure Standard for Acupuncture and Moxibustion).23 It has proved more difficult than expected to find examples of numerals. Numbers greater than one hundred million rarely appear. Unfortunately, neither (one hundred billion) m˘(cid:7)e(cid:5) (R68)24 nor a special synony- mous form of one thousand, rɑr (L80) -nɑw (R19),25 is found in the Xixia Lotus Sutra. However, fractional numbers appear in chapter 15 (Emerging from the Earth), juan 5, as follows:26 "(cid:5)(cid:5)r mb(cid:25)aw khɑ(cid:4) (cid:4)ɑ mb(cid:25)aw rir kh(cid:5) mb(cid:25)aw khɑ(cid:4) ləw (L92) (R20) (L17) (?) (R20) (R72) (R28) (R20) (L17) (L43) one portion of four separated portions one of yi (100,000,000) wan (10,000) portions one quarter (1/4) one trillionth (1/1,000,000,000,000) The combination and ordering of the cardinal directions (east, west, north, south) is a topic of interest. w(cid:5)(R61) le(?) ʁz(cid:5)r2(L86) la(cid:5) (L64) in the Xixia Lotus Sutra are literal translations from the Chi- nese counterparts, and the order follows that of the Chinese text. The proper order of directions in Xixia is thought to have been clockwise starting from east as in Fanhan heshi zhangzhong zhu « »: , east, south, west, north. The intermediate (ordinal) directions between the four cardinal points are also translated clockwise from southeast in the Lotus Sutra in accordance with the Tibetan. Incidental- ly, Huayan jing « » (AvataµsakasËtra) adopts a literal translation from the Chinese based on the axis of east and west.27 Chinese (east-south) (west-south) (west-north) (east-north) XIXIALANGUAGESTUDIESANDTHELOTUSSUTRA(II) 231 Lotus Sutra ( (south-west) (north-east) Huayan jing Tibetan shar-lho lho-nub nub-byang byang-shar (east-south) (south-west) (west-north) (north-east) Although many personal pronouns appear in the Lotus Sutra, including the first person ŋɑ(cid:4)(R14), m˘(cid:7)ɔ(cid:4)(R44), it is difficult to determine how mutual level relationships among them are really set up by examin- ing of the usages in text. However, I note the following: ŋ&a(cid:4)(R17) “to me” ŋ&a(cid:4)-mi(cid:4)(R10) “we two” and ŋ&(cid:5)(R28) -mi(cid:4)“we two” na(cid:4)(R17) “you” and na(cid:4)-mi(cid:4)“you two”28 It should be said that these above-mentioned forms are specific exam- ples, rarely found in texts other than the Lotus Sutra, as are the follow- ing examples of the personal pronoun, genitive, first person: ŋ(cid:25)(cid:27)(cid:4) (R25) “my,” “my son” and “my stature.” Not a few examples of such specific grammatical features appear in the Xixia Lotus Sutra. Inserted Notes in Buddhist Scriptures A small fragment titled… Nianguan men (the beginning portion is lost) is included among the Khara-Khoto documents in the Aurel Stein collection. In the text the word t(cid:7)˘(cid:7)a (L19) dzυ (R5) appears with an inserted note explaining that it means (to put the left leg on the right leg). It is equivalent to the Chinese, jiezuo , jiejia fuzuo “cross-legged sitting,” probably meaning ban jiezuo “half-cross-legged sitting” (jixiang zuo “auspicious sitting”). Such explanations in inserted notes are not rare, enabling the readers to understand the text’s meaning correctly. Explanations of extant phonological works Explanations in the rhyme lexicon Wenhai (Baoyun) « ( )» pro- vide many suggestions for verifying the meaning of the words. For example, təw “egg” belonging to level-tone rhyme 43 is given the following explanation: