TangutReader1st draft

[键入公司名称] Tangut Reader with Introduction Kirill Solonin 2022/4/10 [在此处键入文档摘要。摘要通常为文档内容的简短概括。在此处键入文档摘要。摘要通常为文 档内容的简短概括。]  1 Introduction........................................................................................................................................... 3 1. Disclaimer ................................................................................................................................. 3 2. Purpose of this Reader .............................................................................................................. 3 3. Historical matters ..................................................................................................................... 6 4. Texts in this Collection ............................................................................................................. 7 5. Ambiguity in interpretation ..................................................................................................... 9 6. Tangut Grammar in a Nutshell ................................................................................................. 15 6.1. Numerals ............................................................................................................................. 15 6.2 Pronouns. .................................................................................................................................. 16 6.2.1 Personal pronouns ....................................................................................................... 16 6.2.2Indicative pronouns .............................................................................................................. 17 6.2.3 Interrogative pronouns and interrogative words ....................................................... 20 6.3 Nouns .................................................................................................................................. 23 6.3.1 Patterns of Nominalization ................................................................................................... 24 b. Verb substantives with śjij1 . “Manner of action”. ............................................................ 24 c. Verb substantives with tjị2 ............................................................................................. 25 d. Verb substantives with lew2 ............................................................................................. 25 e. Verb substantitives with sji2 ................................................................................................. 26 6.3.2 Nominal Morphology.............................................................................................................. 27 a. Number ................................................................................................................................... 27 b. Case markers .......................................................................................................................... 28 b.1. Subject (NOM) ..................................................................................................................... 29 b.2. Object Marking. Morpheme ·jij1 ........................................................................................ 29 b.3. Instrumental .......................................................................................................................... 30 b.4. Comitative/ Ablative ............................................................................................................. 32 b.5. Tangut Ergative .................................................................................................................... 32 b.6. Locative morphemes ............................................................................................................. 33 6.3.3. VERB .................................................................................................................................... 37 a. Structure of Tangut verb. ...................................................................................................... 37 b. Verb Prefixes ...................................................................................................................... 38 7. Verb modifiers and periphrastic constructions ..................................................................... 45  2 7.1 Verb modifiers and periphrastic constructions ............................................................................ 45 8. Negatives .................................................................................................................................... 50 9. Tangut modal verbs and markers .......................................................................................... 52 10. Tangut verb agreement ...................................................................................................... 53 SYNTACTIC PATTERNS................................................................................................................ 58 1. Phrase connectors/ conjunctions ............................................................................................ 60 Declarative and interrogative sentences ........................................................................................... 60 1. Texts with full analysis ........................................................................................................... 69 Attack on the town of Yuan ......................................................................................................... 69 Perfection of the Self ................................................................................................................... 71 Reading Materials ........................................................................................................................... 78 Atiśa goes to Tibet ....................................................................................................................... 78 Bodhidharma’s Four Practices .................................................................................................... 80 Introduction to the Sea of Meanings Established by the Saints ..................................................... 81 Judge Li Li .............................................................................................................................. 83 Nāgārjuna and Savari ................................................................................................................... 83 Preface to the Five Sūtras............................................................................................................. 84 Vairocana searches for the Teaching ................................................................................ 86 Fada asks for the teachin g ......................................................................................... 87 Painting an extra leg to the snake ................................................................................... 89 Recommending the Wise ............................................................................................................ 90 Fu Ziqian ............................................................................................................................... 91 Preface to the Rec orded S ayings of Nanyang Huizhon g ................................. 91 Preface to the Wuyin qieyun ................................................................................................ 92 Tian Tan as a comm ander ................................................................................................ 93 Playing in water ..................................................................................................................... 94 Talking with an im mortal .............................................................................................. 95 Ye Gong loves dragons ............................................................................................................ 95 Yue Shifu................................................................................................................................... 96 Zhi cheng aske s fo r i nstructi on ........................................................................................... 96 Youcai Guoshen .......................................................................................................................... 97  3 Introduction 1. Disclaimer The value of Tangut studies is twofold: in terms of comparative and historical linguistics Tangut occupies a peculiar position within the Tibeto-Burman group, providing rich materials for cross-linguistic studies, including deep-level reconstructions of Sino-Tibetan. From a historical perspective, Tangut preserved a rich cache of texts, mainly of Buddhist subject matter, which allow reassessment of the history of Central Asia during 10-13th centuries. Particularly important are the Buddhist texts which only survive in Tangut. The cache of excavated texts associated with the Tangut state in various languages, including Tangut, Chinese, Tibetan is second only to the Dunhuang findings, and its amount far exceeds the written heritage of other Central Asian peoples, including Uighur, early Mongolic, Kitan, Jurchen, etc. Tangut native literature is a trove of knowledge of Tangut own culture. This is particularly important since Tangut preserved probably the richest literary heritage among the Tibeto- Burman languages, second only to Tibetan. Importantly, the available Tangut texts belonging to the realm of both Sinitic and Tibetan Buddhism are chronologically older than most of currently available editions of the texts, from which the scholarship traditionally works. Thus the study of the Tangut sources is helpful not only for the Tangut studies per se, but opens a broader perspective in various fields of language and historical studies. In this repsect it is particularly regretful that Tangut materials still remain partially neglected by the historians of Buddhism, ethnologists, etc. Since the Tangut culture emerged as a combination of indigenous Tangut, Sinitic and Tibetan elements, the Tangut studies should be viewed as a complex multi-dimensional field, which requires equally good familiarity with Tangut, Chinese, Tibetan languages, history of Buddhism and principles of linguistic reconstruction and philological research. Most of current scholars, including present author do not fully qualify for this. This dictates a necessity to compose a generally available textbook or a Reader for Tangut, so as to facilitate further growth of Tangut research world-wide. Tangut language is now taught in several universities, mostly in China, therefore there is a need to standardize the teaching practices. This Reader is one step in this direction. Currently, there is a rich plethora of native Tangut lexicographic and phonetic materials which facilitate the study of the language. Among these, one should mention the so-called Homophones ɣiẹ1 lew2 , the Synonyms ·wo2 lew2 , the Sea of Written Signs ·jwɨr2 ŋjow2 ,the Rhymes of the Five Sounds ŋwə1 ɣiẹ2·we2 bju1 ,Tangut texts with Tibetan transcriptions. These probably are remainders of a once symmetrical system of lexicography, which included other compositions, such as the Sea of Characters dji2 ŋjow2 etc. Only titles of these compositions survive. Surviving Tangut encyclopedia The Sea of Meanings Established by the Saints, is probably also a remainder from this system. Introduction of Tibetan and probably of original Sanskrit materials around mid-12th century produced a variety of transcriptional materials, including Sanskrit mantras, dhāraṇī, personal names of Tibetan and Indian masters, transcribed titles of compositions, etc. Educational needs conditioned compilation and publication of bilingual “readers”, widely distributed within the Tangut realm. The only surviving one is Tanguto-Chinese Timely Pearl in Hand mji2zar1 dzjɨj1bju1pjạ1 gu2 nji composed by Gule Maocai 骨勒茂才. This rich cache of materials and their relatively easy accessability facilitated fast progress in initial reading of the Tangut texts. However, the Tangut materials are not yet sufficiently incorporated into the mainstream of Chinese or Tibetan studies. 2. Purpose of this Reader This Reader and its Introduction do not intend to present a comprehensive academic description on the Tangut language. We are not qualified for this. In this stead we intend to provide a set of practical recommendations for those interested in the language primarily from the perspective of reading and understanding the texts. Therefore  4 we do not discuss the matters of historical and comparative phonology, etymology and subgrouping. As the reader will notice we do not draw clear disctinction between Tangut morphology and syntax, as was once done by K. Kepping. This is generally due to our understanding of the Tangut “auxiliary” morphems as multi-fucntional: i.e. depending on its position, similar morpheme can be used as a meaningful word, or treated as a morphological or syntactic marker. The problem of how these homophones and homographs relate to each other historically remains outside the scope of this Reader. The pathway of grammaticalization in Tangut is not clear: for most of the “auxiliaries” we cannot establish their orginal lexical meaning, whereas syntactic and morphological dimensions are identifiable. Tangut is a Tibeto-Burman language, spoken by the dangxiang 党项 people and the multinational population of the Great State of White and High phiow1 bjij2lhjịj tha2 , otherwise known as Xixia during 11-13th centuries. The absolute chronology of Tangut separation from other TB languages cannot be established with any certainty: what is now known as Tangut language gradually evolved during several centuries until it was finally put into writing in the beginning of 1040-s, although the invention of the writing system itself might date to an earlier period. Tangut language remained in circulation during the Yuan period, and to certain extend during the Ming dynasty. Surviving Tangut native lexicography belongs to the 12th century, generally to the period of a reform of the Tangut literary heritage which took place closer to the end of the century. As observed from early on, e.g. by M. V. Sofronov, versions of the language preserved in the earlier and later versions of the famous Tangut dictionary The Homophones demonstrate mutual deviations, which can be interpreted either as traces of the linguistic evolution, or as the dialect influences. In general, application of comparative method to the available Tangut materials yielded certain results allowing subgrouping Tangut together with various members of Rgyalrong (e.g. Western Rgyalrongic, as recently suggested by G. Jacques and others), or, alternatively within the Qiangic language group (as maintained by Sun Hongkai 孫 宏開). In any case, the evidence indicates towards the Tibeto-Burman languages, and earlier suggestions of an Altaic affiliation remain a curiosity. Matters of genetic and/ or typological subgrouping are unrelated to the reading of the texts, and will not be discussed further. The two resettlements of the dangxiang tribes in the “inner parts”, ended up in the area now known as Ningxia Hui Autonomous Region of the People’s Republic of China. By the end of the 8th century various dangxiang tribes inhabited the areas which later became the Tangut state. Long-term contacts with a variety of Chinese dialects left traces in the Tangut vocabulary. As some scholars argue, Tangut contact with other Tibeto-Burman speaking tribes left traces in Tangut verb morphology as well. Long term contacts with the Chinese language are demonstrated by the fact of existence of the so-called “old” and “new” Chinese borrowings, identified by Hwang-cherng Gong 龔煌 城. Among the criteria for separation between the two classes of borrowings was the retaining or loss of the voiced consonants: borrowings from the so-called North-Western dialect supposedly reflect the merger between voiced and unvoiced stops, which took place around 9-10th centuries. Earlier borrowings retained voiced stops, representative of a different or earlier version of Chinese. A set of sound correspondences allowing to discriminate between the different levels of borrowing is presented in concise form by Nie Hongyin 聶鴻音 in his recent publication.Degree to which linguistic environment influenced Tangut language prior to its standardization in writing remains debated. Surviving Tangut lexicography and study of the texts confirm that Tangut vocabulary is comprised of three parts: one is the indigenous Tangut vocabulary; the other one is formed by various Chinese borrowings and the last one is based on the translations of the Tibetan Buddhist terms, and is thus of limited distribution. Tibetan vocabulary is limited to the translations of Buddhist texts, thus the degree to which it had influenced other layers of vocabulary remains to be determined. Words which can be considered Tibetan borrowings in true sense are extremely scarce, and generally belong to the realm of the so-called “Sino-Tibetan Buddhist texts”. Chinese borrowings are less easy to determine, considering their distribution between all layers of language, and the nature of the Tangut lexicography: since most of Tangut character can be rewritten in Chinese characters, this creats an impression of the total domination of the Sinitic vocabulary in Tangut. That this is not true was realized rather early by both G. Morrisse and Luo Fuchang 羅福萇, who mentioned that there is no one-to-one correspondence between the Chinese and Tangut charcters. However, important part of the Tangut cultural vocabulary is definitely of Chinese origin, some of that related to the local dialect spoken in the Ningxia area during 11-12th centuries. From a grammatical perspective, Tangut texts demonstrate proximity with written Tibetan; however, this similarity  5 is explained by the nature of the textual evidence, and general typology, but not necessarily by a genetic relationship. In any event, as the translations from Chinese demonstrate, Tangut tended to use long sentences, combining separate Chinese clauses into long sentences with the help of various conjunctions, which resembles Tibetan much more than Chinese. Comparison between translated texts and native Tangut compositions, including the Tangut law codes, demonstrates that all layers of textual materials account for one language, which we identify as “Tangut” or “minyag” (mji2 ŋwu̱1 ), contrasting with the so-called “odic” (or “ritual”) Tangut, which is exempt from present discussion. We generally share Nie Hongyin’s take on this language as unrelated to the standard Tangut at all. This Tangut Reader and its Introduction are dependent on the previous scholarship in every department. The purpose of this publication is modest: we do not intend to provide an exhaustive description of the Tangut language, since it was done before, by the scholars more qualified than us, including M. Sofronov, K. Kepping, Nishida Tatsuo, Hwang-chenrng Gong, and lately G. Jacques and Arakawa Shintarō. Therefore the improvement of these descriptions is the responsibility of the scholars who suggested them first.1 Recently, a comprehensive introduction into the Tangut language was published by Nie Hongyin2. In this publication various aspects of the language are presented in a concise and easily accessible format. This publication is more comprehensive than this Introduction, since it dwells on a broader variety of problems, whereas we limit ourselves to offering assistance for reading the texts. We intend to specify a set of grammatical markers which will enable readers to apply punctuation marks to the Tangut texts and analyze them in such a way as to arrive to a more or less correct understanding of their content. Currenly we cannot account for all phenomena observed in various texts with equal accuracy and for some not at all. However, correct punctuation will reveal the sentence structure, and once syntactic relations are established, one can further proceed with the translation. Despite the practice of substituting Chinese characters instead of the Tangut ones is popular in current scholarship, we deem this method as essentially erroneous, since it only creates an illusion of understanding, as many modern publications clearly show. However, when applied with care this method might still remain useful for some initial observations. Several years of teaching Tangut both by myself and the colleagues in other schools demonstrated that there is a need for a comprehensive grammar book on Tangut, and even more so, for an anthology of annotated texts to be used in the teaching practice. Existing accounts tend to formal description of the language. In our opinion a set of tips helpful in reading texts is more needed by those who do not engage in linguistic research, but rather treat Tangut texts as source materials in other fields, such as history, Buddhist studies, etc. Attempt to compile a 1 Various aspects of Tangut language were discussed extensively since the inception of the Tangut studied in early 20th century. The first attempt to deal with the matters of morphology and syntax were undertaken by G. Morisse based on the Tangut translation of the Lotus sūtra (Contribution préliminaire à l'étude de l'écriture et de la langue Si-Hia. Extrait, des Memoires prsentes par divers savants a L'ACADÉMIE DES INSCRIPTIONS ET BELLES-LETTRES, 1 serie, tome XI, IIa partie (Paris: Impremiere Nationale, 1904). Before the discovery of Khara- Khoto, this publication alongside several other works by G. Deveria, S. Bushell, etc., remained the fundamental source on the Tangut language. Following the publication of the Tangut Chinese Timely Pearl by A. I. Ivanov, B. Laufer published his The Si-Hia language: A Study in Indo- Chinese Philology, (Leiden: E.J. Brill, 1916), which remained unsurpassed before N. A. Nevskij’s studies of the Tangut language, published in 1960 (i.e. Tangutskaya Filologia, vol. 1-12, (Moscow: GRVL, 1960). Publications by Nevskij, were followed in the Soviet Union by M. Sofronov, Grammatika Tangutskogo Yazyka, vol. 1-2 (Moscow: GRVL, 1968); and Nishida Tatsuo 西田龍雄, Seika go no kenkyū 西夏語の研 究, (Tokyō: Zauhō kankōkai, 1964). Publication of these works marked the transition between the stage of decipherment of the language and actual reading of the texts. During the 1970-s and 80-s, following the publication of bilingual Tangut-Chinese texts, and native texts of Tangut lexicography (especially Wenhai), the familiarity with the Tangut language grew extensively. This stage culminated in the publication of the Tangutskyi Yazyk: Morphologia by K. Kepping in 1985 (Moscow: GRVL), which was the first work to present systematically Tangut verb paradigms, system of the Tangut postpositions, etc. Shortly thereafter, Hwang-cherng Gong 龔煌城, published series of articles dealing with various aspects of the Tangut language, and offered alternative system of phonetical reconstruction for the Tangut language. (Hwang-cherng Gong 龔煌城, Xixia yuyan wenzi yanjiu lun ji 西夏語言文字研究論集, (Beijing: Minzu chubanshe, 2004). Gong’s “Xixia yu gaikuang” (西夏 語概況，Description of the Tangut Language), along with several introductory papers by Nishida Tatsuo, collected in his posthumous volume from 2012, remain fundamental, yet concise, introductions into Tangut language. Gong’s views on the Tangut phonology and structure of the Tangut language, are generally accepted by the scholarly community, however, some of his observations are challenged. More recently, attempts to provide a complete systematic description of the Tangut language were undertaken by G. Jacques (Esquisse de phonologie et de morphologie historique du tangoute, (Leiden: Brill, 2014), which is a comprehensive study of all aspects of the language from historical perspective; and Arakawa Shintarō 荒川慎太郎, Seika bun Kongo kyō no kenkyū 西夏文金剛經の研究 (Tokyō: Shoukadoh, 2014) based on his research into the Tangut textual corpus related to the Diamond sūtra, which is an attempt of a synchronous description of the Tangut language on the basis of a group of several texts with common subject matter. These works constitute the basis for the current understanding of the Tangut language. 2 Nie Hongyin 聶鴻音, Xixia wenzi he yuyan yanjiu daolun 西夏文字和語言研究導論 （Shanghai: Shanghai guji, 2022）  6 resource book on Tangutology was recently undertaken by Shi Jinbo 史 金波 . 3 This book can serve as an introduction to the general problems of Tangutology but is hardly acceptable as a textbook to be used in class. What follows below, is a brief exposition of the foundations of Tangut grammar, with which one must familiarize himself to facilitate his/ her first steps in reading Tangut. It is fully descriptive, not analytic, or prescriptive; solely based on the texts studied so far by myself and colleagues. We follow “philological” and not “linguistic” approach, implying that subtle (and not so subtle) nuances of linguistic analysis of Tangut do not have direct relation to reading the texts. Thus, linguistic observations were reduced to a necessary minimum, and no historical and comparative materials had been added herein. What follows in not in any way a research on the Tangut language, or on any of its aspects, either synchronic or diachronic. It is rather a systematized collection of tools to help students in their first steps in Tangut. Being an introductory work, this Reader does not observe Leipzig rules for recording the example sentences; we provide only the original and its translation with very rudimentary analysis where necessary. The structure of this Reader is based on the interpretation of the example sentences. Mostly we proceed from the texts either translated from Tibetan, or dependent on the Tibetan subject matter. When the original of the Tangut translation is established we provide references to the Tibetan texts, which are generally to be found in: Garuda in Flight: The Spread of the Old and New Secret Mantras in Xixia4. When this is not possible, we limit ourselves to the sentences of whose translation we are confident. 3. Historical matters Attempts to read Tangut date back to the early 20th century. Original translations by N. Nevskij from the 1930-s, although not as numerous as one would like these to be,5 are indicative that he was on the right track in reading the texts, even though many of his readings proceeded from the educated guess-work, and not from a meticulous linguistic analysis, which was hardly possible at his time. In many respects the situation has improved sufficiently since Nevskij’s time: aspects of Tangut morphology such as the Tangut system of verb prefixation, Tangut verb agreement and the so called “ablaut”, which remained virtually unknown to Nevskij, received scholarly attention. All these research results are of enormous import for the advancement of our ability to read and understand Tangut. What remains mysterious in Nevskij’s readings of the Tangut texts is that his surviving research materials (both published and preserved in his archive) demonstrate obvious mistakes in the analysis of what he called “auxiliary particles”. This term is based on his translation of the Tangut original term, dạ2 bjịj2 , “auxiliaries for speech”. However, his translations demonstrate that he understood them correctly, more or less in the same way as we do now. This is one of many mysteries of Nevskij’s scholarly heritage. These “auxiliaries”, some of which first specified by Nevskij can be to a certain degree equated with the so called “empty words” xuci 虛詞 of classical Chinese, whereas Nevskij preferred Japanese parallels. Such equations have limited explanatory power, since much of Tangut “auxiliaries” have no parallels in either Chinese (most) or Japanese (some). Following the work of Nevskij, M. Sofronov, Nishida Tatsuo and Hwang-cheng Gong published their versions of Tangut grammar. Their publications allowed reading Tangut with more certainty than before. Following advances in the study of Tangut phonology, morphology and syntax were made by G. Jacques and Arakawa Shintarō 荒川慎太郎, and most recently by Nie Hongyin. Generally, phonetic reconstructions of Tangut by Nishida Tatsuo, Hwang-cherng Gong, M. V. Sofronov, Arakawa Shintarō and lately by Nie Hongyin, are steps in the right direction, especially from the perspective of the consonantal system. The reconstructions of the vowel system are less reliable. The questions pertain to the actual phonetic quality of the so called “grades” (deng 等) in Tangut6, presence of tense/ lax and long/ short vowel 3 Shi Jinbo 史金波 Xixia yu jiaocheng 西夏語教程, (Beijing: Shehui kexue wenxian, 2013). 4 Suo Luoning 索羅寧, Dapeng zhanchi: Zangchuan xinjiu mizhou zai Xixia de chuanbo 大鵬展翅：藏傳新舊密咒在西夏的傳播 (Shanghai: Shanghai guji: 2022, in print); when the examples from the Sintic translations are used, these are borrowed from: Suo Luoning 索羅寧，Xixia wen Hanchuan fojiao wenxian yanjiu 西夏漢傳佛教文獻研究 (Lanzhou: Gansu wenhua, 2022, in print) 5 Nevskij’ s early translations are collected in the first volume of his Tangut Philology [Тангутская филология], vol. 1-2 (Moscow: GRVL, 1960) 6 Gong believed that Tangut contrasted three grades, i.e. syllables without medial –i- (1st grade), with –i- (2nd grade) and with medial –j- (3rd and 4th grades).  7 opposition, tonal structure of the language, etc.7 This especially true for the long/ short vowels contrast: surviving materials do not systematically support such an opposition as inherent in the Tangut language, but rather show that it emerged as a secondary development dictated by the contact with Sanskrit materials. Tangut rhyme dictionaries include only “even” and “rising” tones, whereas there are indications that Tangut in fact had four tones, including the “departing” and “entering” tones. Much of Tangut phonetics, morphology and syntax require more research than has been done so far, and even less had been included here. We continue to use Gong’s version of Tangut phonetics not because is more correct than others, but only because it accounts for much of the morphologically, grammatically and semantically meaningful sound alternations observed in Tangut, and presents a typologically acceptable symmetrical system. Gong’s system offers an easier version of orthography as compared with Sofronov and Nishida. Another reason is that the reconstruction by Hwang-cherng Gong was popularized by Li Fanwen 李 範文 in his Tangut dictionary8. One should however keep in mind that Gong’s system is more “generally accepted” than it is “correct”. Both Sofronov and Kepping and later Hwang-cherng Gong proceeded from an understanding that Tangut grammar can be modeled after a scheme ultimately based on the Indo-European languages. Similarly, Nishida and Arakawa modeled their version of Tangut grammar on Japanese. In case of the Japanese scholars, Tangut case markers are generally equated with the respective Japanese particles, as was originally suggested by Nevskij. In any case, the currently available reconstructions of Tangut grammar utilize such categories as “parts of speech”, “tense system” and “case system”. This scheme is generally reproduced by Arakawa Shintarō in his exposition of Tangut grammar. Here we do not suggest any new approach, but try to utilize the results of previous scholarship. Nie Hongyin approach, as he himself formulates, proceeds from the traditional idea of “auxialiary words”, from which he excludes verb prefixes, but includes adverbs. This approach allows escaping the difficulty of formally defining the nature of Tangut auxiliaries, but at the same time makes the exposition a collection of ad hoc observations. In case of this Reader and its Introduction, since we do not intend to provide an academic study of the language, ad hoc objection probably will not stand: any text is just a collection of ad hoc sentences, for which we hope to discover some easy rules. Our understanding is such that the current state of the field of Tangutology allows reliable readings of the Tangut texts: although the problems remain, their nature is such that they do not impede adequate understanding of both native and translated texts, especially when historical context in a broad sense is provided. The problems which persist are in many ways of lexical nature: that is, despite the overall structure of a sentence being transparent, scholars are sometimes not able to adequately understand a Tangut word, which leads in many cases to awkward and clumsy translations. This is especially true in case of the famous Tiansheng Law Code, which despite severa decades of intensive research remains in many respects obscure. This will obviously be remedied with the passage of time and accumulation of empirical knowledge. Another problem for understanding the Tangut texts lies in the ability of current scholarship to handle Tangut texts historically and philologically. Most scholars doing Tangut studies are trained as Sinologists, therefore their familiarity with Tibetan is limited at best, and their command of the Buddhist subject matter is virtually inexistent. Inability to understand what the texts are about invokes all sorts of wild ideas, which impede linguistic analysis. Younger generation of Tangut scholars is better qualified for such reconstructions. 4. Texts in this Collection Scholarship generally recognizes two layers of the Tangut language. One is the so called mji2ŋwu̱1 (miyu 彌語, or fanyu 番語)，which is understood as the language of the dangxiang 党项 people. Most of the extant Tangut texts are in this language. Another layer is the so called “odic” (“ritual”) Tangut, otherwise known as the language of the lhjwịj1 ,the language of the lhi 勒 people as suggested by Nie Hongyin.9 Ambiguity of the Tangut ethnic denominations, as well as scarcity of source materials impede correct identification of the “lhi 7 For critical discussion of Hwang cherng Gong reconstruction see: Li Fanwen 李範文 eds., Xixia yu bijiao yanjiu 西夏語比較研究 （Yinchuan: Ningxia Renmin, 2000）, paragraphs on Tangut phonology written by Nie Hongyin. 8 Li Fanwen 李範文，ed. Jianming XiaHan zidian 簡明夏漢字典 (Beijing: Zhongguo shehui kexue chubanshe, 2011) 9 Nie Hongyin 聶鴻音, “Leni yu: yizhong weizhide gudai Mianzang yu” 勒尼語：一種未知的古代緬藏語, Ningxia daxue xuebao 寧夏大學 學報，4 （1996）  8 language”, which was at different times identified as the Tangut “ritual language”10, or associated with the famous division of the Tangut people into the “red faces” and “black heads”, or the language of the ruling elite of the Tangut state, or a language of a different tribe, which took contraol over dangxiang population, etc. Texts in this language are not numerous, and generally belong to the genre dzjo2 , generally translated as the Chinese shi 詩 or “odes”. This Tangut Reader is limited to the texts in the standard Tangut. Despite genre differences, we are confident that most of the texts, which we deem as belonging to the realm of the standard Tangut attest for the same language, that is, the rules and patterns identified for one text, are relevant for other texts as well. Famous “Preface to the Rhymes of the Five Sounds” contains following paragraph: … Sūtras of the Primordial Buddha, the Law of the Kings, Laws and Commandments, Odes of the Learned People, Pure and Muddy, pairs of Yin and Yang, good and evil omens, seasonal records, Teaching of Salvation (Daoism), Art of Healing, calendars and divination, poetic collections and all that, are the sources from which wen (文) rises. Despite ambiguity of terminology, almost all categories in the above list find their matches among actual excavated texts. However, the selection for this Tangut Reader was based not on actual Tangut understanding of their culture, but rather on the needs of teaching. Thus, the selected texts concentrate on the prose extracts and the Tangut poetry is generally excluded from the selection. Previously, teaching of Tangut language proceeded in most cases through the mediation of Chinese or Tibetan. Although these might be useful devices, Tangut is a language in its own right, thus its study can proceed without linguistic references to either Chinese or written Tibetan. The knowledge of these two languages is imperative for understanding and interpreting cultural vocabulary, and not for the sake of the language itself. Of course, knowing only Tangut, a scholar would not be able to achieve any serious advances in his/ her research, either in linguistics or in historical or religious studies. Using this Tangut Reader requires the ability to use Li Fanwen XiaHan Zidian 漢 典 and familiarity with one of now numerous Tangut input systems.11 Some familiarity with Tangut phonology might be useful, but is not imperative, due to the reasons stated above. At the same time, we presume that the users of this Reader are the students of either Chinese or Tibetan departments. Before using this Reader, one needs to have a general idea of the Tangut language: we originally hoped that this Reader could have been used for the self-study purposes, but now came to understanding that the services of a professor with at least rudimentary idea of Tangut will be required. Most of the texts are extracted from the Tangut composition of zashi 雜史 genre known as The Twelve Kingdoms (十二國, ɣạ2njɨ̱ 1lhjịj ), which is a collection of the “historical precedents” from the Chunqiu and Zhanguo periods. This selection is determined by the personal reasons: The Twelve Kingdoms is the text with which I am most familiar. This text belongs to the same category as more well-known djịj1bo1 Grove of Categories 類林,therefore grammar patterns with which the user of this Reader might be familiar with from reading the translations of these texts, are applicable here. Other important sources for the study of the Tangut language, such as the Tangut versions of Sunzi Bingfa, Mengzi etc., have long since been available; therefore these are not included here. Tangut poetic compositions which employ Sanskrit śloka or gāthā genres are included only occasionally: since grammar markers in these texts are kept to the minimum by the Tangut authors and translators, these texts are not useful for the beginners.12 The selection further includes two fragments from the Tangut version of the Jingde Chuandeng lu 景德傳燈錄, which records encounters between Chan monks and the Sixth Patriarch Huineng. One extract “Expositions of the 10 K. Kepping, “Tangut Ritual Language”, http://www.kepping.net/pdfs/works/Tangut_Ritual_Language.pdf, accessed 26.01.2019 11 To access Tangut input systems as well as electronic dictionary and variety of other materials, visit: 古今文字集成，http://www.ccamc.co/, developed and maintained by Jerry Yu. 12 Lin Yingchin 林英津，Xia yi Sunzi Binga yanjiu 夏譯《孫子兵法》研究》，(Taibei：Zhongyanyuan, Shiyusuo, 1994); Peng Xiangqian 彭向前, Xixia wen Mengzi zhengli yanjiu 西夏文《孟子》整理研究，(Shanghai: Shanghai guji, 2015); Lin Yingchin 林英津, Xixia yuyi Zhenshiming jing shiwen yanjiu 西夏語譯《真實名經》釋文研究, (Taibei: Zhongyanyuan, Shiyusuo, 2006), etc. Such texts as the Tangut translation of the Satyadvayāvatāra, are also not included here.  9 Three teachings” comes from unidentified composition, known as Suiyuanji 隨緣集. Texts of Tibetan subject matter include two selections: one is a fragment on Atiśa Dīpaṃkaraśrijñāna, discussing the reasons why he travelled to Tibet (the so called Upadeśa “Forty Banners of Emptiness” ljɨ̱ r1 njɨ̱ 1 ŋa1 dźjow1tshji1ŋwu̱1 ); still another is a fragment of the description of the early Tibetan translator Vairocana’s mission to India in search of the Rdzogs chen teaching （the so-called Preface to the Five Transmissions in the Dharmadhātu, ŋwə1djịj1tsji̱ r1kiẹj2gu2bu1 ） . We have also included fragments from Tibetan Buddhist ritual literature, e.g. a fragment from Bodhicittotpādasamādānavidhi. Native Tangut texts are selected from the Tiansheng Law Code, prefaces to the Buddhist sūtras, introductory verses to the Sea of Meanings Established by the Saints śjɨj2gu1·wo2 ŋjow2 , Preface to the rhyme dictionary Rhymes of the Five Tones ŋwə1ɣiẹ2·we2 bju1 . We normally only commented upon the personal names, place names and certain words, which in our opinion, a reader might have difficulties interpreting. Other than that, commentaries are kept to the minimum. However, as one will see from below, the example sentences used in the Reader are extracted from Tibetan texts, not included into the Reader. This is done on purpose, since, as we believe, it is easier to identiy the functions of the Tangut “auxiliaries” in the Tibetan sentences. When we provide Tibetan equivalents for the Tangut markers, it is done for the illustration purpsoses, and not because of genetic relationship between them. Aslo, as further exposition will demonstrate, there is no one-to-one congruence between Tangut and written Tibetan. The comparison between Tangut translations and Tibetan texts is further complicated with a general consideration: available Tibetan versions of the texts in Tangut generally postdate the Tangut translations, therefore we cannot be sure how to account for the deviations between the versions: these could either be due to the aberrations in the translation, or determined by the nature of the originals, which the Tangut translators had. Texts were selected so as they can be understood directly, without reference to either Chinese or Tibetan originals. This means, that this Reader is not designed as a self-study book but requires guidance from someone at least remotely familiar with the Tangut language, or with the Qiangic languages in general. 5. Ambiguity in interpretation What follows below is essentially an exposition of the Tangut morphemes which we deem important for the initial analysis of various Tangut sentences. These are equipped with examples, illustrative for the usage thereof. That is, the exposition is based on our reading of actual texts, and intends to help the readers to see the structure of Tangut sentences. Some of the interpretations are fully intuitive, some are purely ad hoc. This is inevitable, for a language which still remains to a large part obscure. Before proceeding to actual material, some preliminaries are due. Scholarship agrees on several important points concerning Tangut morphology and syntax. Morphologically the most productive device in Tangut was the phonological alternation, which signaled the transformation of the roots between the “parts of speech”, e.g. between nouns and verbs, between transitive and intransitive verbs etc. Adjectives are not specifically marked in Tangut, and are indentified according to their position in a sentence. Adverbs are marked by specific morphemes, which can be sometimes but not always equated with the Tibetan la don (terminatives, according to M. Hahn) particles. In this respect Textbook of Classical Literary Tibetan by M. Hahn is an indispensable source of reference for our Reader. One of the basic deviations from the standard descriptions of Tangut is that we treat verb prefixes in various capacities, and generally proceed from an understanding that the prefixes should be treated as the adverbial markers, in a sense close to the Tibetan “terminatives”. Phonological alternation was universal in all divisions of Tangut language, and various paradigms involving the transformation of vowels, initial consonants, rhyme and tone are found throughout all Tangut word-classes, as was originally shown by Nishida, and later developed by Hwang-cherng Gong and G. Jacques in an array of publications, and recently systematized in a concise manner by Nie Hongyin. That is, proceeding from these observations, we can conclude that Tangut language in fact did possess inflectional morphology, which was represented orthographically. Scholarship reconstructs several morphologically productive patterns of the sound change in Tangut. However, one again should keep in mind, that some of these patterns are better supported than others. This is particularly due to the deficiencies in the phonetic reconstruction of the contrasts within the vowel  10 system. That is, while some patterns of sound changes fit well into Tangut phonology, some remain scholarly attempts to account for phonemic contrasts which are not recognizable through standard analysis. In theory, analysis of the Tangut sound changes enables an arrangement of the members of various classes into contrastive cognate stems. This further allows suggesting that various Tangut word classes (verbs, pronouns, nouns) derive from a set of roots through established paradigms of sound change, which can also be accounted for from cross- linguistic and historical perspectives, as was demonstrated by G. Jacques and his research in a series of publications. However, this type of systematic research remains a desideratum. Besides its linguistic importance, the reconstruction of cognate stems is important for the historical research, especially in interpreting of the Tangut native titles, official ranks, names, etc. Scholars have long since discovered that the Tangut script was created in a manner that the sound changes were represented in orthography, but not always in a straightforward manner. Therefore, in many cases, comparing orthography with phonetic reconstruction, one can observe partial or complete correspondence between the graphic and phonetic representation of a sound chance of a specific root. However, the connection between sound and meaning, which was clearly visible for the Tangut native speakers, is not so obvious for us now. For reading the texts knowledge of basic “grammar” morphemes is imperative. Although one might choose to treat Tangut grammar morphemes as analogous to the Chinese “empty words”, one must be aware that most of the Tangut “empty words” (e.g. verb prefixes, agreement markers) cannot be adequately reproduced through the Chinese characters. There is only partial congruence between the standard list of Chinese “empty words” (e.g. as inferred from Wang Li 王力, Guidai Hanyu 古代漢語) and the Tangut ones. Although Chinese equivalents can be assigned to the Tangut grammar morphemes, such an equation is context dependent, and not universal or formal, as can be observed from the treatment offered by Nie Hongyin, of found in Han Xiaomang 韓小忙 Tangut Dictionary. While interrogative pronouns and some of the “case marking” morphemes are adequately rendered through the respective Chinese morphemes, most of the Tangut “auxiliaries” (using the translation of the actual Tangut word) resist such an analysis. Again the easiest way to interperet these, is not from the Chinese equivalents, but from the Tibetan particles. Comparative analysis demonstrates that Tibetan la don particles, case markers, conjunctions, etc., are more easily associated with the Tangut morphemes than the Chinese “empty” words. Since all of them emerge in both Tibetan and Sinitic texts, we have chosen Tibetan based exemplar sentences, which are more straightforward in revealing the fucntions of these particles. Tanguto-Chinese substitutions, which are common in modern scholarly Chinese translations from Tangut, remain at best arbitrary or even misleading, when applied to the syntactic or morphologically significant morphemes. Adherence to this practice produces aberrations in the analysis and is in fact an obstacle for the correct understanding of the Tangut sentences, especially in cases when applied to the texts which are not translations from Chinese. However, the substitution method remains convenient, especially for the scholars originally trained in Sinology. The problem is thus not the substitution itself, but a point of view where the substitution equals translation and understanding. From the syntactic perspective, Tangut language is SOV. Apart from the sound alternations, such functional categories as the “parts of speech” (i.e. pronouns, nouns, verbs, adjectives, adverbs etc.) can be identified in Tangut texts generally on the basis of their position in a sentence and on the presence of specific markers, or absence thereof. Contrasts between the parts of speech are established on the basis of phonological alternations and affixation of specific morphemes, indicative of the change of the word-class, e.g. nominalization of verbs, or verbalization of nouns. Finally, position in a phrase or in a sentence or in a clause is indicative of a part of speech. Syntactically, the second element in a two-component phrase is identified as a verb or as an adjective, given the identification of the first component as of a “noun”, which can either be a “subject” or “object” of a verb, or a referent of an adjective. E.g.: ·ụ2 tsəj1 ，“Small Vehicle”（lit. 乘小）. In some cases the order is reversed, e.g. 2 2 in tha ·ụ ，i.e. 大乘, which is explained through historical and cultural reasons: this last term is a direct calque of the Chinese dasheng 大乘. Verb normally occupies the final position in a sentence, and in most cases is identified with the presence of a verb prefix and/ or an agreement marker. Since prefixes and agreement are inexistent in Chinese and Tibetan (as separate morphemes), the Tangut translations from these languages are of limited value for the grammatical analysis of Tangut verb system. Verbs without prefixes are rare in standard Tangut, and mostly occur in poetry, where the prefixes are omitted for the sake of maintaining the meter. However, as long as the same verb roots emerge both with and without prefixes, they are easily identifiable with experience.  11 Tangut possesses rich (or even redundant) pronominal system, which is currently under scholarly scrutiny. Tangut pronouns can be analyzed from the perspective of their distribution and combinatory characteristics. Generally, Tangut pronouns can be distributed among the classes of personal, demonstrative, interrogative and so on, and according to the layers of ordinary and honorific speech. The criteria separating “ordinary” and “honorific” speech are at best vague, and the concept itself is an ad hoc interpretation of a situation when pronouns of allegedly various layers emerge in similar contexts. However, for the reading purposes, this is of limited relevance. Out of basic pronouns, plural forms of nouns and pronouns can be formed with the help of specific morphemes, such as njɨ2 and nji2 .This last one often occurs as the agreement marker for 1st and 2nd PL. Specific pronouns for 1st PL existed, based on the contrast between “inclusive” and “exclusive” forms. These will be introduced below. DUAL also existed in Tangut, and was marked with a morpheme mji2 . Depending on the analysis, various Tangut morphological and syntactic morphemes can be assigned to more than one functional category, and can emerge in morphologic and syntactic capacities. This is partially explainable by the fact, that Tangut, just as written Tibetan, is a high context language with a high degree of homonymy and large number of homographs. The problem of the so called tongjia zi 通假字 was originally observed for the meaningful words, but as we suspect this situation was even more common for the “auxiliaries”. The problem of whether these “ambiguous” morphemes represent varying functions of the same morpheme in different contexts, or they are realizations of originally different morphemes is beyond the scope of present description. Suffice it to say that one cannot be always sure, if one is dealing with a singular morpheme in different capacities, or with different morphemes, which had been pronounced and written in a similar manner, i.e. partial grammaticalization. Adherence to the Chinese equivalents of Tangut graphs provided by Gule Maocai in his celebrated Tanguto- Chinese Timely Pearl in Hand is another reason of ambiguity. One should keep in mind, however, that the Chinese and Tangut texts in the Pearl are not the translations of each other, but probably had been composed independently by a person equally fluent in both languages; therefore both versions follow the grammar of the respective languages, and thus cannot be treated as example sentences in a grammar textbook. Below we present several cases of such ambiguity; more textual examples are in the respective paragraphs of this Reader. 5.1 To illustrate the above we can provide several examples. Some of the Tangut “markers” were assigned to one specific Chinese graph and vice versa: e.g. Tangut tsjɨ1 , to which a variety of Chinese equivalents is provided (e.g. in Han Xiaomang’s Tangut-Chinese Dictionary). Some of these cases are relatively easy to solve: e.g. for tsjɨ1 , one has to discriminate three possible positions of this morpheme. Between the nouns it has meaning similar to English “too, also”; after verbs and in the end of clauses it can either be a “concessive marker” (“although”, that is marking a concessive clause, directly congruent with the Tibetan kyang or yang, which in turn translates Sanskrit api), or marker of 2nd DUAL agreement (i.e. morphological function, which has no Tibetan equivalent. We still need to carry out more research, if it emerges in the cases of the DUAL number in Sanskrit). Apart from this, tsjɨ1 can be suffixed to an interrogative pronoun, meaning “whatever”, “whichever”, “whatsoever” (i.e. Tibetan: gang yang). This is again due to its concessive function. Again, there is no way of telling if we are dealing with similar morpheme in different capacities, or with a two different morphemes with different etymology. Symmetrical situation is observed for the morpheme kjɨ1 , for which one also has to discriminate between three possible positions: Before a verb it functions as a prefix, after a verb it is an agreement marker for DUAL. The third case is kjɨ1 after an interrogative pronoun, of a noun where its function is less clear. Wenhai glosses this morpheme as “similar with nja1 and rjɨr2 ”, i.e. accounts for both prefix and adverbial functions. Generally, as our example sentences demonstrate, when interrogatives are suffixed with kjɨ1 , they perform in standard interrogative capacity, whereas suffix tsjɨ1 adds concessive meaning, “whoever”, “whatever”, etc. This is probably connected with “concessive” aspect of the morpheme tsjɨ1 . Further, as noticed by G. Jacques, kjɨ1 can also have an adverbial usage, indicaitive of the manner in which an action takes place. Its position in an interrogative is probably due to this last function. Hence this morpheme can be interpreted as a la don. If proven correct this will explain the problem of the double prefixation, which is troubling scholarship for some time. If this point of view is correct, we are not dealing with the “double prefixation” at all, but with a verb and its adverb. That is, we observe the congruence in the usage between the two morphemes kjɨ1 and tsjɨ1 , whereas each one has its own peculiarity: tsjɨ1 cannot be a verb prefix.  12 5.2 Further, the Tangut language possesses a set of “auxiliary morphemes”, correctly interpreted as markers of “nominalization”, “complementation”, “verbalization” etc. One can provisionally determine these as suffixes. Out of these śjij1 ，tjị2 ，lew2 are the most frequent, and the most productive. Chinese equivalents for these are shun 順，ke/ suo 可/ 所 and ying/ suo 應/ 所 respectively. Chinese equivalent suo 所 assigned to both last morphemes demonstrates that they had partially equivalent distribution with the Chinese suo 所, whereas the first one is unrelated with the two. In fact, the first of the three nominalizes a verb or is suffixed to a pronoun, and is indicative of a manner in which an action occurs, e.g.: thja1 śjij1 translates Tibetan de nyid, i.e. “naturally”, in Chinese contexts: ziran 自然. In this particular case, the direct substitution will not be helpful: although Tibetan de can be equated with the Tangut indicative pronoun thja1 , śjij1 is unrelated to the Tibetan nyid. Other Tibetan equivalent for this Tangut is de ltar. In this circumstance, thja1 śjij1 contrasts with ljɨ1̣ kjɨ1 , Tib.: ji ltar…de ltar, based on Sanskrit yasmāt…tasmāt or more generally yat… tat. When śjij1 is suffixed to a verb, it nominalizes it, creating new lexical words. Other examples are: kjụ1 śjij1 （ verb root “to strive for” + NOM suffix ） is a translation of the Tibetan sgrub thabs, “method of attainment”, tśjɨ̱ 1 śjij1 ，（verb root “to follow”+ NOM suffix）, is a translation of the Tibetan rim pa, “manner of arrangement of ritual actions in order”, Sanskrit karma; ·o2 （“to enter”）+ śjij1 returns “entry” (cf. Tib: ’jug pa); lja1 śjij1 ，“manner of coming”, i.e. “arrival”, but also a more general meaning “situation”. But directly equating śjij1 with Tibetan suffix pa is again incorrect, since it does not indicate an “agent”, as does the Tibetan morpheme, and there is a variety of other ways to express nominalization with pa/ ba in Tangut, as will be clear from the example sentences. That is, Tibetan morpheme remains more polysemantic that the Tangut one. There are some more examples: (“Transmission of the Dharma in the Bod lands”, here śjij1 modifies de̱j1 “to transmit”); (“I understood Masters’ intention/ desires”; here śjij1 modifies 2 śia̱ ). This shows that manner of derivation (or word production) through nominalization is productive in Tangut. Treatment of the last two morphemes tjị2 ，lew2 is again ambiguous: from the Chinese equivalents one can infer that both have partially equivalent distribution, since suo 所 equally applies to both. The distribution of both tjị2 and lew2 is rather broad: they apply to both nouns and verbs.Ambiguity is caused by the semantics of the Chinese suo 所 in the traditional materials. Importantly, only tjị2 was identified as an “auxiliary” by the native materials. Being suffixed to a verb root both morphemes transfer a verb into a nominalized verb phrase. Tangut tjị2 is suffixed to the verbs of “abiding” and “movement”, e.g.: dźji̱ j1 tjị2 > verb root “to live + NOM suffix” = place of abiding. Quite contrary with this definition, when suffixed to the indicative pronouns, tjị2 maintains its actual meaning “place”: thjɨ2 tjị2 ，“this place”> “here” (lexical meaning). This particular 2 locative function is not identified for lew . Further, tjị2 is contrasted with the locative suffix do2 , which indicates direction: ， (“It is similar to the situation, when one maintains the direction, he will arrive to the place, which is intended”). But： , ，“if one strives to attain the true wisdom while there is no desired (conceived) object…”. Easy to see that on the surface level both śia̱2 tjị2 and śia̱2 lew2 emerge in similar context and contrast between the two phrases is impossible to establish. Similarly, in certain contexts lew2 remained a meaningful word, with modal meaning, this particular feature is not observed for tjị2 . Both tjị2 and lew2 equally participate in the negative passive constructions, similarly rendered through the Chinese bu ke 不可, as well as combine with the morpheme sji2 , as observed by Duan Yuquan 段玉泉.This particular function is again due to the nominalizing function, i.e. “something which cannot be done”. The above means that although tjị2 and lew2 are different morphemes, but possess partially equivalent disctribution. However, the Chinese calques are indicative of their contrasts. Currently we intend to consider these two morphemes as representing similar pattern of nominalization, generally for the lack of a better option. More examples provided below. So, as above we observe partial congruence between the two morphems, but at the same time each of them has its own specific capacity. 5.3. Subordinate clauses with temporal meaning are marked with series of “words” equally meaning “time/ when”, “while”: zjịj1 ，dzjɨj1 , zjọ2 . Despite several attempts, scholarship was not able to identify a clear  13 distribution pattern. Therefore, on the grounds of partial homonymy between the three, and observabale sound change pattern: z/ dz; jij/ jo, we intend to determine them as mutually interchangeable variants. In some cases, judging from the context, zjịj1 seems to have distribution equivalent with one for the conjunction ku1 ,which coordinates two clauses, where the subordinate clause is a condition of the main clause (“if…then”). E.g.: , . (“If (when) one enters the sea of samsara, it will be as of one taste”.) Neither dzjɨj1 , nor zjọ2 appear to have this particular distribution. Currently, only dzjɨj1 can be identified as a “noun” meaning “time”, since it occurs in phrases with adjectives “long”, “short”. Usage of zjọ2 seems to indicate basic temporal clause: .(“During previous time, although I was reciting many mantras, I was often afraid.”). However, as in the previous case, we consider the three morphemes as synonyms, sharing similar function, leaving the subtleties of distribution to more qualified researchers. 5.4. In some cases, Tangut “empty words” were only partially grammaticalized, retaining some of their original lexical meaning, whereas others could no longer function as free forms. The first situation is observed for tjị2 and lew2 . Another example is a set of Tangut morphemes, which all share Chinese calque zhong 中. These are Tangut kha1 , gu2 , ·u2 : all three are assigned Chinese zhong 中, whereas gu2 , ·u2 can also be 1 2 understood as Chinese 內, whereas kha cannot. Out of these three, gu can operate as a postposition “inside”, “within”, and as an adjective “middle”; ·u2 is again postposition “inside”, and as an adjective “inner”, as opposed to “outer” ( ). These two can be related to the domain of “spatial” markers. Finally, only kha1 has a function of marking a dependent clause. In this case it is translated as “in the circumstance”, and seldom performs in the “spatial” capacity. It is specific function is to mark quotations, in this capacity it is directly equated with the Tibetan locative la.That is, quotations are mostly marked with kha1 , and not with gu2 , ·u2 . That is, the three Tangut words are not mutually related, among these only gu2 is legitimately translated as zhong 中 in an adjectival sense. Their treatment as synonyms was determined by arbitrary assignment of a similar Chinese equivalent, and not by an existing or recognized semantic proximity. 5.5 The most peculiar feature of the Tangut language is its verb morphology. One important feature of the Tangut verb system is “verb agreement”, which can take place either between subject and predicate, or between object and predicate, and is arranged differently depending on the Person of both subject and object (detailed study by G. Jacques). Based on the phonological alternations and their ability to accept agreement markers, Tangut verbs can be distributed between the two morphological classes, one of which is understood as the “basic stem” (i.e. the one un-marked with agreement markers and sound alternation, stem A) and the other one as the “derived stem” (stem B), as suggested by Nishida, Gong and Kepping. Division manifests in phonologic alternations, including the changes in the root vowel, rhyme, tone or initial consonant and affixation of a specific marker to the stem. Some etymological observations concerning the origin of the root vowel Ablaut were suggested by G. Jacques. Distribution of verbs between the stems reflects contrasts between transitive and intransitive classes. As mentioned above, the verb stems contrast by their ability to accept the markers of verb agreement: “the basic stem” has a zero agreement marker (or, the stem itself is a marker), whereas the “derived stem” is able to take agreement markers for the 1st and 2nd singular, dual and plural. The agreement markers are homophones (and sometimes homographs) of the respective personal pronouns, i.e. of 1st, 2nd SG. There are doubts as for the genesis of the agreement markers, however, as orthographic evidence demonstrates, the Tanguts believed personal pronouns and agreement markers to be cognate (see the table below). As is clear, the subject of the “basic stem” is 3rd singular and plural, and for the “derived stem” 1st and 2nd SG and PL respectively. The markers and prefixes were identified by Nishida Tatsuo, and systematically introduced by K. Kepping, dual markers have recently identified by Zhang Yongfu 張永富. Verbs of both stems normally appear with prefixes; however, these are not mandatory and there are cases where verbs are not prefixed. Prefixes demonstrate a clear pattern of phonological alternations, involving rhyme, tone and/ or main vowel change, while initial consonants remain stable. On comparative grounds with other TB languages, Tangut verb prefixes are treated as “directional”, however from a semantic perspective their “directional meaning” had almost completely faded. Thus, treatment of the prefixes as directional is “generally accepted” rather than “correct”. Proceeding from an ad hoc semantic analysis, scholarship suggested opposition between the prefix series as indicative of “perfective” and “irrealis” aspects. Although such a division is maintained here, it is not always  14 supported by the actual texts. Even more importantly, regardless of the origin of the prefixes (i.e. “direction markers” or “direction verbs”), their etymology is not discoverable by means of historical reconstruction. Treatment of prefixes is complicated by the nature of our sources: e.g. Gule Maocai maintained “one-to-one” translation principle. Thus he provided Chinese equivalents for each Tangut character in the text, including prefixes which have no direct analogs in Chinese. Chinese equivalents he chose to translate both prefixes and other Tangut auxiliaries are context dependent, therefore Chinese translation actually further obscures the problem. This situation is reproduced in the newly published Tangut-Chinese Dictionary by Han Xiaomang 韓小忙13. Although written Tibetan demonstrates verb prefixation, there are only a few cases where we can identify Tangut prefix with a Tibetan prefix e.g. of the past tense. 5.6. Agreement markers and stem sound alternations have clearly defined function. The problem of how Tangut expressed “tense” and “aspect” remains. Scholarship generally believes that verb prefixes contrast as expressing “perfective/ irrealis” aspect opposition. By this logic, one can suggest that the prefixes of the “perfective” series (series A of G. Jacques) could be understood as the markers of preterit and present tenses, whereas the “irrealis” (series B of G. Jacques) indicate the future tense. This clearly an extension based on analogy, and should be understood not in terms of Tangut actually possessing category of “tense”, but rather as indicative of a specific manner to express “time”. This interpretation is rather ad hoc, and we still have not reconstructed distribution pattern for the prefixes and their connection with verb stems. Such connection is probably semantic, and not phonological, but this remains to be proven statistically. Apart from prefixes, scholarship had identified certain “auxiliaries” which are indicative of the “future tense” and “past tense”. Similarly with the above, their distribution pattern is not yet discerned and their presence in a verb phrase is again not mandatory. These “particles” more or less obviously derive from the verb stems meaning “to wish for” (e.g.: kiẹj2 ). By this logic, marker of the “past tense” should be “to complete” (e.g.: dźjwa1 ), however, this does not appear to be the case. In any event, even kiẹj2 was not fully grammaticalized when Tangut language was put into writing. One possible exception is a pair of the “future marker” ·jij1 , and “perfective marker” sji2 . Out of these, ·jij1 derived its phonetics from ·jij1 ，and meaning from lja2 , “to come”. This word was originally identified by the Tanguts as an “auxiliary” 14. E.g.: (“He will return to the Bod land”). However, this 1 morpheme is scarce in the Tangut texts. While ·jij is clearly a morpheme of the “future tense”, the function of the morpheme sji2 is less clear. Contexts seldom allow direct opposition between these two morphemes, and alternative interpretations are possible. In certain contexts sji2 is obviously an equivalent of the Chinese neng 能, as originally observed by Nie Hongyin. Such combinations as ljij2 sji2 ka̱r1 , “eye endowed with an ability to see” are not easy to interpret: Here sji2 can either an auxiliary verb “to be able”, or a member of an apposition: “seeing+eye”= “seeing eye”, i.e. a nominlization. If there were no sji2 , the phrase would mean something like “seeing of an eye”. E.g.: ， (“When one starts to practice on the basis of his understanding, and until he had reached the Buddhahood, it is the direct path.”) Here is again ambiguity between possible modal and nominalization functions. Similar situation is observed in: , which can be translated as both: “is able to adorn the realm of the Dharma nature”, or “the adornment of the Realm of the Dharma nature”. In certain contexts sji2 acquires syntactic function. Such an ambiguity means that the indications of “tense” in the Tangut language are generally derived from context, and, as a rule, are not expressed through verbal morphology. Physical time is inferred from context. 5.7. There is no exact evidence as to confirm the existence of a complex case system in the Tangut language. In Tangut, case relationships are expressed through postpositions, studied by Nishida, Sofronov, Kepping and other scholars. Case system for Tangut was suggested e.g. by Nishida and Arakawa, probably as an analog to the grammar of Japanese. The division into “core” and “local cases” in Tangut suggested by Arakawa is useful for expedience sake, and has little explanatory power. The list of “local case markers” is variable, while the so called “direct” and “indirect/ oblique” cases in Tangut are somewhat clear. Marker of nominative is zero (and not tja1 , which specifies general topic of a discussion); direct object is again marked by zero, whereas the indirect object in 13 Han Xiaomag 韓小忙, Xixia wen cidian (shisu wenxian bufen) 西夏文詞典（世俗文獻部分）. Vol. 1-9 (Beijing: Zhongguo shehui kexue chubanshe, 2021). 14 Both these morphemes are traditionally treated as suffixes.  15 variety of sema/ntic connotations is represented by·jij1 , traditionally equated with the Chinese zhi 之. When this graph emerges between the nouns, it represents genitive (attributive), in other cases it codes various relationships between nouns and verbs, i.e. accusative, dative, etc. Two clearly identifiable case markers are rjir2 ，which combines ablative, commutative and dative; and ŋwu2 , instrumental, which also has causative and sometimes adverbial meaning. Many postpositions belong to the “local category”, indicative of “direction”. These will be discussed below. Finally, one should mention the so-called Tangut “ergative” case. It is represented by a compound dźjɨ ·wji1 ，translated into Chinese as xingwei 行為. Another major issue is the well-known homonymy of Tangut language. This leads to a phenomenon recognized as tongjia 通假 characters. The markers will be represented in the respective section of this Reader. All of the above creates a paradox when we are able to read texts with an ever-increasing accuracy, but are not always able to analyze the texts linguistically. The nature of the Tangut script is such that phonetic alternations are not directly visible in graphic form to an outsider, although they might have been obvious to the native speakers. Therefore we cannot always correctly apply our knowledge of Tangut morphology to the textual analysis. Generally, we are still not able to correctly utilize the prefixation system and morphologic alternations in our reading of Tangut texts. Juxtaposition between Tangut and Tibetan and Sanskrit materials had just begun; hopefully it would yield reliable results, allowing clearer observations concerning the connection of the prefixation system with the tense system. 6. Tangut Grammar in a Nutshell NB: Treatment of Tangut morphology offered here sometimes deviates from the interpretations offered by scholarship, e.g. by G. Jacques and Arakawa Shintarō. Since we believe our Reader be to be a reading guide, we adhere to our interpretations, and refrain from an in-depth discussion of scholarly subjects. Example sentences only comment on the discussed morpheme, larger context is introduced only when necessary. 6.1.Numerals Standard system 一 lew1 ·a 二 njɨ̱ 1 三 sọ1 四 ljɨ̱ r1 五 ŋwə1 六 淚 tśhjiw1 七 śjạ1 八 ·jar1 九 gjɨ̱ 1 十 ɣạ2 百 ·jir2 千 tụ1 萬 khjɨ2 There are three independent systems of cardinal numerals in Tangut. The table above contains standard system of the Mi language, which presents no difficulties. Apart from this, odic Tangut and the Homophones preserved other systems of numerals, based on the sequence of months and sequence of sons (男) in a family. These systems were studied by Nishida. However, in most of the texts, alternative numerals appear very seldom and only in specific contexts: e.g. sọ1 ·ụ2 , i.e. sancheng 三乘, “three vehicles” (standard), but rejr2 ·ụ2 ，i.e. sanzang  16 三藏 (odic) ，a monastic title; ɣạ2 njɨ̱ 1 djịj1 lwər2 ，shier bu jing 十二部經 (standard) ，but śja1ljijr2 (odic) ，i.e. shifang 十方. Place name tśjɨ̱ r2 thej1 ŋər1 ，Wutaishan, uses an odic numeral for “five”. As native literature confirms, the alternative numerical systems were mutually compatible, and the such characters as rejr2 and sọ1 were glossed though one another. Ordinal numbers in the standard system were formed through suffixing of a morpheme tsew 2 to the respective numeral. 6.2 Pronouns. Tangut pronominal system is complex. Below we list basic pronouns necessary for initial familiarity with the texts. Homonyms (and in some cases homographs) of personal pronouns are used as the markers of verb agreement in Tangut. 6.2.1 Personal pronouns Tangut pronominal system includes a variety of morphemes, which in some cases resist straightforward interpretation. The texts in this selection generally contain the set of three basic pronouns: 1st SG, 2nd SG/ PL (probably used in the honorific speech), there was no specific 3rd SG pronoun: indicative pronouns were used to address 3rd SG. Some morphemes homographic and homophone with the personal pronouns participate in pattern of verb agreement. Tangut verb agreement will be discussed in a separate section. pronoun Translation ŋa1 (plural either unmarked, “I, me”; 1st SG; 1st SG subject or formed by addition of njɨ2 ) and object agreement mjo2 “I, me” 1st SG, honorific speech? nja2 (plural formed njɨ2 “You”, 2nd SG; 2nd SG with or with mji2 ) nji2, plural formed with njɨ2 “You”, 2nd SG, elevated speech; nd 2 PL Examples below illustrate the distribution of these pronouns. These examples give general idea of the usage of the pronouns. Some example sentences feature verb agreement, which will be dealt elsewhere in this introduction. a. ŋa1 Apart from being 1st SG personal pronoun, Tangut ŋa1participates in variety of terms, involving Sanskrit ātma, or Tibetan nga and bdag, e.g. ŋa1 mjij1 (無我, bdag med)，ŋa1 khwẹj1 （我慢, nga rgyal）, etc， that is it is compatiable with both Tibetan bdag and nga and also in some contexts with kho bo. All the pronouns can perform the capacities of subjects and objects. In this last case they are suffixed with ·jij1 , or emerge independently. E.g.: ， ， : “If in all the dharmas, there is no attachment to “You” and “Me”, then all-equal realization is attained” (KU); “ ， ”: “All the Great Compassionate Ones think and remember me.” (SAM, Tib: bdag la dgongs su gsol). (1) : “I am the one who is being taught about, and I am the one who teaches (lit.: “speaks”) (HE, Tib: ’chad po nga chos kyang nga) (2) ， : “I have inserted here so that to make happy the hearts of those others who have recently entered the [Path] of learning”. (GALO, Tib: ’on kyang kho bos ni gzhan dag las dang po pa phal cher yid ’jig par bya ba’i don du ’di bcug pa yin te)  17 b. mjo2 (3) : “I am Vairocana.” (FIVE) (4) : “I have never had such an inauspicious dream as today.” (FIVE) (5) : “Thus all my words and sayings… ” (ZHANG, Tib: de phyir kho bo’i lab lob…) (6) , : “Who is Shun? A man. Who am I? A man.” (THREE) (7) ， : “While I was travelling, I saw this.” (HUIZHONG) (Above contexts do not support “honorific” or “elegant” speech) c. nja2 (8) : “You two should use skillful means” (FIVE) (9) : “I do not need gold from you” (FIVE) (10) , : “[The mantras] which you recite, whom you teach?” (FIVE) d. nji2, judging from the structure of the character, it originally represented dual number, however, this meaning has almost completely faded. Observations show that in the texts translated from Tibetan nja2 is generally used as a transcription character, whereas 2nd SG is occupied by nji2, which emerges as the translation of both Tib khyed and khyod, i.e. its identification as honrific is not supported by the sources. (11) : “You go to the realm of Bod to turn the Wheel of Dharma” (FIVE) (12) : “You (PL) [and your followers] do in whichever way it suits you.” (GALO, Tib: khyed rjes ’jug dag gar mos par gyis) (13) : “I [will] truly tell you.” (GALO, Tib: khyod la yang dag bstan par bya) (14) , , , : “As soon as you two (PL) go to listen (DUAL AGR) [to the Teaching] and this Dharma will prosper, then not only my happiness, but also your (PL) happiness will be beyond measure.” (FIVE) (15) : “ I give you (SG) this little gold” (FIVE) (16) , , : “Indians, since they are evil, [they will] come for your (2nd PL) lives.” (FIVE) e. Alongside the aforementioned pronouns, Tangut specifies 1st PL inclusive and exclusive pronouns: gja2 mji2 and gjɨ2 mji2 . These two seem to be in complementary distribution. Morphemes gja2 , gjɨ2 and mji2 do appear independently in the texts known to us. (17) … ， ， : “We [EXCL] composed [] treatises and texts, and although [we have] exhausted our minds and hearts, [these are all] useless.” (MComm) (18) ：“It does not belong to us (INCL), it belongs to you, Great Master.” (GAMPOPA) (19) : “If today we kill (1st PL AGR) these people, then…” (FIVE) Above (8) exemplifies similar mode of the 2nd PL, however it is the only known example so far. Here morpheme mji2 is indicative of the DUAL, see also (8). 6.2.2Indicative pronouns There is a variety of indicative pronouns in Tangut, generally implying opposition between “this”, “that” and “further that”, similar to the Japanese koko, soko, asoko. The complete paradigm is in the first row of the table below; other pronouns below obviously form paradigms in terms of the alternation of initial consonant, vowel or  18 tone; however, their mutual relationship is not completely clear. In sentences thja2 substitutes subject in 3rd SG.; other graphs with meaning “that, other” etc. are lexical words and not pronouns. Generally, the pronouns can be modified with the case marking morphemes, most importantly with ·jij1 . thjɨ2 “this” “it” thja1 “that” “it” thja2 “that” “they”, indeterminate thju2 “here” thja1 thja2 “plural” tśjụ1 tśhju2 tśhjɨ1 tśhjiw1 The distribution presented in the table above is imaginary, and the system of indicative pronouns should be studied further, both from the perspective of the sound correspondences and semantically. One should keep in mind that the pronouns are cognate and follow the patterns of sound alternation. Morphemes can perform in both pronominal capacities, and as modifiers to the nouns. Contrasting pairs or triads are inferred on the basis of inconclusive textual evidence, so the distribution and contrasting stems should be treated as hypothetical before more research is conducted. One example of the contrast between thja1 and thja2: “ ”: “that 2 1 before this”. Contrast between thjɨ and thja is clearly observable and will not be discussed further. a. thjɨ2 “This”, Tib: ’di ( ..: “[Those] who desire to abandon this world…” Tib: jig rten ’di las ’byung ngo zhes; , “E MA , this realm of Complete Enlightenement”, Tib: E MA HO, sangs rgyas spyod yul di…), by extension “here”, when suffixed with tjị2 . When modified with a verb sju2 ， thjɨ2 forms an adverb “like this”, or an adjective “similar”, other affixes include. (20) : “Rakśita behaved (lit.: did) according to this manner (i.e. “like this”) (FIVE) (21) : “At this place there is a meaning, which is desired.” (FIVE) (22) : “Once [I] slept, and in a dream I saw this image” (FIVE) (23) , , : “These (implied plural), are they gods, or apparitions, or whatever, I cannot understand”.(FIVE) (24) : “This is called Nature of Enlightenment.” (ULTIMATE) (25) … “If amanasikāra, [mentioned] here, is measured with the “true mind”, then…” (MComm) (26) : “If practitioners contemplate their bodies in this manner, then there is no [concept] of “man”. (ULTIMATE) b. thja1 “that”, “this” (Tib: ’di, de. , Tib: ’di ni Ā dang mdzes pa’i TA, “This Ā with the adorned TA”; : “Its essence is like this, then…”, Tib: de ni snying po de bas na); standardly contrasts with thjɨ2, by extension “there”, when suffixed with tjị2 . (27) : “That Vairocana could not have received the genuine upadeśa of the Ultimate Realization” (FIVE) (28) , : “If, for example, [one] sees rūpa, who is that seer? ” (ULTIMATE) (29) : “The meaning of that is also similar to this” (ULTIMATE) (30) : “On its lower side the King’s sleeping place was established” (FIVE) (31) , : “By the power of my some of [my] previous vows (lit.: there are previous vows), I went there.” (FIVE) (32) …: “Having presented maṇḍala there… (MComm) Unlike thjɨ2, thja1 can be suffixed with śjij1 ,as mentioned above. (33) : “Thus, there will be violation of the victorious upadeśas.” (FIVE)  19 (34) ， ， ， : “When discriminations do not emerge, and one abides naturally in this manner, [one] does not traverse the bhūmis, but is able to attain the bhūmi of Complete Enlightenment.” (MG) (35) : “It exists in the manner it is conceived.” (VAJRA, Tib: ji ltar brtags pa de ltar snang NOTE adverbial ) When suffixed with sju2 , thja1 creates a phrase congruent with Tib de bzhin. (36) : “In this way moving and unmoving particles.” (VAJRA, Tib: de bzhin gyo rdul mi gyo ba) c. thja2, “that” Tib: de, when suffixed with nji2 -“they”, when unmarked “he/ she”. Contrasts with thja1 (40) 15 One particular usage of thja2 is an indeclinable , translation of the Tibetan de rjes su, “subsequently”, and sometimes as a substitute for the Tibetan de kho na: “ : “Now it is also just as in the dream” (lit: in the manner according to the dream, Tib: byas pa cig rmis te de kho na bhzin du byung (GAMPOPA) (37) , : “[The families] of the guards live badly, give [the gold] to them”. (FIVE) (38) , : “After that, when the Tibetan king Khri srong had passed away, his son Ral pa can…” (FIVE) When suffixed with sju2 , thja2 creates a phrase congruent with Tib. de bzhin, similar with thja1. (39) : Thus rises joy of kunduru (HE) (40) : “If a practitioner does that, this appears before him, so that he will attain sovereignty.”(HUIZHONG) d. thju2 . Currently assign spatial meaning, probably “here”, also “this”. Contrasts with thja2 (45) (41) : “to strive for here and there” (VAJRA), here thju2 translates Tibetan der. (42) For meaning “here”: (3) for the Tibetan ’di. However, for this further study is needed. (43) : “Because of the power of previous vows, I came here”. (FIVE) (note wrong order in ) (44) ， : “There is neither attachment nor defilement, neither this/ here, nor that/ there.” (MIRROR) (45) ， : “Now, these three gates, these are brief, but explain broadly” (MIRROR) e. Other pronouns tśjụ1, tśhju2, , tśhjiw1 do not emerge in the texts used for this Reader, or the Introduction. One exception is tśhjɨ1, which emerges in temporal adverb：tśhjɨ1 dzjɨj1 ， tasmin samaye and in negative constructions which be discussed below. On the basis of the above, one can infer that Tangut probably did not discriminate between indicative and locative pronouns, but this has to be the subject of a specific study. f. Other than the above, Tangut seems to have a reflexive pronoun ·jij1 ，to which Chinese zi 自 is assigned. In many cases meanings “self” and “reflexive” pronoun are confused. Similarly, morphem ·jij1 often operates in the capacity of a verb, similar in meaning to the Chinese jishi 即是 (43 below). The usage of ·jij1 as of reflexive pronoun is rather limited. The distribution is partially equivalent with the Tibetan rang. When prefixed to other nouns, this morpheme means “own”: ·jij1 tsji̱ r2 ，“self-nature”; ·jij1 tshji̱ j1 “自說”, i. e. 優陀那 uddāna. 1 When combined with twụ , it means “each and every one” ( …: “Each of them understood that their inetentions were completely wrong...” MC). 15 For reading purposes only a, b, c are necessary.  20 (46) ， : “This was transmitted by Bodhidharma himself” (MIRROR) (47) : “This Way abides in self-attainment”. (THREE) (48) , : “The teachings, spoken according [to the abilities] of the people are beyond measure.” (THREE) (49) : “[Afflictions] naturally depart by themselves.”(DIA) (50) ， “：“The two truths, which are revered by everyone, I [will explain] first, following what was collected by others…”.(TWO) (51) : “I am the one who holds my own assembly” (HE, Tib: rang gi tshogs ldan nyan pa nga) (52) “ ” … : “Then from the symbol RĀM, which emerges in one’s own mind…” (GALO, Tib: rang gi snying gar rām las byung ba…) g. Tangut quantifier zji1 with “generalizing” meaning “everyone, everything” (Tib kun, also yongs) can also be considered as a pronoun, with meaning “all of them”, i.e. 3rd PL. In certain contexts zji1 nominalizes ŋowr2 2 ŋowr (“everything”, Tib: thams cad): “ , (TWO): “All the dharmas are similar to the flowers in the air”. (Lit.: “[understood as] similar to the flowers in the air”; : “All dharmas are the mind”, (GAMPOPA, Tib: [chos] thams cad rtog par ’dug). (53) : “The elements of the Bodhicitta are the gates to everything” (VAJRA, Tib: byang chub yan lag kun gyi sgo) (54) : “The most important among all, the wish-fulfilling tree” (FIVE) (55) : “What is similar with this is the Way for everyone” (VAJRA, Tib: de dang ’dra ba yongs kyi lam) 6.2.3 Interrogative pronouns and interrogative words Tangut includes several morphemes which can be interpreted as interrogative pronouns, and several lexical items which can be interpreted as interrogative “words”, generally meaning “what”, “who”, “how” and an “alternative question” (“or”). This last one occupies position between two clauses or words, whereas others as a rule emerge in the initial or in the final position. One specific case is the interrogative particle ·a ，which has two homonyms and homographs: a verb prefix and numerical “one”. In its prefix and interrogative capacities, ·a is directly prefixed to a verb root, and its meaning can be derived from context. In some cases interrogative pronouns can be suffixed with tsjɨ1 and kjɨ1 . In the first case, the meaning is extended to a generalized form: “whoever”, “whatever” etc. This is determined by the function of morpheme tsjɨ1 as of marker of concession.These two morphemes seem to be in complementary distribution. a. Interrogative wa1 “what”, probably can be interpreted as a general question: “is there?”, “what?” In 1 combination with dju (to be, Chin you 有) this morpheme can be translated as “is there” depending on the context. (56) : “What virtues [he] has?”/ “Does he have any virtue?” (BCComm) (57) ?: “Now are there any benevolent friends there?” (HUIZHONG) (58) : “Is there a reason for that?” (MComm) (59) : “What (are there) hindrances? What (are there) obstacles?” (HONG) Suffixed with morpheme sju2 , means “how”, in negative constructions: “whatever”, “whoever” (Cf Tib.: gang gyis, ji ltar, or Chinese ruhe 如何). (60) : “Whatever [one might] say, [one will] not be able to attain [this]” (CM)  21 Suffixed with tsjɨ1 in negative phrases: “whatsoever”, (cf.: Tib. cir yang, gang yang). In this particular 1 situation, tsjɨ seems to be in complementary distribution with kjɨ1 . Unlike interrogative ljɨ1̣ , wa1 is seldom suffixed with kjɨ1 . Interrogitives in Tangut can be used in the raltive constrctions. (61) , : “Now, whichever wise scholar I might meet, I am no longer afraid”. (GAMPOPA, Tib: da lta ji lta mkhas pa cig sla ’o zer yang shes rab kyi go phye nas nyams nga med; Tibetan and Tangut not fully congruent) (62) : “It does not emerge and does not perish, and is nothing whatever.” (VAJRA, Tib: mi skye mi ’gags cir mi dgongs) (63) ：“[Since] the mind and the dharmas are nothing whatsoever…”(VAJRA, Tib: sems dang tshos ci yang ma yin) Interrogative wa1 can be prefixed to lexical words: wa1 ɣiej1 （Is this true?）, wa1 nio̱w1 ， 1 1 1 1 2 wa ·jiw nio̱w , （both meaning “why”）etc. Distribution and meaning similar to wa sju . (64) ， ， ? : “Although I say “it is not similar to anything”, how can there be no virtue?” (HONG) (65) “Why cannot [one] understand the pure Dharmakaya?” (ULTIMATE) (66) , : “Why does one attain the unweathering powers?”(DIA) b. Interrogative ljɨ1̣ “what”, accepts various suffixes. ljɨ1̣ do2 , locative “where?”, cf. Tib: gang la (67) ： “There is no place where it can be heard” (VAJRA) ljɨ1̣ kjɨ1 , “what”, “where”, “how”. This is by far the best-attested interrogative. It can be suffixed with “locative” do2 , meaning “where to” (see (67). (68) : “Cannot go anywhere” (ZHANG, note syntactic tsjɨ1 , lit.: “Although where, it is impossible to go”, here Tib: gang du) In cases when ljɨ1̣ is affixed with the morpheme kjɨ1 , it acquires adverbial meaning, similar with the Tibetan gang du, gang gyis. However, evidence is inconclusive and sometimes supports gang rather than gang du. This form emerges in the sentences with the relative clauses. (69) : “Wherever one goes (whatever one does) it is the practice of Bodhisattva” (VAJRA, Tib: gang du dga’ bar byang chub spyod, Tangut and Tibetan not fully congruent here; adverbial ) (70) : “By what powers are the masses of devils collected?” (GALO, Tib: gang gis stobs chen bdud kyi sde, INS with bju1 ) (71) : “It appears in whichever way as it is conceived” (VAJRA, Tib: ji ltar brtags de ltar snang) = (Relative construction, adverbial) In many cases, especially in the translations from Chinese, or in the texts of “Sinitic” subject matter, ljɨ1̣ kjɨ1 is prefixed to the copula ŋwu2 （to be）. (72) : “What is the way to get to Vajrāsana?” (FIVE, adverbial) (73) : “What is the ultimate teaching of the Master?” (HONGZHOU) (74) , : “What are the three immeasurable kalpas?” (HUIZHONG) (75) : “What one must do?” (HUIZHONG) (76) : “What is substance?”(HUIZHONG) Unsuffixed. (77) , : “What is “song”; what is “dance”?” (HE)  22 NB: Since ljɨ1̣ kjɨ1 often emerges in combination with the copula verb ŋwu2 (to be), there is a tendency to attribute kjɨ1 as a verb prefix. This distribution is based on the Tangut equational pattern: tja1 …ŋwu2 , see below. But its co-occurrence with nouns, dictates combining ljɨ1̣ kjɨ1 together as a standardized 1 1 formula. Other combinations are possible as well. Again, ljɨ ̣ kjɨ and wa kjɨ1 1 emerge in similar contexts, therefore their comparative distribution is not clear. c. Personal interrogatives sjwɨ1 ，sjwɨ2 , “who”. The exact contrast between the two remains vague. In our texts sjwɨ2 emerges only in a compound sjwɨ2 nji̱ 2 , “someone” (Cf. Chin.: moujia 某甲). (78) : “In your realm, who is the best scholar in the teaching of the Ultimate Attainment?” (FIVE) (79) ? ? : “Who are you? Do you have friends? Where you come from?” (FIVE) (80) , 属 : “If it so happened that they are the holders of the upadeśas, then who taught [them]?” (FIVE, note ergative marker, lit.: “By whom they were instructed?”) (81) , : “One is not able to attain the [place of] Bodhisattva, who is the one to become the Buddha?” (HUIZHONG) (82) 務 , : “The mantras which you recite, whom you teach [it]?” (FIVE) = (9) (83) : “Whoever [one] is (i.e. no one), [he] cannot exceed the commandments of the Tathāgata” (FIVE) (84) : “Who and how (i.e. “no matter what and how”) does not understand?”, i.e. “No matter by whom, it will not be understood.” (CM, Tib: gang gis kyang ni de mi shes) d. Interrogative ljọ2 . This morpheme, alongside standard meaning “what”, ljọ2 prefixes verbs, including 2 modal, with a meaning “how possible”, “is it possible”, etc. Simultaneously, ljọ seems to have “spatial” aspect, i.e. “from where”, especially when suffixed with locative do2 . In this context it is in complementary distribution with ljɨ1̣ kjɨ1 do2 . (85) , : “Dharmas did not come from wherever, and cannot go anywhere.” (ZHANG, here Tib. has chos kun gang las kyang mi ’ongs gang du yang ni ’gro la), here and are in complememtary distribution. (86) ：“ ?”: “The Great Master asked: “The Great Virtue (voc., Chin, 大德), where you come from?”(HUIZHONG) (87) : “The principle of equality, which does not abide anywhere (i.e. all-pervaiding)” (ZHANG, Tib: gang yang mi gnas mnyam pa’i don) (88) : “Is there a way to remedy it through the help of the Path? (VAJRA, Lit.: “is there such a principle?” or “is it appropriate”, Tib. here has lam gyi bcos su ga la yod). (89) : “Is it appropriate not to meet [them]?” (FIVE, lit.: “is there a law not to meet them?”) (90) : “Where you come from?” (FIVE) (91) : “Is it not similar with our [saying]: “Eye moves and the Way stays”? (THREE) (92) : “How can miracles be attained in the state of no-mind?” (HUIZHONG, this phrase can be translated: “In the state of no-mind, is there any attainment of miracles?”) (93) , , : “Although there is nothing which is not “awareness”, while “absence” is it by itself absent, is it necessary to talk about “presence”? (ULTIMATE) e. Interrogative morpheme mo2 and its allograph mo2 ，“either or”. Tibetan: ’am. When combined with a verb root prefixed with mja1 , probably expresses a possibility. (94) : “As said below, either with sixteen arms, or with six arms” (GALO, Tib: phyag bcu drug pa’m phyag drug pa’m…)  23 (95) ， […] “The Mount Sumeru, four rivers, either twenty three, or nine tress with flowers, should be arranged in order as there are attained”. (GALO) (96) , , : “These, are they gods, or apparitions, or whatever, I cannot realize.” (FIVE) (97) , : “By the speeches of these evil officials [here in this] realm, I am afraid harm will come [to me].” (FIVE, subject in ERG, lit: [Because] of what the evil officials speak in the realm, I am afraid harm will come [to me]). (98) ? ? ? ?: “Is it not emptiness and tranquility?Or is it there attachment? Or is it there vision? Or is it not similar to the emptiness of space? ” (ULTIMATE) Distribution of mo2 is identical with the above, however, in most cases it is a transcription morpheme. Another widely attested interrogative pattern in Tangut is based on the interrogative word thjij2 sjo2 . Although its discussion probably is more appropriate in the Syntax section, the logic dictates to place it here. In the Tanguto-Chinese bilingual sources this word is glossed as ruhe 如何 or yunhe 云何, Tib: ji ltar, ji bzhin, etc. However, the meaning is broader than that. (99) ， ， ， ? : “The first, vajra is originally compared with prajñā, why is that?” (KU) (100) ， ， … “Through which dharmaviddhi one enters the maṇḍala of Śri Hevajra, and according to which krama one receives empowerment…” (SAR, Tib: cho ga ji lta bar bzhin du….rim pa ji lta bar bhzin du dbang bskur ste) (101) : “[Then my] understanding increased, and why it increased?”(GAMPOPA, Tib: rtogs pa la bogs cing byung ji ltar na) (102) ， : “Just in the manner in which whichever (i.e. all of) Supremes in World of the Three time periods awakened their mind of True Enlightenment…” (GALO, Tib: ji ltar dus gsum mgon po rnams rdzogs sangs rgyas nges byas pa) (103) : “In whichever way, it cannot be conceived (i.e. there is no way too conceive it)” (VAJRA, Tib: ji bzhin pa zhes pa mi rtog kyang) Generally, even on the basis of the above example sentences one can infer the existence of discernable combinatory and distribution patterns for the interrogative pronouns and words. These patterns involve matters of co-occurrence of specific pronouns and specific morphemes, currently identified either as verb prefixes or as case markers. However, the identification of these patterns is a research task, which will not be addressed here further. 6.3 Nouns Nominal morphology in Tangut presents little difficulties per se. From the perspective of their structure, the Tangut nouns can be monosyllabic, disyllabic and polysyllabic. Series of nouns can be freely attached to each other, forming appositional strings. E.g.: : “The son of the cakrvartin king, [who is] the Bodhisattva king Khri srong”. Such constructions are easily identifiable in the texts. Disyllabic nouns can be analyzed as noun phrases, where the first constituent is defines the last one. E.g.: lhjịj njij2 , lit.: “the king 2 2 of a state”, i.e. “the king”; ·jiw dzjwo , lit.: “control man”, i.e. “policeman, guard”. Sometimes compound is formed of two monosyllabic nouns: gja1 rjijr1 , lit. “army and horses”, i.e. “the army”. In some cases, the composite nouns demonstrate extended or metaphoric meanings: zjɨr1 dźja2 , lit.: “wisdom sharp”, i.e. “scholar”. A group of nouns, especially technical vocabulary based on Tibetan, demonstrate a more complex structure, which can be analyzed as “complement + verb” phrases, e.g. dzju2 lhjịj : “Lord + bestow”, i.e. “empowerment”, which was translated in the Chinese Khara-Khoto texts as zhushou 主授. This word is a direct translation of the Tibetan dbang bskur. Similarly, some polysyllabic nouns are formed according to the Tanguto- Chinese pattern: dziej2 de̱j1 njij2 : “wheel + turn (Complement+verb: Definition, Tangut pattern)+ king”, i.e.  24 a “Wheel-turning King” (zhuanlun wang 轉輪王). In many cases nouns can be modified with the morphem gjɨ1 , which can be interpreted as an indefinite article. E.g.: ， : “This third root teacher was a person from the royal family from Southern India in the Western Heaven.”; ， “ ” 。 ， “ : “In the Northeastern corner there is a country, it is called the “Country of Tanguts”; in there there is a mountain, it is called “Helan shan”. Similarly with other morphemes, gjɨ1 has multiple functions, among these is the marker of DUAL agreement. Although there are some other functions, we cannot account for them at the moment. 6.3.1 Patterns of Nominalization Tangut employs several productive patterns of nominalization, briefly discussed above. Below we provide more details. Since the derived words perform as nouns, we put them in the Noun section. There is a set varying set of morphemes, identified as nominalizers. Below we specifically discuss five of these. a. Nominalization with mjijr2 . Productive procedure was suffixing verbs and adjectives with mjijr2 , normally understood as the “actor suffix”, equivalent with the Tibetan pa/ ba. E. g.: mjɨ2 lhe̱w2 mjijr2 , 2 2 2 “endowed with composedness”, i.e. yogin, based on the Tibetan rnal ’byor pa; se̱w dza mjijr , “the ones who practice discrimination”; ·wji2 mo2 tha2 mjijr2 : “The Great person Vimalakīrti.” Suffix mjijr2 can be attached to verb phrases, turning them into nominal phrases: : “Those who wish to practice sādhana of Śri Hevajra”; : “Fundamental nature of the one who transmits the ultimate One-Vehicle”. Generally, this pattern presents little difficulty. b. Verb substantives with śjij1 . “Manner of action”. Morpheme śjij1 suffixed to verb roots. Closest parallel is the Tibetan nominalizing particle ba/ pa, or sometimes suffix thabs. Modern Chinese translations fa 法 and shun 順 are misleading. (104) , , , : “Inside of it there were bone śarira of Master Dīpaṃkara, commandments (lit.: “bases of teaching”, rten), practice instructions (lit.: “ways of practice” NOM) and upadeśas.” (GAMPOPA, Tib: bka’ gdams kyi mang ngag, Tangut differs here). (105) ：“ ， ”: “Bodhidharma says: “This practice of calming the mind us similar to contemplating the wall.” (MIRROR) (106) , : “This “cultivation of practices” includes four ways.”(MIRROR) (107) ， ， : “The intention of this text, is that all evil should be extinguished, and all practices must be carried out completely.” (MIRROR, here śia̱1 śjij1 should be translated as “intention” (dgongs pa), e.g.: , : “Casting away self-arrogance and all else, [I] understood the meaning, and collected the intentions of the āgamas and tantras”. (ZHANG, Tib: don rtogs lung rigs dgongs pa rdzogs). (108) , : “The reason for inability of the sentient beings to attain the true Way is that they possess deluded mind.” (THREE, here śjij1 modifies the verb in the main clause.) (109) 《 》 : “This Dharmaviddhi on Initiating the Bodhicitta”. (GALO, Tib: byang chub sems bskyed pa’i cho ga) (110) : “The point of view (lit.: manner of seeing) of Saraha” (ZHANG, Tib: Sa ra ha yi rtogs pa) (111) : “Explained his (i.e. Hevajra) ritual practice” (SAR, Tib: de yi sgrub thabs cho ga bshad) (112) : “Not to make explanation of the meaning of the phrases” (HE) Suffixed to pronouns (adverbial usage) and interrogatives (113) ：“essence of mind abides naturally”(GAMPOPA, here ·jij1 śjij1 represents Tib.: rang bzhin dag du) (114) ， : “Not departing from hearing and seeing, thus one would necessarily attain the Buddha land.” (MG) Here suggested translation for thja1 śjij1 is “thus”.  25 (115) ， : “Accordingly, karmic fruits and retribution will leave by themselves” (DIAMOND) Tangut ·a śjij1 normally translates as “together”, Tib. dang bcas, lhan gcig (116) ， ， …: “Endowed with Six seals, together with the dakinīs, on the Lotus Lion throne” (GALO). (Cf. Tib. dang bcas) (117) …: “After explaining all together the basic meaning of Mind-seal…” (FIVE) (118) : “All together agamas, tantras and upadeśas…” (ZHANG, Tib: lung rigs man ngag dang bcas kun) c. Verb substantives with tjị2 When the example sentences are compared with the Tibetan original, we observe that tjị2 corresponds with the Tib. pa/ ba. In most cases such constructions should be translated with a subordinate clause with “which”. However, one should keep in mind that tjị2 was not completely grammaticalized, and in many cases retained its original meaning “place” (e.g.: 125, 126, 127). Therefore its morphological (syntactic function) is dependent on its spatical meaning, i.e. of an object to which an action is applied. It is treatment as of the equivalent of the Chinese ke 可 is based on the Chinese texts，and is context dependent. Suggested translation should either use subordinate clause (that what), or with verb noun. (119) : “As the limit of Space, it cannot be reached” (VAJRA, Cf. thob pa med, nominalizing construction, lit.: “there is no reaching”, “no place to reach”). (120) : “Because it is one, there is no striving to one side” (VAJRA, Tib.: gcig pas phyogs bcu bsngo dang bral, Tangut and Tibetan not fully congruent here) (121) , : “Since it does not change and cannot be changed, there is not object of attachment” (VAJRA, Cf. Tib.: mi gyur ’gyur ba med pas chags pa’i gnas med) (122) , : “Since One Body spreads all over, there are no dharmas for explaining” (VAJRA, Tib.: sku gcig kun tu khyab pas bsnan pa’i chos med de). (123) [] , : “There is no Way to attain, [] if the reliance (i.e. that on which one relies) and the Way are wrong, [one] will be engulfed by discriminations.” (VAJRA, Tib: phyin pa’i lam, …. rten dang lam nor rtogs pas zin ta re) . (124) ， : “Who can explain verbally that what is cannot be spoken about or conceived?” (HONGZHOU) (125) ， ， : “Great Master, where is the origin of your basic intention?” (lit.: that from where your basic intention comes? HONGZHOU) (126) “ ?”: “Great Master, where is your abode?”(lit.: “dwelling place”, “that where you live” HONGZHOU) (127) ， : “[They] themselves do not know where they came from, how to explain where [they] go?” (HONGZHOU) (128) ， : “In the mountains there are no people, but there is a place from which sound emerges.” (HONGZHOU) d. Verb substantives with lew2 Tangut lew2 has two meanings: This morpheme can be a modal verb, and can be suffixed to a verb stem, forming a verb substantive. As mentioned above, we are not always able to discriminate between these two functions, or to trace regularities in its contrasting distribution with tjị2 . One possible way to interpret constructions with lew2 is passive participle, which corresponds with one specific usage of lew2 as an 2 antonym to sji in combination sji2 lew2 ，i.e. “subject/ object”, as in (137). Translation implies subordinate clause with passive meaning (“that, what”, or “to be+Verb”), or with a verb substantive. In some contexts we translate verb roots suffixed with lew2 as passive verbs, however, there are no straightforward  26 reasons for that, except for stylistic considerations. Similarly with tjị2 , lew2 translates Tib. pa/ ba, and is partially equated with verb constructions with bya, or as adverbial (i.e. for the la don particles). Use of lew2 as of a modal morpheme will be discussed separately. (129) : “Completely abandon the mind [discriminating between] “presence” and “absence”, subject and object”. (ZHANG) (130) ， , …: “With whatever teaching one has been trained, and attained a benefit, in the samsara ....”(ZHANG, Tib: rang gi gdul byar gang bab la ’di yis mi phan pa srid pa’i mtha’… adverbial, Tangut and Tibetan not fully congruent). (132) : “The un-manifested Dharma nature […]” (VAJRA, Tib: snang ba med pa’i chos nyid ni; lit.: the Dharma nature which is un-manifested.) (133) : “I am the one whom the teachers in the world desire for.” (HE, lit.: “that for what” Tib: ’jig rten ston pa bsgrub bya nga) (134) : “The Way, which has no abode, in which there is not object of discrimination (lit “to be discriminated”), and which cannot be conceived.” (VAJRA; here Tib：mi gans dmig pa’i yul med pa mi rtog chos kyi lam). (135) : “Thus, there is no object of attachment (lit to be attached to), and there is no abode for the mind.” (VAJRA, Tib: de bzhin bzung pa’i yul med chos kyi gnas kyang med). (136) : “In the period of contemplation, nothing whatsoever is contemplated.” (VAJRA, Tib. contrasts sgom pa and bsgom)… (137) : “For it cannot be cured with a remedy which can cure.” (VAJRA, Tib: de la gnyen pos bcos shing dgag du med) (138) : “This will be your karma, and there will be no attainment [for you, lit. “nothing to be attained”] whatsoever.” (FIVE) (139) : “I have fully completed the vows, for which you, the King, have desired.” (FIVE) (140) : “Commandments of the Tathāgata should not be exceeded by anyone.” (FIVE, lit: “are not something for someone to exceed.”) (141) , : “This is not something which can be expressed by language, or conceived by mind.” (THREE) (142) , ， , , : “Bhikṣu! Suffering is for understanding; Accumulation is what is cut off; Extinction is what is realized; Way is what is cultivated.” (TWO, here modal usage is possible). e. Verb substantitives with sji2 Some examples demonstrate that the nominalizing function might be connected with the modal meaning “to be able”. In some cases, however, it might be indicative of the tense of a verb, and will be treated below. (143) … ， : “The Great Dharma is the remedy for the most subtle” (VAJRA: chos chen po chung du rnams kyis gnyen po yin). I.e. “is something able to remedy”. (144) : “The ability to contemplate (Verb+NOM) and that what is contemplated are non- dual.” (VAJRA: bsgom dang bsgom par bya ba gnyis med) (145) , : “Not to abandon the three poisons, five poisons and five desires, and make them into adornments (Verb+NOM).” (FIVE) (146) , : “Delusions of the sentient beings become the harm (Verb+NOM) to the foundation of the State.” (FIVE) (147) ， : “Then, bodhicitta is the cause of attainment of the Ultimate Enlightenment.” (BSam) Again, as in the case with the pronominal system one observres partially equivalent distribution between various morphems participating in the nominalization patterns. In some cases one can observe formal and sometimes  27 semantic parameters, but in some cases we are not able to correctly distinguish between the patterns. This remains another research task for the future. As long as all the morphemes discussed above are not fully grammaticalized, they can still be translated though their lexical meaning without violating the message of a sentence. NB: Peculiar way of forming verb substantives is the duplication of verb root. E.g.: dźji̱ j1 dźji̱ j1 ，“stay stay”, i.e.: those who stay (live), i.e. “population”; mji1 mji1 ，“hear hear”, i.e. “what is heard”. This pattern presents little difficulty, and had long since been discussed by the scholarship. 6.3.2 Nominal Morphology As of now, we are not able to postulate the existence of grammatical gender in Tangut. When required by context, feminine nouns were formed through suffixing morpheme mja1 (lit.: “mother, feminine”) or dẹ1 1 1 1 （“feminine principle”）to the root: ŋa dźjij mja (Lit.: “female traversing the space”; based on Tib: 1 1 mkha’ ’gro ma; dakinī ); swew dẹ ，Sanskrit: vidyarājñī. Existence of the so-called “ergative case” in Tangut indicates the existence of discrimination between animate and inanimate substantives. a. Number Number in Tangut is expressed through the affixation of specific morphemes, which are interpreted as indicating number for nouns. a.1．Plural nji2 ，homophone and homograph of the verb 1st and 2nd PL verb agreement marker. When affixed to nouns, indicates plural. E.g.: phjo2 kar2 nji2 , Sanskrit: Vaiśeṣikas, lit. “the discriminators”; tśja1 1 2 nu nji : Tīrthikas (lit. “The ones who refute the Path’); phə1 nji2 ,Tibetans. Suffixed to pronouns: 1 2 thja nji : “they”. Distribution of morpheme nji2 as the plural marker seems limited to the Tibetan texts. a.2． Plural prefix rjur1 . Analogous with the Chinese zhu 諸, “many”; Tib: rnams. (148) : “I should be revered by many people”. (FIVE) (149) : “It is the original root of all of the many teachings.”(FIVE) (150) : “All evils will increase.” (MIRROR) a.3. Plural affix zjịj1 ,“several”, “as many as”. The Chinese equivalents ruogan 若干 and jihe 幾何 are not helpful in interpreting this morpheme. As in many cases, we suggest double function for this morpheme: one is indicative of “several”, “as many as”, i.e. similar to the Tibetan gcig, tsam, ji tsam, zhing，cung zad, etc. Frequent occurrences of adverb zjɨ̱ r1 zjịj1 （ “briefly”, “a little”， etc.）are probably also connected with this usage (151). “Brief” (151) : “Poor monk Zhang G.yu Brags pa briefly composed this text.” (ZHANG, Tib.: cung zad). This is by far the most attested usage. “A few” (152) ， : “[I have not transmitted] a single phrase, which was not said by the masters Tilopa and Nāropa.” (ZHANG, Tib: Tai lo pas ni Nā ro la tshig gcig tsam yang ma gsungs te) (153) ：“The there is not even a slightest particle in this, which can be expressed verbally.” (VAJRA, Tib: de la brjod du rdul tsam med) (154) : “I give you this little gold.” (FIVE) Adverbial: Tib: -tu, -du, -r (155) : “That is, although my understanding became firm, there was no where to advance.” (GAMPOPA, adverbial, Tib: rtogs ba brtan du song ba las bogs ’don rgyu cher mi gda’)  28 (156) : “[It] is an illusion expressed verbally.” (VAJRA, adverbial, Tangut inverses Tib.: ming tshig rgyu mar snang ba ste). (157) ：“ ， , : “That is why the Master Chengguan said: “If every mind is the Buddha, then there is not even one mind which is not the Buddha mind.” (MIRROR, lit.: “as one”) (158) ， : “If taken as not one, it will differ from “cohesion.” (HONGZHOU, lit.: “as not one”) (159) , : “After staying there for a month, they asked the people” (FIVE, lit.: “monthly”, adverbial usage). a.4. Plural suffix ŋewr2 , “many”, “all”. Function of plural suffix derives from the lexical meaning “number” and “to count.” This is preserved in ŋewr2 ljɨ1̣ , “Teaching of calculation”, i.e.: sāṃkhya, grangs can pa. (160) …: “In autumn time, when the people have no leisure....” (FIVE) (161) : “We need to discuss with many scholars what [we] must [do].” (FIVE) (162) : “I reach everywhere here.” (HE, Tib.: kun) (163) : “The masters of old also transmitted only this teaching.” (MIRROR) a.5. Word ŋowr2 ŋowr2 (“all”, “everything”). (164) ， : “At all times, without assistance, it will be impossible to attain supreme enlightenment.” (BCSComm) (165) : “All the sentient beings emerge from me.” (HE, Tib: nga las ’gro ba thams cad ’byung) (166) : “For the benefit of all sentient beings….” (SAR, Tib: sems can thams cad kyi don du) (167) : “All the dharmas are the Mind.” (GAMPOPA, Tib.: thams cad). Here is nominalized with . a.6. Morpheme njɨ2 Morpheme njɨ2 is probably cognate with nji2 , however its distribution is different. It is normally attached to nouns, implying the end of enumeration. Its meaning is clear from context, and will not be specifically discussed here. b. Case markers Latest attempt to arrange Tangut case system in a systematic manner belongs to Arakawa Shintarō. Valueable observations on the nature of some of Tangut auxiliary morphemes belong to G. Jacques. Most of the latest work in this field is generally dependent on the earlier achievements by Nishida Tatsuo, M. V. Sofronov, and most of all by K. B. Kepping. It is a general understanding that the Tangut language does not possess a system of inflectional cases, in the manner the Indo-European languages do. Broader typological considerations, such as absence of gender, allow suggesting that there really were no inflectional cases in Tangut. Functions of various cases are represented by a variety of morphemes, which can be to a certain degree equated with the “particles” of written Tibetan. However, there is no direct congruence even when we are dealing with the texts translated from Tibetan. There are specific markers, partially coinciding, to express subject (NOM), object (ACC/ DAT, direct and direct), and attributive (GEN), instrumental (INS), agentive (ERG) and comitative (COM) functions. Out of these, when marked, ACC/ DAT and GEN are represented by the same morpheme ·jij1 . Its interpretation is thus dependent on its position in a clause: between the nouns ·jij1 represents attributive relation, between a noun and a verb it is an object marker. Further, specific interpretation depends on the semantics of the following verb. Thus in Tangut contrast between ACC and DAT is generally context dependent rather than formal. Most of these morphemes are not free forms, but only emerge as affixes, whereas some of these, such the ERG marker, retain their lexical meaning and sometimes occur as lexical words. At the same time, there is a variety of morphemes, which can be assigned to the so called “local cases”. In some cases, we observe that morphemes, identified as marking the cases, in fact are attached not to a specific word, but can form subordinate clauses. That is, sometimes it is hard to delineate between morphology and syntax in Tangut. This dictated the choice of the example sentences: rather than to quote Tangut nouns in “inflected  29 forms”, we used complete sentences. In this way, the reader might acquire initial knowledge about how Tangut sentences work. Generally, Tanguto-Chinese bilingual sources are of limited value for the reconstruction of the case system in Tangut. Here priority belongs to the Tibetan texts. However, there is no direct congruence between Tibetan particles and Tangut “case morphemes”. What is called Tangut case markers are morphemes suffixed to the respective noun, as in Tibetan. Thus, there is a tendency to identify these as “suffixes” (e.g. G. Jacques). This identification is probably correct for some of morphemes, but not for others, whose identification as “postpositions” is equally legitimate. This means that the Tangut “case markers” probably belong to the realm of syntax and not morphology. However, for a Reader with limited purposes, the issue of exact definitions is of secondary importance. b.1. Subject (NOM) Tangut does not have a specific marker (or, has a zero marker) for the subject of a clause, i.e. NOM coincides with the root. Morpheme tja1 , which is sometimes treated as the indicator of NOM is equivalent with the Tibetan isolating/ topical particle ni, including its usage in the verses to maintain the meter. Here its positions are completely symmetrical with the Tibetan ni. Literal translation should thus be: “concerning …”. However, its distribution in the Tangut texts do not fully reproduce Tibetan pattern (e.g. 150). Syntactically, tja1 participates in a well-known equational pattern with ŋwu2 (“to be”): “Something tja1 is ŋwu2 something else”. There is evidence that tja1 could perform the capacity of the semi-final particle (169). (168) , : “Concerning all the dharmas, [they] are became like outer skin.” (GAMPOPA, Tib: sngar gyi chos thamns cad phyi shun du song nas…) (169) : “Concerning that sometimes understanding becomes broad, that is truly the understanding itself.” (GAMPOPA, Tib: res ni nyams la rgyangs song ba snyam byed pa ste de nyams rang du ’dug). Here tja1 coincides with ste. (170) , : “However, what concerns me, in order to make joyful the minds of those who are beginners in learning….” (GALO, Tib. ’on kyang kho bos ni gzhan dag las dang po pa phal cher yid ’jug par bya) (171) ， : “Contemplation of all the karma of mind as of the purity is the wisdom consecration”. (ZHANG, Tib: yid kyi sgrib pa dag pa ni dbang gsum pa ye shes shes rab dbang ngo, Tibetan and Tangut not fully congruent). (172) ， : “Concerning the original foundation, it is the mind of the sentient beings.” (MIRROR) b.2. Object Marking. Morpheme ·jij1 Tangut object is suffixed either with zero or with the morpheme ·jij1 . Here one must keep in mind that object marker (ACC) and attributive marker (GEN) and (DAT) coincide in Tangut, i.e., there is an opposition between direct and oblique cases. Morpheme ·jij1 is suffixed to noun, it objectifies, but can apply to a clause as well. Application of ·jij1 seems to be limited to the indirect object, and further with the issue of transitivity of the Tangut verbs. This requires further research. Zero (173) : “Master Saraha learnt the art of mixing medicine (ACC) from Nāgārjuna.” (MComm, here in fact adeverbial: learnt how to mix medicine) (174) : “This king had two queens.”(FIVE) (175) […] , : “[The king] had [the horse] loaded with two dou of gold sand and two golden patras.” (FIVE, causative verb)  30 ACC with ·jij1 (indirect object, Tib la) (176) : “Great Compassionate Ones, all of you keep me (ACC, Tib.: bdag la) in your thoughts” (BCP, Tib: thugs rje chen po rnams bdag la dgongs du gsol) (177) : “With its help, not-one not-different illuminating contemplation of the “true sign” (ACC) is transmitted.” (KU) (178) “ ” ， ， : “The word “metal” [means] permanent, [that is] it (ACC) cannot be destroyed with other things.” (KU) (179) : “Homage to (ACC) Arya Avalokiteśvara .” (GALO, Tib: la) (180) : “The Master consecrated (accepted) Rakṣita (ACC) and others.” (FIVE, ergative sentence with causative verb: lit.: “made them accepted”; object marked with ·jij1 ). DAT with ·jij1 (181) : “This little gold I give (DAT) you (PL).” (FIVE) (182) : “[He] gave Kumara （DAT）a golden garland.”(FIVE) GEN Tangut GEN (attributive) can be expressed either with or without morpheme ·jij1 . GEN unmarked (183) : “Cakravartin king of the Bod realm” (FIVE) (184) : “Ear organ of the king Khri-srong…” (FIVE) GEN with ·jij1 (between nouns and/ or pronouns). Distribution of the Tangut GEN is not fully congruent with the Tibetan kyi: , translates yid du ong pa’i gnas. Tangut does not require attributive here: “a place of mind enjoying”. (185) ， : “Transmission from ear to ear of (GEN) these is not to be exposed to others” (VAJRA, ’di rnams….kyi rna ba nas rna bar brgyud pa las gzhan du gleng bar mu bya ’o) (186) : “[I am] Flaming jewel, lord of (GEN) all scholars.” (FIVE) (187) : “Rakṣita paid homage at the feet of (GEN) the Master.” (FIVE) (188) : “Intention to the benefit of (GEN) all the sentient beings…” (FIVE) (189) , : “He will delude the sentient beings and will be able to harm the foundations of (GEN) the state.” (FIVE) (190) , ：“That what becomes the Buddha, is the fundamental nature of (GEN) the sentient beings.”(HUIZHONG) (191) : “Only the Buddha can know his virtue.”(HUIZHONG) b.3. Instrumental Tangut instrumental is represented by two morphemes ŋwu2 and bju1 .Among these, ŋwu2 represents instrumental in true sense, however, its application is quite broad: Alongside indicating “instrument”, it can also indicate purpose or reason for an action, or a manner in which an action proceeds (adverbial), not unlike Tibetan kyis. Note, that although it is not incorrect to translate ŋwu2 with the Chinese yi 以, their distribution is not identical, e.g.: (197), where yi 以 is not necessary in Chinese. The second morpheme, bju1 indicates the “basis” on which action takes place, or its reason, therefore Chinese translation sui 隨 is appropriate. The morpheme bju1 was not completely grammaticalized, and preserves its lexical meaning “foundation”, “basis”, and is used to translate Tibetan rten in specific contexts. In the examples sentences we mark the position of the morpheme. Instrumental with ŋwu2 Manner of action (la don)  31 (192) : “Unconsciously (lit.: “without self-control”, INS), love and faith rose in him.” (FIVE) (193) , 谍 : “Then, having attained (lit.: “through attainment of”) the mind of the Ultimate Realization (INS), he spoke a gātha about the hardships of the Master.” (FIVE) (194) : “The Master came fast (INS).” (lit.: “with fast feet”. FIVE) (195) , : “Praising (INS) one by one the king, his assistants (?), his temples, etc., he said these words” (FIVE) (196) : “Maintain me with the true mind (INS).” (GALO, Tib.: yang dag nyid du bzung bar bgyi) (197) , ?: “If one understands in this manner, it is necessary to practice traversing the bhūmis (INS)?” (HUIZHONG) Instrument (198) , : “Intending (lit.: “with the mind”) that those who have not yet understood this truth, [could] realize, [the teaching] is transmitted (INS) with words and phrases (INS).”(FIVE) (199) ..: “With (INS) this light, Bhagavān emerges from the vajra.” (GALO, Tib: ’od zer des bcom ldan ’das rdo rje las byung pa…) (200) :[It] should be understood through (INS) the wisdom of self-awareness (VAJRA, Tib.: rang gis mtshon te shes par byos) (201) , : “The discriminations dissolved into the light, [and the light] was no longer disturbed by (INS) the discriminations” (GAMPOPA, Tib.: de nas rtog pa ’od gsal ’ba’ zhig tu song nas der rtog pas mi gnod pa cing byng) (202) : “Adorned with (INS) the garlands of human bones.” (GALO, rus pa la sogs pa’i rgyan thams cad kyis brgyan pa) (203) : “Through [its] merits (INS), let me and all sentient beings, be delivered from all wrongs.” (GALO, Tib: bsod nams gang des bdag sogs ma lus skyon bral nas) (204) , : “In general, things (INS) cannot hinder the principle, and principle (INS) cannot hinder things.”(HUIZHONG) (205) : “To have true self established through the signs of illusory body (INS)”(ULTIMATE) (206) ， : “For example, to make a Buddha image from (INS) pure gold,[one] should first understand [what is] the pure gold.” (MIRROR) Instrumental with bju1 In many cases represents Tibetan bju1 , which Tanguts probably understood as an indication of something on which one relies in doing something else. Its “causative” meaning arises from this. Again, there is not direct congruence between Tibetan and Tangut. Suffixed to the indicative pronouns, turns them into adverbs, “thus”, “in this manner”, etc. (207) ， 《 》 ， : “(INS) Following the creation of Nang chung zhabs, and relying on the ritual manual by the Lord of Merit, Yerpawa, poor monk Zhang G.yu Brags pa briefly composed this.” (GALO, here bju1 stands for la of the original; instrumental ŋwu2 represents nas, Tibetan modal elative, indicative of the manner of an action). (208) : “Abide in the natural manner (INS).” (VAJRA, Tib: lhun gyis gnas pas gzhag pa yin) (209) : “Abide by (INS) one’s own nature.” (VAJRA, Tib: rang bzhin gyis gnas pa) (210) , : “Through most subtle and profound understanding, [it is] able to emerge by itself.” (VAJRA, Tib: ji bzhin rtogs pa phra ba yis de ma thag du rang las ’byung) (211) : “Truly play because (INS) of the all-embracing powers.” (VAJRA, Tib: kun tu chub pas rnam par rol, Tangut here is a literal translation of rnam par rol, we follow it)  32 (212) : “Although, because of the mistakes, perversions as explained as (INS) the Way…” (VAJRA, Tib: ’khrul pas nor ba’i lam du bstan pa yang) b.4. Comitative/ Ablative In Tangut comitative is represented by the morpheme rjir1 ,whose Chinese equivalent is normally yu 與. Judging from contexts, we infer, that the Tangut rjir1 represents both comitative and ablative, both functions are equally well-attested. Contrast between the two functions is contextual, determined by the meaning of the following verb. There is no complete congruence between Tangut and Tibetan: : “Not in accord with the instructions and upadeśas.” (ZHANG, Tib.: lung rigs gdams ngag mi mthung). Tangut ABL does not seem to occur with the verbs of movement or direction, i.e. does not have spatial meaning. Comitative (213 , : “Upon reaching the border with India, they met with (COM) the guards.” (FIVE) (214) : “Naturally contradicts with the most profound upadeśas.” (FIVE) (215) , …: “Concerning Kumara, the head of speed walkers, because [he] originally made heartfelt friends with (COM) Vairocana…” (FIVE, NOTE bju1 in causative meaning) (216) : “There came one person, [who looked] similar with (COM) Vairocana” (FIVE) (217) : “While [he] played with (COM) [other] children…” (FIVE) (218) ， : “If a wise person understands the truth of this, then [he/ she] should proceed in harmony with (COM) the Dharma.”(MIRROR, NOTE ŋwu2 meaning “manner of action”.) (219) : “Together with (COM) the yoginis…” (SAR, Tib: rnal ’byor ma lhan gcig dur) (220) …: “This is similar to meeting with (COM) an old friend….” (GAMPOPA, Tib: snga ’dris kyi mi dang phrag ’dra ba cig byung) Ablative (221) : “Abide in the realm where the prapāñca (ABL) are abandoned.” (FIVE) (222) : “The nature of self-mind is separated from what it relies upon (ABL)…”, (GAMPOPA, Tib.: rig pa’i ngo bo rten med du song…, Tangut and Tibetan not fully congruent) (223) : “Defilements are not separate from this mind (ABL).”(MIRROR) (224) ， “ ” : “Because it [means] separation form the defilements, it is called “vajra”.(DIA, 1 NOTE bju in causative meaning) In general, we can suggest that ABL function is based on the identification of rjir1 with the Tibetan conjunction dang. In most cases, Tanguts used rjir1 ka2 to translate Tibetan dang bral. b.5. Tangut Ergative In some contexts Tangut resorts to the so-called “ergative constructions”, originally identified by K.B. Kepping. The distribution of the ERG in Tangut is connected with the issues of animate/ inanimate and transitive/ intransitive opposition. Theoretical conclusions arising thereupon will not be discussed here. ERG appears in cases when the agent of a sentence is 3rd SG/ PL, and is marked with a specific word dźjɨ ·wji1 ,which translates as “action” or “”deed”. Verb ·wji1 (to do) is the basic stem, thus implying the agreement with agent in 3rd SG/ PL. Object of such a sentence is unmarked or unmarked with the morpheme ·jij1 (indirect object), or sometimes omitted, if it is clear from the context. As is often the case in Tangut, this word was not fully grammaticalized, and  33 is sometimes used in its original meaning: : “Having heard of these actions of [these] two people…” (FIVE). The genre distribution of the Tangut ERG is limited: it generally occurs in the literary narratives, such as (FIVE) and is not frequent in doctrinal or ritual Buddhist compositions, where 3rd SG or PL seldom emerges. (225) , : “In Rakṣita’s dream, the two maidens (ERG) came and presented two cups of wine…” (FIVE) (226) , : “The Master (ERG) consecrated (lit.: accepted) Rakṣita and others, and called them names of Rnam par snang mdzad, i.e. Vairocana and Legs grub.” (FIVE) (227) : “In the middle of the journey the guards (ERG) killed [him].” (FIVE) (228) : “The Master gave Vairocana one spoon of gold.” (FIVE) (229) : “The king (ERG) sent for someone, who believed and loved the Buddhadharma.” (FIVE) (230) ， ， : “Tathāgata (ERG) verbally transmitted the realization of the Dharma realm of the One-Vehicle and the nature of mind of non-discrimination of the One-Vehicle to the Great person Vimalakīrti (ACC/ DAT·jij1 ).” (MComm) (231) : “Vairocana gave the gold to the guards (ACC/ DAT).” (FIVE) One object is marked with ·jij1 , whereas the other is not, i.e., contrast between direct and indirect object applied to one verb. (232) , , , : “Thus, the Master (ERG) bestowed vinaya [on him], and taught the Great Attainment, the Great Master made clear everything, the best way he could.” (FIVE, NOTE: both predicates in causative, hence the translation). b.6. Locative morphemes There is a variety of morphemes indicative of the direction of an action. Generally, Tangut locative particles can be equated with the Tibetan la don, but their distribution is different, therefore there is no one-to-one correspondence. In some cases we observe situations when Tibetan requires allative, and Tangut does not. Contrary situations are also widely observed: : “I pay to the Great One, Bhagavān.” Tangut uses allative do2 , while Tibetan does not. Cf.: Mgon po chen po bdag ’tshal. Also: : “From now on I will truly hold the supreme Three Treasures: True Enlightenment, the Dharma and the Great Assembly (LOC).” Cf.: Tib: sangs rgyas chos dge ’dun te/ bla na med pa’i dkon mchog gsum/ deng nas brtsams te bzung bar bgyi. a. Locative with do2 Miscellaneous cases (233) ， : “[Yogin] truly serves upon guru Bhagavān.” (GALO, Tib.: bla ma bcom ldan ’das yang dag par bsnyen te, Tib without allative) (234) : “Prepare a great donation of food for the spirit protectors of the realm (LOC).”(GALO, Here “locative of object”: Tib: phyogs skyong gi thor ma rgya chen bsham mo). 2 Ablative/ Loc with do (Tib. nas) (223) : “This is an erroneous Way, heard from (LOC) another Buddha.” (VAJRA, Tib. Sangs rgyas gzhan nas thob byed nor ba’i lam) (235) : “[They] came to (LOC) the Master to listen to the Dharma.” (FIVE) (236) : “I do not need gold from (ABL) you.” (FIVE)  34 2 Interrogatives with do (237) ， : “The dharmas did not come from anywhere and cannot go anywhere.” (ZHANG, Tib: gang nas; gang du). NOTE the contrast between and . (238) : “The Truth of equality, wherever it abides…”(lit: “wherever it does not abide”, ZHANG, Tib: gang yang mi gnas mnyan pa’i don…) (239) : “It is unheard of anywhere, and transcends the verbal expression.” (VAJRA, Tib. gang la) La dong usage (this usage cannot be clearly discriminated from the LOC/ ABL. Example sentences demonstrate congruence between do2 and la.) (240) ， : “While staying in a satisfying place (LOC), such as a charnel ground etc., together with the dakinīs, [a yogi]…” (GALO, Tib: yid du ong pa’i gnas su ’dug nas…). (241) : “While I abandon the desires, give me protection (lit. “following after” or “foundation in”) of the ultimately tranquil Dharma.” (GALO, Tib: ’dod chags dang bral ba rnams kyi mchog zhi ba’i chos la skyabs su mchi ’o. Tangut here is not a literal translation of Tibetan, but congruence between do2 and Tib. la is clear, this sentence applies also to the ABL/ LOC category) (242) ， ， : “Then, with (INS) the belief into the Lama (LOC), similar with the intention of seeking the consecration…” (i.e. as if asking for consecration from a lama. GALO, Tib: bla ma de la gus pa dang bcas pas dbang bskur ba’i don du gsol ba btab par bsams la…, Tangut here should be corrected: … ) (243) : “[It] abides naturally in all the sentient beings.” (VAJRA, Tib: ’gro ba kun la rang bzhin gnas) (244) ， : “The all-powerful Vajrapanis without exception gather at my place (LOC)” (ZHANG, phyag na rdo rje mthu stobs rnams ma lus bdag la rdzogs par shog) (245) ， : “One person from Dunhuang came to present a vajra treasure to the emperor Wu (LOC).” (KU) b. Locative with ɣa2 (Tib. la/ las/ nas/ la don) Miscellaneous cases (246) ， …: “Relying on the samayas (LOC) and the five ambrosias etc., the ones who wish to practice Śri Hevajra…” (SAR; Tib: …bdud rtsi la sogs pa’i dam tshig bsten pa, dpal kyi mdzad rdo rje bsrgrub par ’dod pas…Tib no LOC) (247) : “Those who are skilled in words and names (LOC), do not understand its meaning.” (ZHANG; Tib: tshigs tshogs mkhas kyang don ma rtogs, no LOC) (248) : “In the broad immesuarability of the all manifest heart” (VAJRA, Tib:mngon sum kun tu (la don) mi rtog pa yangs pa’i snying po la… (LOC) Locative la (249) : “When it manifests inside (LOC) of the mind, it is understood.” (MAH, nang du sems la shar na rtogs pa) (250) : “Even to (LOC) this there were no attachements.” (GAMPOPA; Tib: de la yang ma chags bya ba yin) (251) …: “Because Indians wanted make harm to (LOC) the Dharma…” (LOC of object; lit.: “harm to the Dharma”) (FIVE) (252) : “The Master entered into the state of equipoise (LOC).” (FIVE) (253) ，《 》 : “I have composed Ganacakrakrama as the due reverence to (LOC, do2 ) the gurus and the Buddhas, relying on (LOC) the oral transmission.” (GALO; Tib:  35 bla ma lha’i mtshogs mchod ’os rnams la gus btud nas mtshogs mchod pra khrid gsung sgros la brten ci nus bri gsung sgros la brten. NOTE use of both do2 and ɣa2 in the capacity of la) (254) ， : “All instructions, tantras, upadeśas, all together without exception collect in the Mind of Nāropa (LOC). ” (ZHANG; Tib: lung rigs man ngag dang bcas kun ma lus Na ro’i thugs la rdzogs) Ablative las (255) “ ” ， : “From (LOC) this letter HŪM, spreads the light of five colors and encompasses the space” (SAR, Tib: hum yig de las) (256) : “All the sentient beings arise from (LOC) me.” (HE, nga las) (246) : “Far transcends the realm of verbal expression…” (2885; Tib: spyod yul lam las ’das) (257) ， …: “[Since it] arises from the Diamond mantras (LOC), the unequalled true wisdom can be obtained, and thus…” (GALO; Tib. rdo rje sngags pa las rab byung bas…) (258) ， …: “The purity of the Lotus family, which truly arises from the Supreme Bodhicitta (LOC) ” (GALO; Tib: byang chub chen po las byung ba’i pad ma’i rigs mchog dag pa….) Ablative nas (259) : “For the Dharma, it is not appropriate to arise from (ABL) another [place].” (lit.: “it is not appropriate for the Dharma to rise from another.” VAJRA, Tib: chos ni gzhan nas mi ’gyur) c. Spatial/ Temporal with tśhja̱1 . Generally, tśhja̱1 can be equated with a directional morpheme. However, we are not always able to establish its distribution or its congruence with a specific Tibetan particle. In certain contexts this morpheme appears to coincide with Tibetan semi-final particle, or a conjunction with temporal meaning “when”, “as soon as”, and thus relates to the realm of syntax. This needs further research. As in other cases, tśhja̱1 sometimes emerges in contexts where it is not required by a Tibetan original: ， , : “At this stage, as soon as the disciples enter the stage of “coming out from contemplation”, they give rise to pride in their tutelary deity….” (GALO, Tib: skyes bu thun mtshams kyi lha’i nga rgyal dang ldan pas). Directional morpheme tśhja̱1 In general, this morpheme occurs as a conjunction between main and subordinate clauses therefore should be discussed more extensively in the syntax section. (260) ， ， : “This is as if one ascends to (DIR) a nine-level tower: while he is ascending higher and higher, he sees things which are further and further away.” (KU, quotation from Zongmi) (261) : “While one sun shined on the mountain (DIR) Haspo in Tibet…” (FIVE) (262) : “[They] will come for (DIR) your life.” (FIVE) Temporal conjunction; semi-final particle tśhja̱1 , “allative after verb substatives” (263) “ ” ，“ ” : “While (CONJ) in the Sun platform which emerges from the letter RAM, one should contemplate green letter HŪM” (GALO, Tib: rang gi snying gar RAM las byung ba’i nyi ma la HŪM yig sngon po bltas te; Tangut differs from Tibetan in this instance, ABL las is rendered by bju1 ). (264) , , , ” , “ ”: “While (CONJ) inside of it [the box] there were bone śarira, bases (i.e., fundamental texts, bka’ gdams), training instructions and upadeśas of the Master Dīpaṃkara, [he] urged “Let me see.” (GAMPOPA, Tib: ’di na jo bo rje’i gdung ring bsrel dang bka’ gdams kyi man ngag yod pa yin byas pa kho nan che bas kha phye ba cig rmis te; here tśhja̱1 emerges in the context: yod pa yin byas pa la. Here la allative of time.)  36 (265) (236) …: “While (CONJ) [seated] on the Lion Lotus throne…” (GALO, Tib: khri seng ge’i khri pad ma dmar po’i gnad pa la) (266) : “While contemplating, there is nothing to be contemplated” (VAJRA, Tib: sgom pa’i che na) (267) , : “As soon as (CONJ) [Vairocana] understood everything, Srīsiṁgha said:...” (FIVE) (268) , : “After (CONJ) the seven scholars and the King spoke their speeches, Vairocana answered: ” (FIVE) d. Locatives with ·u2 Generally congruent with Chinese nei 內. Some examples indicate directional meaning. (269) : “Came into (LOC) the Dharma realm free of hindrances” (FIVE) (270) 谍 : “Enter into (LOC) the maṇḍala of Sri Hevajra… ” (GALO, Tib: dkyil ’khor du zhugs) Adjectival usage (271) …: “If it appears in (LOC) the inner mind” (MAH, Tib: nang du sems la) (272) : “Inner, outer and secret, three types of collections…” (GALO, Tib: phyi nang gsang pa’i theg pa gsum) However, basic meaning is “inside”. (273) : “The master practiced while [he] was staying at (LOC) that place for a year.” (FIVE) (274) , , …: “In the meanwhile, the dragons, yakśas and evil spirits in (LOC) this place…” (FIVE) (275) : “In here (LOC), did you see the two Tibetan monks who came in search of the Dharma?” (FIVE) (276) : “This was made known in (LOC) the realm” (FIVE) (277) : “Now, in (LOC) the realm of Tsang there is no such a trumpet.” (GAMPOPA, Tib: dbus gtsang gi rgyal khams na gzangs dung ’di tsam yod de mi yong) e. Locative with kha1 Partial congruence between kha1 and la don reflects in the formations such as zji1 kha1 , Tib shin tu, “the most”. In most cases this morpheme marks subordinate clause in a sentence, therefore will be discussed in more detail in the syntax section. Meaning “among”, “within” (278) : “In the realm, among the Tanguts, Chinese and Tibetans…” (LAW, note contrast between kha1 and ·u2 ) (279) : “Within the Great Vehicle again there are to two [teachings]…” (TWO) (280) : “It is supreme among the Five Sciences”. (281) : “In the world, there is nothing treasured more than children”. (GAMPOPA, Tib: jig rten na bu las gces pa med) (282) : “It is the biggest among the big, and the most profound.” (VAJRA, Tib: che bas che ba’ zab mo yin) (283) : “Those who wish to abandon this world” (VAJRA, Tib: ’jir rten ’di las ’byung ngo zhes)  37 In some cases Tangut requires kha1 while Tibetan does not require a locative (284) …: “The Enlightened Ones from (LOC) Three Time Periods…” (GALO, Tib: dus gsum ’byung pa’i sangs rgyas) (285) : “Supreme among (LOC) the accumulations, this maṇḍala…” (GALO, Tib: bsod nams mthar thug tshogs ’khor lo) Meaning “into” (286) : “Entry into (LOC) the practices of Bodhisattva” (Tib: spyod pa la ’jug pa) (287) : “The discriminations entered into the light.” (GAMPOPA, Tib: rtogs pa ’od gsal ’ba’ zhig tu song, Tib and Tangut are not fully congruent). f. Locative with gu2 This morpheme mostly emerges as an adjective: “central”, “middle”, etc. Locative/ directional usage: Meaning “within” (288) “ ” : “From within (LOC) one’s own mind emerges letter RAM…” (GALO, Tib: rang gi snying gar ram las byung ba’i nyi ma) Meaning “in” This distribution is limited to the words meaning “space” and “dream”. (289) : “Then, in (LOC) a dream there were a golden bottle and a lacquer box” (GAMPOPA: Tib: rmi lam na…) (290) : “In (LOC) the space” (GALO, Tib: nam mkhas la) (291) : “Yogi abides in the way of birds in (LOC) the sky” (VAJRA, Tib. implies attributive usage) 6.3.3. VERB Acccording to previous research Tangut verbs are distributed between the stems A and B. Phonetically the disctribution is marked by a variety of sound changes, including the vowel alternation (Ablaut), consonant alternation, tone alternation etc. Out of these, stem A is identified as the “basic”, that is accepts prefixation and the so called “verb modifiers” (see below). Stem B, is believed to derive from stem A. In addition to the above, accepts personal agreement, i.e. 1st and 2nd SG and PL. Typoligically, the stem alternations account for the Transitive/ Instransitive classes, etc. Tangut verbs roots are generally monosyllabic modified with a number of markers (prefixes, agreement markers, causative suffixes, etc.), indicative of TAM, voice and other features; if a verb is composed of two roots, it is not marked thus its meaning must be established from context. However, in for some disyllabic verbs, the verb prefix is inserted between the two roots, this developing a verb-adverb structure. Tangut uses a variety of “verb modifiers” (verbalizers) which allow transformation between a noun and a verb; if attached to a verb stem, the verbal modifiers and prefixes signal variety of stages of an action demonstrated by a stem: i.e. if an action is complete, or is in process, or is desired (belongs to the future tense), or its result is caused by an external agent (causative/ anticausative). a. Structure of Tangut verb. Tangut verb consists of following constituents: Verb stem, prefix, negative, modifier, tense marker, causative marker, agreement. Position of agreement may vary according to context. In case when a verb consists of two syllables, prefix is inserted between the syllables, marking the main syllable. Tangut verbs can discrimanted into  38 various classes. The main ones are transitive and intransitive, which are marked with specific patterns of sound change. Another group is modal verbs, which generally behave as standard ones. (PREF) (NEG) VERB STEM (MOD) (PAST/ FUT) (CAUS) (AGR) (after G. Jacques). ( )( ) ( )( ) b. Verb Prefixes Tangut language possesses a rich system of verb prefixes. The pairs of prefixes in most cases constitute minimal pairs, and display obvious pattern of sound alternation. This allows suggesting that the prefixes derive from a set of common roots. As far as we can judge there is no obvious etimology for the prefixes in Tangut, and most research on these is comparative and typological, rather than genetic. Tangut “directional prefixes” are the most common elements of the verb morphology, and therefore represent all functions (TAM) except for the personal agreement. That is, it is generally considered that prefixes of A-stem are indicative of perfective aspect, whereas B-stem is identified as “irrealis” (after G. Jacques). Similarly with other Tangut “auxililaries”, apart from being verb prefixes, the morphems in question acquire other functions, i.e. emerge as adverbial markers, as can be inferred from the comparison with the Tibetan originals of the Tangut translations. That is, the prefixes emerge as the equivalents of the termnatives in written Tibetan. However, as of now, we are not able to present a phonologically valid pattern of distribution of prefixes, similar to the one which applies to the Tibetan terminatives. Adverbial function, as our examples demonstrate, is generally limited to the so-called “perfective” stem prefixes, which are in general more widely attested than the “irrealis”. The appropriation of adverbial function might have been a later development. This further leads to a discussion of the origins of the prefixes and their primary function in the language, which will not be pursued here. Below we present examples for the usage of the prefixes, for which we can provide more or less adequate translation. As seen from the table below, the prefixes of different stems contrast in the main vowel, rhyme and tone, at least two contrasts are observed simultaneously. The distribution of the pairs of prefixes is determined either by the distribution of verb stem vowels, or is prescribed by the sematics of the respective verb stems; but in exactly which way, remains uncertain. A stem B stem Translation a single action (interrogative) ·jij1 (durative ) high to low nja1 njij2 low to high kjɨ1 completed action kjij1 desired action Cislocative ·wjɨ2 completed action ·wjij2 desired action Translocative dja2 completed action djij2 desired action Deprivation djɨ2 completed action djij2 desired action Appropriation rjɨr2 neutral (also adverbial) rjijr1 Neutral b.1 Examples with a and ·jij1 Phonetically a and ·jij1contrast, and form a minimal pair. However, the available texts don’t provide enough data to actually observe the contrast between the two. a is one of the most frequent morphemes in Tangut, due to its numerous functions (numeral “one”, interrogative morpheme, verb prefix, the first constituent of the indeclinables with the meanings of “similarly”, “together”, etc.). Distribution of ·jij1 is limited to the verbs of “abiding” (dźji̱ j1 ) and “arising” (śjwo1 ). There examples when ·jij1dźji̱ j1 is followed by a concessive marker tsjɨ1 , suggesting a concessive construction “despite”, “in spite of”. Other than that, we suggest for ·jij1dźji̱ j1 “durative” meaning, similar to the Tibetan gnas pa. a  39 As verb prefix (292) : “Although [Khri srong] (PREF) translated the two Teachings of Cause and Effect…” (FIVE) (293) : “We (PREF) discuss [it] with the scholars” (FIVE) (294) : “I fully (PREF) completed the vows, which were desired by the king” (FIVE) (295) : “This just as (PREF) to meet a person known from before” (GAMPOPA, Tib: sngar ’dris kyi mi dang phrad pa ’dra ba…) (296) ， : “Once [it was] (PREF) understood clearly that the self-nature of Enlightenment has no basis…” (GAMPOPA, Tib: rig pa’i ngo bo rten med du song nas…, Tangut corresponds with song). (297) : “I (PREF) understood that also there is no antarābhava.” (GAMPOPA, Tib: bar do med mnyam pa cig byung) (298) ， ， : “One person (PREF) realizes, and others do not realize, saying so is improper.” (HONGZHOU) As an interrogative morpheme (299) “In here, did you see the two Tibetan monks who came in search of the Dharma?” (FIVE, verb stem additionally marked with demonstrative pronoun tśhjɨ2 ) (300) :“ ?”: “The Master asked: “Do you know what is cessation and continuity?” (HUIZHONG) (301) :“ ?”: “Someone asked: “Do sentient beings attain Buddhahood?”(HUIZHONG) ·jij1 Durative/ concessive (302) ， : “(PREF) Abiding in the world for 517 years, his appearance remained like that of an adolescent…” (KU, here we suggest “durative”) (303) …: “What the the master practiced (PREF) while staying there for a year….” (FIVE, durative) (304) ：“Although the mind of attainment of the supreme Bodhi (PREF) is arising…” (BSAM) (305) , : “(PREF) Standing outside the gate, (they) waited for those who enter and exist” (FIVE) (306) ， : “Although industrious practice (PREF) is arising, it is all in vain.” (MIRROR) Specific usage: “et cetera” (307) : “After [one] truly discards self-arrogance…” (ZHANG, Tib: nga rgyal la sogs gtan spang nas…) Given very limited distribution of ·jij1 , we suggest to independently consider ·jij1dźji̱ j1 寧 as an independent word with syntactical function. Again, dźji̱ j1 , simultaneously with retaining its original meaning “to be in”, “to abide” (Tib gnas pa), was undergoing the process of grammaticalization. E.g. VAJRA uses ziejr2 to translate both bhzugs pa and gnas pa in lexical meaning. b.2 Examples with nja1 and njij2 Similarly with a , nja1 , emerges in the contexts where Tibetan original use cig (gcig). (308) 貼 : “One time [I] (PREF) slept and saw this image in a dream.” (FIVE) (309) “ ” : “The master (PREF) wrote the “Teaching in Five Parts” of the glass jade (i.e. dark blue) clean paper with the golden ink” (FIVE)  40 (310) ， : “Once [he] (PREF) looked north and saw hot lands and dangerous mountains.” (FIVE) (311) : “[My] understating [PREF] increased.” (GAMPOPA, Tib: rtogs pa la bogs cig byung) (312) : “[Since] the self-nature is not [PREF] determined as one…” (VAJRA, Tib: rang bzhin gcig tu ma nges pas) (313) : “Because [one] traverses higher realms, his mind is [PREF] engulfed in the afflictions.” (VAJRA, Tib: mtho ris ’grod pa’i rgyu de nyon mongs zin par ’gyur) (314) : “One sun immediately [PREF] disappeared in the space.” (FIVE) In our texts we generally have no example sentences with njij2 , except for one instance which we cannot interpret adequately. Therefore, we will refrain from discussion here. b.3 Examples with kjɨ1 and kjij1 Morpheme kjɨ1 is one of the most frequent “auxiliaries” in Tangut language. It occurs at least in two attested capacities: verb prefix and the marker of DUAL agreement. It also emerges in Sanskrit transcriptions and with indeclinables, such as kjɨ1 djɨj2 , “necessarily”, “certainly” and in interrogatives. Translations from Tibetan do not allow observing its semantics; therefore we limit ourselves to several examples with relatively straightforward interpretation. In the example sentences below we translate verbs with kjɨ1 through various English tenses 1 generally proceeding from the context. As seen from examples, application of kjɨ as verb prefixe has no direct relationship with the stem of the verb in the Tibetan original, however in the majority we suggest simple Past tense. Following the suggestion of G. Jacques, apart the aforementioned two functions, we suggest the “adverbial” usage of kjɨ1 . As in other cases, we cannot be sure if in its adverbial function kjɨ1 as other morphemes for which we suggest adverbial function, are verb prefixes at all. This function is discussed here as a tribute to the traditional understanding. Examples with kjɨ1 1. Perfective (PS) stem in Tibetan (315) : “These words (PREF) entered the ear organ of the king Khri srong.” (FIVE) (316) : “Rakṣita (PREF) transmitted [his] dream [of] two maidens who came with two cups of wine.” (FIVE) (317) ， ， : “Embark the ship of the wisdom of the Middle Way, cross the stream of the Four Extremes, and (PREF) arrive to the Opposite shore of the Victorious Prajñā.” (MComm, note parallel use of nja1 and kjɨ1 ). (318) ， : “Then [it is said] that [he] (PREF) overshadowed other schools with the teaching of the Middle Way, this is as if the great Sun eclipses the shining of the small stars and the Moon with its light.” (MComm, note: ·jwɨ̱ 1 is prefixed in the first clause, but not in the second). (319) , : “One night I had a dream, in which I cut off head from a child.” (GAMPOPA, Tib: nub cig rmi lam na bu cig yod pa’i mgo bcad pa cig rmis (PS). Here Tibetan rmi ba is in the perfective stem, thus congruent with kjɨ1 , however in this case adverbial interpretation is possible) (320) 《 》 ， : “Following the ritual prepared by Rje btsun Yerpawa, poor monk Zhang G.yu brgas pa (PREF) composed this brief text.” (ZHANG, Tib: rje brtsun mal yer pa bas len mdzad pa’i rim pa la brten nas sparng ban zhing gis cun zad gsal bar byas nas bris pa’ o (PS). Tibetan and Tangut not fully congruent) (321) : “Arranged according to how they (PREF) were viewed.” (VAJRA, Tib: blats pas (PS) tshogs ni bkram pa’o)  41 2. Future stem in Tibetan (322) : “Bhagavan, (PREF) introduce me (CAUS, OBJ AGR) into the fortress of the Great Enlightenment.” (GALO, Tib: thar ba chen po’i grong mchog du mgon pos bdag ni gzhug (PrS) du gsol) Suggested adverbial usage (following G. Jacques) Most of examples for the adverbial usage of kjɨ1 were discussed in the interrogatives section, see also (319). (323) : “In the Dharma Realm of Great Joy, (PREF) which is similar to the Realm of Space…” (FIVE, here the predicate sju2 is prefixed with kjɨ1 , so far this is the only example we were able to find) (324) ， ， : “Not attaining the original sources, studying [while] (PREF) being attached to the heretical ways, [one’s] mind of concentration definitely will not exceed the three realms.” (MComm) Examples with kjij1 These examples are not very numerous, and demonstrate uneven distribution of the two morphemes. We suggest translating it with the present tense, or as an indication of probability. 16 (325) … “Although the words and names (Pref) rely on something else…” (VAJRA, Tib: ming tsam gzhan la brten (Present stem) ’dra yang) (326) …: “In a female body [one] will/ could reach the Eighth bhūmi…” (FIVE) b.4 Examples with ·wjɨ2 and ·wjij2 Translation of ·wjɨ2 as of the marker of perfective aspect is confirmed by the word ·wjɨ2 rar2 (“the past”, i.e. something which is gone) also: ，“ ”: “After this, my name so and so…”, Tib: ’di rjes su bdag ming ’di zhes bya ba… (GALO). Here·wjɨ2 tśjɨ̱ 1 translates Tib rjes su, Sanskrit anu, “afterwards”. However, the Tibetan original word implies an adverbial implication in ·wjɨ2 . 1. Examples with ·wjɨ2 The example sentences reveal double function of ·wjɨ2 : it can function as a verb prefix, and participate in specific adeverbial constructions with meaning “similarly”, “in the same way” or “consequently”, where Tibetan uses bzhin with or without la don. Also, as an adverbial, it might emerge in similar contexts with kjɨ1 including the interrogatives. However, this usage is not frequently attested in our reference texts. Adverbial (327) , ， : “(PREF) Consequnely, all Buddhas and all Bodhisattvas, (PREF) keep me in your thoughts.” (GALO, Tib: de rjes su sangs rgyas dang byang chub sems dpa’ thams cad bdag la dgongs su gsol. Here PREF+ directly translates rjes su; : PREF+ thought+ think. Here the usage is adverbial: “think of me thoughtfully” in literal translation). (328) : “Then, if one’s mind (PREF) thinks in the same way about seeking the secret consecration…” (Lit.: “thinks seekingly” GALO, Tib: de nas gsang ba’i dbang bskur zhu ba’i dons du gsol ba btab par bsams la…Here introduces the manner of thinking, i.e. lit: “think seekingly.”) (329) : “(PREF) Enjoy in here and enjoy in the future…” (VAJRA, Tib: de ltar bde dang phyi mar bde) (330) : “[It] exceeds [all] doing, (PREF) similarly to the emptiness of space.” (VAJRA, Tib: bya bral nam mkha’ bzhin tu gnas) 16 Here there is a possibility of a mistake: kjij1 for tsjij1 “another”.  42 (331) : “(PREF) Similarly with the great Ocean, it abides forever.” (VAJRA, Tib: rgya mtsho bzhin tu ye gnas la) (332) : “Whichever [one] (PREF) thinks of, can be achieved.” (FIVE) Verb Prefix In our texts we do not have parallel examples allowing the usage of ·wjɨ2 in relation to the Tibetan verb stem. 2 Generally, we suggest that the verbs with ·wjɨ should be translated with the Present or Past Tense depending on the context. Generally, we do not accept recent idea that ·wjɨ2 expresses the imperative mood, which is, in our opinion, derived from context, and not expressed formally. (333) : “(PREF) Listen, my friend Grub” (FIVE) (334) : “Then Rakṣita (PREF) revealed his original body.” 17 (335) : “You two now wait (DUAL AGR) [because] the master (PREF) entered the contemplation.” (FIVE) (336) , : “Before a long time passed, the Master (PREF, here directional meaning can be observed) came out of the door…” (337) , : “Not by human actions, [so happened that] the Dharma (PREF) was transmitted to the two Tibetan monks.” (FIVE) (338) : “By the power of original vows [he] (PREF) was born in the Western Heaven.” (FIVE) 2.Examples with ·wjij2 Examples are not numerous in our texts. Suggest translating with Present Indetifinte, or Future. (339) ， : “If even one of my words, which I say, is not in accord with the āgamas and upadeśas, [my] head (PREF) will cleave (cleaves).” (ZHANG, Tib: lung rigs gdams ngag mi mthun pa tshig gcig byung na mgo bo gos.) 18 (340) : “[The master] by his virtue (PREF) will deliver me and the sentinet beings from evil…” (GALO, Tib: bsod nams gang des bdag sogs ma lus skyon bral nas). b.5 Examples with dja2 and djij2 These two morphemes are one of the most frequent in the Tangut texts. Unlike previous pairs, in which the member of the perfective stem enjoys wider distribution than the other, these two morphems emerge with much frequency. 1. Examples with dja2 For the morpheme dja2 we were again establish functions of a verb prefix and an adverbial marker. (341) : “Because of striving towards the great joy, [it] (PREF) became the disease of attachment.” (VAJRA, Tib: bde ba chen po ’dod pas chags pa’i gnas yin te). Adverbial. (342) : “The discriminations could no longer obstruct it” (lit: [It] (PREF) became impossible to be obstructed by the discriminations.” (GAMPOPA, Tib: der rtog pas mi gnod pa cing byung) Adverbial. (343) ...: Then [I realized that] a yogin (PREF) ceases [to exist] but there is no cessation [for him]. (GAMPOPA, Tib: rnal 'byor pa la shi 'am shi med...) Prefix (344) : “I dreamt that I (PREF) killed a big black man” (GAMPOPA, Tib: mi nag po cig bsad pa cig rmis) Prefix 17 In the text·wjɨ1 is used instead of ·wjɨ1 , which in turn is a part of a compound thjɨ2·wjɨ1 “right now”. 18 dźjị2 ne̱ w2 , meaning not clear.  43 (345) : “Thinking of the karma of speech as of (PREF) purified is the secret consecration.” (GALO, Tib: ngag gi sgrib pa thams cad dag par bsams pa gsang pa’i dbang ngo) Prefix. (346) : “Although for the sentient beings (PREF) there are causes and ayatanas…” (VAJRA, Tib: ’gro ba rgyud drug snang ba yang, Tangut and Tibetan not fully congruent) Prefix (347) : “[He] gave [them] two golden coins.” (FIVE) Prefix 2．Examples with djij2 There are examples where this morpheme directly indicates Tibetan optative. Thus we infer two functions: Future tense and Optative. (348) : “In this, if [the discriminations] had been none, then…” (GAMPOPA, not found in Tibetan) (349) ， ” : “Thus, you, the most profound disciple, today will be the one who (PREF) will attain all which there is to know.” (GALO, Tib: de phyir shin tu bu khyod kyis thams cad khyen nyid thob par mdzod) (350) ， : “All the powers of Vajarapani without expections (PREF) should collect on me.” (GALO, Tib: phyag na rdo rje mthu stobs rnams ma lus bdag la rdzogs par shog). Here: djij2 = shog. (351) ， : “(PREF) let me obtain the powers of the master Virūpa without exception.” (ZHANG, Tib: Bir wa pa’i nus mthu rnams ma lus bdag gis thob par shog) Here: djij2 = shog. (352) ， “I realize the empowerements of the secret tantras without exception, and (PREF) attain all practices without obstacles.” (ZHANG, Tib: sngags rgyud ma lus sbyor ba shes las tshogs mtha’ dag thogs med ’grub) 3. Examples contrasting dja2 and djij2 / (353) : [and the light] was no longer disturbed by (INS) the discriminations” (GAMPOPA, Tib.: de nas rtog pa ’od gsal ’ba’ zhig tu song nas der rtog pas mi gnod pa cing byung) (354) ， : “I will attain command of the immesurable upadeśas, and make the all the dakinīs rejoice.” (ZHANG: man ngag dpag med mnga’ bdag byed mka’ ’gro ma lus mnyes byed shog) Here contrast as Tib byed/ byung, semantically different verbs in present and future stems. But in Tangut contrast dja2 / djij2 perfective and irrealis. Despite obvious contrasts, we are not able to clearly disctribute the verbs with dja2 / djij2 between the stems of Tibetan verbs in the originals. 4. Specific usage of dja2 and djij2 Morpheme djij2 prefixes copula ŋwu2 (“to be”) followed by the concessive tsjɨ1 in “concessive clauses” or in the clauses expressing impossibility (although, even if, etc.) (355) , : “(lit.) Although the tongue organ might be skillful, but for this it cannot express even a small particle” (VAJRA, Tib: de bzhin ltse’i dbang pos kyang de la brjod du rdul tsam med. Tangut and Tibetan are not fully congruent). b.6.Examples with djɨ2 and djij2 This pair is imaginary and created on the basis of sound contrast alone. 1. Examples with djɨ2 (356) : “After the unmistaken wisdom of the ultimately pure true nature is attained and (PREF) [its] characteristic obtained, then…” (SAR, Tib: shin tu rnam par dag pa’i de kho na ma nor ba’i ye shes cing mtshan ma rnyed pa dang). Tib perfective stem, translation with Past Indefinite. (357) ， : “Then (PREF) having invited yogins and masters whoever is suited to come…” (GALO, Tib: de nas rnal ’byor pa pha mo ’du bar ’os pa bsogs la)  44 Most examples with djɨ2 are in combination with either ɣju1 (invite) and rjir2 (attain). Constrasive usage of these verbs with djij2 is not observed. b.7 Examples with rjɨr2 and rjijr1 Two morphemes constitute a minimal pair. Out of the two, rjɨr2 is much more widely attested. Again it is probably connected with the double function of the morpheme: as of a verb prefix and an adverbial marker. Examples with rjɨr2 Verb prefix (358) : “You (PL) (PREF) do it in the manner which pleases you.” (GALO, Tib: khyed rjes ’jug dag gar mos par gyis) (359) : “All the dharmas (PREF) did not come from anywhere.” (ZHANG, Tib: chos kun gang nas kyang mi ’ongs) (360) : “[He] [PREF] received the sixteen ascetic practices.” (FIVE) (361) : “[The King Rahula said]: (PREF) explain to me the upadeśas of the Ultimate Realization.” (VAJRA) (362) : “[He] (PREF) went to the master and said…” (FIVE, here adverbial is possible: “walking towards the master, [he] said…”) Adverbial 19 (363) : “What concerns the dharma of the supreme meaning, there is nothing special [PREF] to be said…” (VAJRA, Tib: don dam chos zhes lhag par bshad pa med) (363) : “There is not a tiniest particle (PREF) to be said about the immaterial Dharma body” (VAJRA, Tib: gzugs med chos sku bshad pa rdul tsam med) (364) , : “What (PREF) is said and perfection, what is analyzed and what exists, [are] like an illusion and a picture.” (VAJRA, Tib: gsungs dang rgyas dang dmigs dang snang sgyu mkhan ri mo’i rnam pa bzhin, Tibetan and Tangut are not fully congruent) (365) : “We did not see the two seekers for the Dharma (PREF) leaving.” (FIVE) (366) : “Then the Master (PREF) instructing the King, said:…” (FIVE) (367) , …: “Because Rakṣita did not believe in what (PREF) was said…” (FIVE) Examples with rjijr1 Some of the example sentences seem to indicate to the Future tense of the verb stem. Again, in our texts, the example sentences with rjijr1 are not numerous. (368) , ： “If there is no power of non-attachment and immutability, even if [something](PREF) will be done, it will become bondage.” (VAJRA, Tib: ma chags mi gyo’i stobs med na byas na ’ching ba chen por ’gyur) (369) , ,“ ” : “Halfway, after you give this gold to the guard who is causing trouble, [he] (PREF) will say: “I did not see the passing of the Dharma seekers.” (FIVE) (370) , : “When the speed-walkers (PREF) will catch them, [they] will kill them (lit: “make dead”; FIVE). From the above one can observe that the distribution of prefixes in Tangut is non-proportional: some occur with much more frequently than others. In ours understanding this is connected with the complex nature of prefix morphemes. Some of them tend to combine a variety of functions, whereas others do not. From this, a variety of historical conlcusions might follow. However, its further pursuit is beyond the scope of present publication. 19 西夏译文 śiəj2 可能为 phju2 （“上”）之讹，待考证。  45 7. Verb modifiers and periphrastic constructions Tangut does not seem to have a clearly defined tense system. Theoretically, the two morphemes, ·jij1 and sji2 can be considered the markers of the FUT and PAST, or rather as indicative of an action which is either about to occur, or is happening in the process of speech, or was completed. While the function of ·jij1 is clearly difined, the nature of sji2 is complex: It participates in the patterns of nominalization and in syntactic structures. However, the morpheme is still periphrastic, since it implies an intention to do something. Examples with ·jij1 (future) (371) , : “Later, after Vairocana completed bowing at the Suprmeme master Srisimgha’s feet, he was (FUT) going to return to the land of Bod.” (FIVE) (372) 二 …: “If [one] [FUT] intends to explain this sūtra, there are two parts…” (KU) (373) …: “If [one] (FUT) intends to explain this Collection of the Two Truths According to varios Traditions,…” (TWO) Examples with sji2 (perfective) In case of sji2 , it is not always easy to determine, if it is the marker of “perfective”, or participates in a nominalization, or marks a subordinate clause. The last two cases are mutually interconnected. E.g.: (374), (375). That is, we consider the “perfective meaning” of sji2 as secondary. (374) ， : “That what (PERF/ NOM) Vimalakīrti has become the first root master, is there a reason for that?” (MComm, nominalized clause, verb prefixed with ) (375) , , : “Because risining of the strong vigor is the reason that the fetters increased…” (VAJRA, Tib: rnam rtsol ’bad pa’i sa kun tu zin ta re; Tangut and Tibetan not fully congruent) . 7.1 Verb modifiers and periphrastic constructions We suggest the existence of periphrastic constructions to express a variety of functions. Periphrastic constructions employ a variety of verbs with original lexical meanings “to do”, “to make”, “to become”, “to initiate”, etc. Some of these constructions are better supported by the evidence than others, thus our examples are limited here. Most of the verbs participating in these constructions have not yet been fully grammaticalized; therefore it is not always to easy to determine if the verb in question is used as an auxiliary or in its actual semantic capacity. Periphrastic future is expressed through combination of a lexical verb or nouns with verb modifiers meaning “to strive for”, “to intend”, “to attain” such as kiẹj2 or gji̱ 2 . These are indicative of attaining a new quality, which was not there before, rather than expressing grammatical Future tense. The verbs “to come” and “to go” can also perform in this capacity. The complexity of the situation is such that these verbs have not lost their lexical meanings, and thus it is not always easy to determine if we are dealing with a periphrastic construction or with the verb in its ordinary sense. To indicate this specific phenomenon, we adopt the term “verb modifier”, generally for the lack of a better option. In any case, the translation is dependent on the original meaning of a verb modifier. Therefore, the so called “verb modifiers” remain a hypothetical category, and can be neglected in translation. a. FUTURE Examples with kiẹj2 (lexical “to wish”) (376) … “In order [VM] to collect all the beneficial accumulations…” (GALO, Tib: tshogs rnams bsag bar bya ba…; lit: “willing to collect…”)  46 (377) , : “While [Vairocana] wanted to attain “speed-walking”, the people in the realm asked him…” (FIVE) Examples with gji̱ 2 (lexical “to strive for”) (378) : “In order to abandon [VM] this world…” (VAJRA, Tib: ’jig rten ’di ’byung ngo zhes) (379) : “Even if [one] will be walking along the Path, …” (VAJRA, Tib: ci ste lam la bgrod ’gyur na; verb in the future stem) (380) : “While there is no object for discrimination, strive to attain (FUT) the true wisdom in it…” (VAJRA, Tib: bza’ ba’i yul med de la ye shes ’byung bar btsom) (381) : “In order to traverse bhūmis and places…” (VAJRA) b. Verb modifiers b.1 Tangut causative / phjo2 / phji1 Tangut causative is formed from a verb stem, to which a maker phjo2 / phji1 / , is attached depending on the verb stem. This pair follows well established i/ o sound change, therefore its origin might be traced to a lexical verb “to command” (ling 令). Examples with phji1 This morpheme is generally congruent with the Tibetan causatives. (382) “ ” , “True characteristc”, [means] that the nature of original enlightenment cannot be altered (CAUS) with afflictions.” (KU, lej2 , to alter, stem A) (383) : “So as to abide in the understanding (i.e. to be made to understand) that there is nothing to be attained (CAUS)…” (MA, Tib: ci yang ma yin pa la blo ’jog pa yin te, ’jog pa in causative meaning, Chinese versions of this text from Khara-Khoto uses 安其心). (384) , , : “Again, one night I had a dream, [in which] two people made me blow (CAUS) a paired horn…” (GAMPOPA, Tib: nub cig gi rmi lam na mi gnyis la zangs dung gtan nas ’bud du bcug pa la…, OBJ AGR, mə̱1 , to blow. There some deviations between Tangut and Tibetan here. Tangut =bcug pa) (385) , ：“What concerns me, I have added it here so that to make happy (CAUS) the minds of other beginners.” (GALO, Tib: on kyang kho bos ni gzhan las dang po phal cher yid ’jug par bya ba’i don du ’di bcug pa yin te, Tangut = ’jug par bya) (386) …: “[This sādhana was composed] so that yogins cleanse away [CAUS] vasana and other defilements.” (GALO, bag chags la sogs pa’i ’dri ma dag par bya ba’i don du…Tangut = dag par bya) (387) : “[This sadhana was composed] to make dakinīs happy (CAUS) with this offering…” (GALO, Tib: mkha’ ’gro ma rnams gtor mas mnyes par bya…) (388)   : “Should make (CAUS) offering of the water from the vase of pure karma…” (GALO, Tib: las kyi bum pa’i chus gtor zhing) (389) : “So that all dakinīs rejoice (CAUS)…” (ZHANG, Tib: mkha’ ’gro ma lus mnyes byed shog) (390) : “So that all good intentions of the inner nature of mind are fulfilled…” (VAJRA, Tib: nang gi rang bzhin legs pas re ba thams cad skong) (391) : “[The king] had two golden patras loaded [on the horse, CAUS].” (FIVE) Examples with phjo2 This is a derived stem of phji1 , and thus accepts agreement marker. In our texts examples with phjo2 are less frequent. However, we can observe its congruence with the Tibetan causatives.  47 (392) : “Bhagavan introduce (CAUS) me into the fortress of the great liberation.” (GALO, Tib: thar pa chen po’i grong mchog du mgon pos bdag ni gzhug du gsol. Lit: : “make me enter”, Tib causative: gzhug du) (393) ， 。 [] [] ， ， ， ” : “For the sake of the sentient beings, I will maintain all precepts without exception (CAUS), those who are not delivered should be delivered (CAUS), those who are not liberated should be liberated (CAUS), those who are restless should be relieved (CAUS), and I will introduce (CAUS) the sentient beings into nirvana.” (GALO, TIb: sems can kun gyi don gyi phyir bdag gis sdom pa ma lus bzung ma bsgral ba ni bsgral bar bgyi ma grol ba ni dgrul bar bgyi dbugs ma phyin pa dbugs dbyung zhing sems can mya ngan ’das la ’god. This sentence contrasts in the subordinate clauses and in the main clause.) b.2 Modifiers with meaning “to do”, “ to become” Judging from the Tibetan versions of the example sentences, one can observe, that these modifiers apply to a variety of the Tibetan verb stems. Although we suggest that theses modifiers generally indicate past and present tenses, but there is no absolute certainty to this effect. a. Examples with we2 Auxiliary verb suffixed to either main verb or terminative, indicating that the action expressed by the main verb or terminative is completed, or will be completed. There is obviuous congruence between we2 and Tib. ’gyur ba, byed pa, and song. Generally, one might infer passive meaning for such constructions, as our examples show. Again, the Tangut verb modifiers emerge in the stead of various stems of the Tibetan verbs, as they emerge in the original therefore it is difficult to treat them as formal tense markers. (394) : “It is said: “This will be become visible (VM) the eye of wisdom” (MA, Tib: shes rab mig gis mthong bar ’gyur. Terminative (adverbial) of the result+future/ present stem in Tibetan) (395) : “All the dharmas became similar (VM) to the outer skin.” (GAMPOPA, Tib: sngar gyi chos thams cad phyi shun du song) (396) : [The light] will be no longer obstructed (VM) by the discriminations (lit: the light will become impossible to obstruct with the discriminations) (GAMPOPA, Tib: der rtog pas mi gnod pa cig byung) (397) : “Achievement which became (VM) visible” (VAJRA, Tib: mthong bar gyur pas dngos grub) (398) : “Samayas which emerged [VM] from rising joy” (VAJRA, Tib: dga’ bar gyur pas dam tshig ste) b. Examples with wji1 / wjo1 (to do, to make) In this pair wji1 is considered as basic stem, and wjo1 as derived stem, accepting agreement markers. Both can be affixed to the verb stems and nouns. Both wji1 and wjo1 preserve much of their lexical meanings, therefore in many cases phrases with these verbs should be translated just as verb+complement compounds. Their inclusion into the category of verb modifiers is a mere convention. When affixed to a noun, wji1 and wjo1 should probably understood as “performing such and such capacity”, sometimes a verb noun (393, 391). Generally these can be treated as the markers of Present and Past tenses. Examples with wji1 (399) : “This again does not satisfy me (VM, object agreement, thus basic stem).” (FIVE) (400) : “[The Buddha] orally (lit: “using speech INST”) transmitted (“transmit+VM” to him) the nature of mind of no-thought of the One-Vehicle.” (MAComm)  48 (401) , [] : “Again, one night there was a dream of in which [I] cut off (PREF+cut+VM) a child’s head…” (GAMPOPA, Tib: nub cig rmi lam na bu cig yod pa’i mgo bcad pa cig rmis) (402) : “Tilopa and Naropa did not say (speak+VM) even one word”. (ZHANG, Tib: Tai lo pas Na ro pa la tshig gcig tsam yang ma gsungs te) (403) : “On the way give this gold to the guards harm doers.” (FIVE, here: , harm+to do+NOM suffix) (404) : “If there is examination, it is [VM, lit: “act as”] contemplation of discriminations.” (VAJRA, Tib: brtags na rnam par ’gyur bsgom) Examples with wjo1 Derived stem of wji1 , lexical meaning “to do”, occurs with 2nd SG. In our texts this morpheme mostly occurs in lexical meaning. (405) : “[The master said]: “You make [VM] friends [with him]”. (FIVE, here is indicative of 2nd SG) (406) 《 》 : “At first, [you should] do [VM, 2nd SG] according to the Ritual of Ganacakra for woods, mountains and charnel grounds.” (GALO, Tib: dang po re zhig gans khrod la sogs pa khang pa zhung pa mthung pa der) (407) : “[The Master said]: “Awareness is like this. When you abide by awareness, should not be mentally engaged.” (HUIZHONG) (408) : “What are you going to do?” (FUT, 2nd SG) c. Examples with zjịj2 and zji2 (to do, to make) This pair is hypothetic, since in our texts we have no examples with zji2 , thus its pairing with zjịj2 is based on 2 formal criteria only. As for zjịj , it has limited disctribution, and occurs only in the Recorded Sayings of Nanyang Huizhong in interrogative clauses. (409) ”: “If one has attained the liberation in this way,when [one] encounters cold and famine, what [shoud on] do? ” (HUIZHONG, Chin: 曰：“既得无物自在，饥寒所逼，若为 用心？) (410) : “Who is this doer?” (HUIZHONG) Other pairs, such as dźiəj2 and dźjow1 underrepresented in our texts, therefore their discussion remains for the future research. b3. Examples with śjwi1 and śjwo1 (to begin, to intiate) As in the above case it is not always easy to discriminate between lexical and auxiliary functions of these verbs. As the texts demonstrate, śjwo1 is much more frequent than śjwi1 . This might be due to the nature of the texts from which our examples are derived. Two modifiers clearly constrast as the basic and derived stems of the same root as in: (411) ， ， ， : “I will now generate the unsurpassed mind of enlightenment in the same manner as it was generated by the Bhagavans of the three time periods.” (GALO, Tib: ji ltar dus gsum mkhon po rnams rdzogs sangs rgyas nges byas pa byang chub sems ni bla na med dam pa bdag gis bskyed par bgyi). Here the contrast: (3rd PL, basic stem) and (1st SG, derived stem with agreement marker). As long as both morphems occur with verb prefixes, we can conclude that they retained their original meanings. a. Examples with śjwi1 (412) : “First, I will generate the supreme mind of complete enlightenment.”(MIRROR)  49 b. Examples with śjwo1 (413) …: “The Lous family, which emerges (lit: emerge+VM) from the Great Englightememt…” (GALO: byang chub chen po byung ba’i pad ma’i rigs mchog dag pa…) (414) ， ， ， : “I initiate [initiate+VM] the mind of englightenment, which unsurpassed and most marvelous, for the sake of the sentient beings I will maintain all the precepts without exception.” (GALO: byang chub sems ni bla na med sems can kun gyi don gyi phyir bdag gis sdom pa ma lus bzung). Despite the subject of the sentence in 1st SG, since in the original bdag is in ERG, the verb root modified with the basic stem śjwo1 As mentioned before, the cases where the two morphemes emerge as VM are rare; therefore these should be generally treated as regular verbs. b.3 Examples with śji2 and śjɨ1 (to go, to leave) These two morphemes again can function as both lexical verbs and as modifiers. Mortheme śji2 is generally suffixed to the verbs of motion, such as njɨ2 , dźjij1 meaning “to arrive” and “to set forward”. Morthem śjɨ1 is suffixed to the verb roots and nouns and tends to indicate the purpose of movement, similar to the English constructions with the infinitive. Being derived, it accepts agreement markers. Examples with śjɨ1 (415) : “Went to practice (VM) in the Helan Mountains of the Minyag Realm.” (MComm) (416) : “You two go (listen+VM) to listen” (FIVE, main verb in derived stem, VM in derived stem, DUAL agreement) (417) : “[He] went to have rest (lit: went to go [rest+VM] to rest) in the palace.” (FIVE) (418) : “You go to turn (turn+VM, 2nd SG agreement) the Wheel of Dharma in the Bod realm.” (FIVE) Examples with śji2 Morpheme śji2 is basic stem derived stem: śjɨ1 . (419) : “[They] reached the border of the Western Heaven”. (FIVE, root +VM basic stem, Subject in 3rd PL) (420) : “Half-way, (VM) having reached Arya Palo gling [they] stayed [there].” (FIVE) (421) : “[Vairocana] left for Takri Tsawarong.” (FIVE, root +VM basic stem) Another possible pair dźjwa1 to1/ dzia2 (to complete, to finish) is not supported by our current evidence. b.4 Morpheme djij2 Morpheme djij2 is relatively well researched. As other Tangut auxiliaries it combines morphological and syntactic functions. One of the main functions is marking of the “continuative”: “while he was speaking, I was listening”; marks the predicate of the subordinate clause in complex sentences, i.e. “that” or “although”; in this last capacity can be understood as a conjunction. This last aspect will be treated in the Synatax paragraph. This morpheme can also be interpreted as a nominalization marker. The distribution of djij2 is partially similar with that of sji2 : “ ”: “Discrimination (what is able to discriminate) and discriminated object are in reality the mind.” (MComm). Here is similar to . Nominalazing examples In such cirxumstances djij2 nominalizes not the preceeding verb, but the preceeding clause in general. Thus the translation should be either in gerund, or through a subordinate clause.  50 (422) ， ， ， : “That is, during his staying in the world for 517 years, since his appearance was similar [to that of] an adolescent, he was called “adolescent, i.e. Kumara”. (KU, durative represented by , nominazlized with ). (423) :“ ” ， : “What the masters who practice according to the Āgamas (i.e. the Bka’ gdam pas) say about faith rising from realizing of the truth of emptiness, it is truly correct indeed, [as I realized].” (GAMPOPA, Tib: bka’ gdams pa kun gyi zhal nas stong pa nyid rtogs pa na las ’bras la ’dzem du ’gro…Tangut and Tibetan not fully congruent.) (424) …: “If one thinks in the way of invoking (VM) the secred consecration, then…” (ZHANG: de nas gsang ba’i dbang bskur zhu ba’i don du gsol ba btab par bsams la…) Continuative (425) ， ， : “What concerns the Great Master reverend Galo, sometimes he was doing according to this, and sometimes was not doing according to this (VM)”. (ZHANG, Tib: dpan chen po Rgwa lo ni res ’ga’ ’di mdzad res ’ga’ mi mdzad par gdo, here djij2 directly corresponds with gda’ ba) (426) , : “The king is not alone, [he is] talking with someone.” (FIVE) (427) , , : “The son of the younger queen became able to recite following whaterver the Master was reciting (VM)…” (FIVE) (428) , …: “Then, [while he] was paying (VM) homage to the masters’ footprints, the master asked the Royal Heir…” (FIVE) 8. Negatives Standard Tangut has a set of morphemes with negative meaning. One of these is nja̱2 “not to be”, in contrast with two ŋwu2 meaning “to be” respectively. Other than that, there are negative morphemes mji1 mjɨ1 and 1 1 tji .Out of these, mji is the most frequent, applying to the verbs, adjectives etc. Morphrme mjɨ1 is generally applicable to the so-called “modal verbs”, and is actually the sole marker of the Tangut modals, but the distribution is not exactly clear. Or, as some examples indicate, mjɨ1 adds modal meaning to the verb it modifies. One specific usage of mjɨ1 is combination with the demonstrative pronoun tśhjɨ1 and verb stem. This combination indicates that the action represented by the verb stem is impossible or will never occur (as suggested by Kepping; this, however, is a conjecture). Finally, tji1 is an adverb marking the so called prohibitive mood. Finally, there are verbs mjij1 and m 1̣ ，equally meaning “not to have”. From a phonological perspective, mji1 mjɨ1 and mjij1 are probably cognates. Among the negatives, mji1 has the widest distribution. a. Examples with mji1 (429) : “[All the dharmas] do not exceed the mind” (MA, Tib: sems las phyi rol tu gyur pa med de) (430) : “It is not similar with the previous concentration…” (GAMPOPA, Tib: snga ma’i sgom mi ’dra bar song) (431) , : “Then though arose [in me] that no can harm can [be done me].” (GAMPOPA, Tib: …’dres gnod pa cig mi nus par ’dug) b. Examples with mjɨ1 In most cases occurs in combination with the verb njwi2 , “can”, “to be able”. This case is simple and will not be addressed here. (432) ， : “I will truly uphold this from now on, and (NEG) never abandon [it].” (GALO, Tib: deng nas brtsams te bzung bar bgyi nan du yang ni mi btang ngo) (433) ， : “The dharmas did not come from anywhere and (NEG) cannot go anywhere.” (ZHANG, Tib: chos kun gang nas kyang mi ’ongs gang du yang ni mi ’gro la) (434) : “It was impossible for us to see those who came searching for the Dharma.” (FIVE)  51 (435) : “There is (NEG) no grasping [of it], [it is] like the moon in the water.” (VAJRA, Tib: bzung bas mi zin chu zla ’dra) (436) : “The joy itself (NEG) cannot be obtained.” (VAJRA, Tib: zhi ba de nyid mthing ba min) c. Examples with tji1 , Tangut prohibitive (437) : “(PROH) don’t you listen to these words!” (FIVE) (438) ：“ ”: “Master Vagisvara said: “(PROH) don’t you go to another realm!” (MComm, agreement expressed through the derived stem) d. Copula verb nja̱2 “not to be”. The distribution of nja̱2 is relatively simple: It behaves as a standard copula verb in equational sentences. (439) : “There is no birth and there is no extinction, it is nothing at all (lit: is not anything).” (VAJRA, Tib: mi skye mi ’gags cir mi dgongs, not contrastive usage of mjij1 and nja̱2 ) (440) : “Liberation comes by itself and (NEG) not from the other (lit: is not the other).” (VAJRA, Tib: grol ba rang ’byung gzhan las ming) (441) : “Self-generated true wisdom (NEG) is not existent.” (VAJRA, Tib: rang ’byung ye shes yod ma yin) e. Copula verb mjij1 , “not to have” Again, this morpheme is relatively easy to interpret through its Sinitic counterpart wu 無,whereas in Tibetan it corresponds ma, ma yin, med, med pa etc. (442) : “Since it is not born (lit.: is not having birth), there is no extinction.” (VAJRA, Tib: ma skeys pas na ’jig pa med) (443) : “It is unparalleled true wisdom (lit: wisdom which has no equal).” (VAJRA, Tib: mtshungs med pa’i ye shes kyis) (444) : “Since there is equality, there is no sequence/ gradation.” (VAJRA, Tib: mnyam pas rim par bkod pa med) Constructions implying “impossibility” with mjij1 Morpheme mjij1 participates in negative adverbs, meaning “impossibility” or with a modal meaning of prohibition. Formation of these constructions is connected with the adverbial and nominalizing functions (“that what”) of the nominalizing morphemes lew2 and tjị2 .Although for the Tangut speakers these two constructions had been obviously different, from our current perspective they are not easy to discriminate. a. Examples with tjị2 mjij1 (445) , …: “Although it is (NEG) impossible to go to India, staying in my own realm is also unpleasant…” (FIVE) (446) : “Whatever is said about it, [for the words] it is (NEG) impossible to reach [it].” (MA, Tib: brjod pa’i tshig gang gis kyang thog tu mi phebs pa, Tangut and Tibetan not fully congruent) (447) : “As the limit of space, it is impossible to be reached (lit: there is no reaching).” (VAJRA, Tib: nam mkha’i mtha’ bzhin thob pa med) b. Examples with lew2 mjij1 (448) : “(NEG) Unmanifested (lit: which cannot be manifested) Dharma nature, when it is established without striving, is contemplation.” (VAJRA, Tib: snang ba med pa’i chos nyid ni btsal gzhags pas bsgoms pa yin) (449) : “In the calm contemplation originally (NEG) cannot be contemplated.” (VAJRA, Tib: bsam gtan nyid pas bsam du med).  52 (450) : “In the Dharma realm, whose limit is unreachable, there (NEG) is nothing to decrease.” (VAJRA, Tib: mtha’ la thug pa med cing chos dbyings la bri ba med) (451) : “In it, there is nothing to remedy with remedies (lit: “what can remedy”).” (VAJRA, Tib: de la gnyen pos bcos shing dgag tu med) 9. Tangut modal verbs and markers Most widely attested Tangut modal verbs include ·wo2 and njwi2 , meaning “must” and “can” and wjị2 , “can”, “know”. Apart than that, modal meaning can be expressed by affixiation of the morphem lew2 to the verb stem. One criteria to identify modal verbs is the usage of mjɨ1 in negative phrases, however this rule is not always observed: negative mji1 also occurs with mjɨ1 . In some contexts morpheme ·wo2 preserves its lexical meaning, and translates Chinese “principle” (Chin li 理). Generally, as the example sentences demonstrate, alongside their modal meanings, Tangut modals can express imperative, optative etc. One more way to indicate that an action should or should not be done is through suffixing morpheme śjwo1 . ( , : “[The king]” said: “Since this teaching of Ultimate realization is also beneficial to the sentient beings here it should (MOD) be made prosperous (verb in causative)”. a. Examples with ·wo2 (452) : “The scholars (PL), you discuss what you (MOD) must [do].” (FIVE, 2nd PL AGR) (453) : “All the people must pay homage to me.” (FIVE) (454) : “Masters the gods, (MOD) worthy of reverence…” (GALO, Tib: bla ma lha’i mtshogs mchod ’os rnams) (455) : “If there is no mind, then (MOD) should be no speech?”(HUIZHONG) (456) ， “[One] must not do all evil, but completely excersize all the practices.” (MIRROR, Chin: 斷一切惡，備修萬行). b.Examples with njwi2 As the example sentences demonstrate, while it is correct to translate njwi2 as “can”, it has other application, such as expressing general possibility in the future or present, and optative. (457) : “[Ultimate wisdom] (MOD, CAUS) can make [reach] the ultimate place of Bodhi and nirvana”(KU) (458) : “If one truly realizes the all-abiding principle of equality and hold firmly to it, then [one] (MOD) might attain realization.” (ZHANG, Tib: gang yang mi gnas mnyam pa’i don nges rtogs rab brtan gzhan phan shog). Here construction is used for the Tibetan optative. (459) : “It (MOD) can swiftly appear from itself.” (VAJRA, Tib: de ma thag rang las ’byung). Here MOD implies future tense, Tib ’byung. (460) : “One will not be able to transcend the three wordls where there are no objects to be attained.” (VAJRA, Tib: rnyed pa’i yul med ’jig rten gsum las yong mi ’byung) (461) : “All the dharmas (MOD) can/ will emerge.” (VAJRA, Tib: chos rnams sna tshogs ’byung bar byed) c.Examples with lew2 This morpheme can express imperative, whereas sometimes there seems to be no direct equivalent for it in the source sentences. (462) : “The scholar (MOD) should understand in this way.” (VAJRA, Tib: mkhas pas de bzhin shes par gyis). Here represents Tib gyis, imperative of byed pa.  53 (463) : “In the time of contemplation, [one] should contemplate nothing at all.” (VAJRA, Tib: sgom pa’i tshe na ci yang mi bsgom mo) (464) , : “I also have some supernatural powers; (MOD) should make them manifest (CAUS).” (FIVE) (465) ：“The body which was born, must perish, thus [one] must be diligent in practicing the True Dharma.” (FIVE) (466) : “The king Khri srong must abide peacefully.” (FIVE) (467) , ， , , : “Again, bhiksus, suffering must be known, accumulation must be cut, extinction must be attained, and the Way must be attained.” (TWO). This sentence allows translation through nominalizing. (468) ， ， ， ，《 》 : “The second, concerning the master who wrote [this], as in the text: “composed by Śramana Dehui from Helan mountains”, the ten supreme virtues of this master should be known from what is said in the Record on the Entry of the True Dharma.” (MComm) (469) “ ” ， : “This letter HUM should be seen as green” (GALO, Tib: HUM yig sngo bo blats te) (470) : “It is something which must be realized through self-awareness.” (GAMPOPA, Tib: rang gis mtshong te shes par byos) (471) : “Must be recited for three times.” d.Examples with wjị2 (472) : “It cannot be obstructed by discriminations.” (GAMPOPA, Tib: der rtogs pas mi gnod pa cig byung) (473) : “The adversaries will not be able [to affect it]. ” (ZHANG, Tib: rgol ba kun gyis mi thub shog, Verb in optative) (474) , ..: “Whichever [mantra] the master was reciting, [he] became able to recite accordingly…” (FIVE) (475) : “The nature of original enlightenment cannot be changed by the objects and afflictions.” (KU, verb in CAUS) 10. Tangut verb agreement Verb agreement is a relatively well studies aspect of Tangut morphology. Kepping, Hwang-cherng Gong, G. Jacques among others, contributed substantially into the study of the subject, which is generally clear now. The distribution pattern can be consulted in Jacques 2014. Theoretically, Tangut verb agrees with the subject in 1st, 2nd PL and SG, and similarily with the object (e.g.: : “A friend said to to me.”). Object agreement is not as frequent as subject agreement. Verb agreement is a complex procedure, involving variety of patterns, depending on the properties of subject and object of a sentence as well as valence of the predicate verb. Generally, when the SUB is 3rd SG, and an action is addressed to 1st or 2nd SG and PL, Tangut applies object agreement involving the basic stem. In other cases the derived stem is used. In most cases, when verb agreement is used, the subject is often omitted. Recently identified is also the agreement in DUAL. Agreement is expressed by the alternation between verb stems A and B, as well as by suffixing of the agreement markers, which may or may not be cognate with the respective pronouns, as mentioned above. For 3rd SG and PL the agreement is expressed by the verb stem, while no specific markers are attached. Generally the agreement markers are as follows: ŋa2 ，nja2 / ，nji2 ，kjɨ1 ，gjɨ2 ，tsjɨ1 .Written Tibetan does not have the agreement, we provide the source sentences for reference. We only introduce the cases with more or less obvious AGR pattern. a. Examples with ŋa2 , subject and object agreement, cases with no marking  54 20 (476) ， : “(AGR) I will truly tell you the secret mantras, ways of practice and the viddhis.” (GALO, Tib: gsang sngags spyod tshul cho ga ni khyod la yang dag bstan par bya) 1st SG, SUB AGR (477) ， : “The unequalled mind of enlightenement, I will in this manner genenrate (AGR) [it] in my mind.” (GALO, Tib: byang chub sems ni bla na med dam pa bdag gis bskyed par bgyi). 1st SG, SUB AGR (478) : “I wrote (AGR) [this] in accordance with the Āgamas and upadeśas.” (GALO, Tib: lung dang man ngag ji bzhin bri) 1st SG, SUB AGR (479) : “This again does not satisfy me (AGR).” (FIVE) 1st SG, OBJ AGR (480) : “Tell me the ultimate upadeśas.” (FIVE), marked as object, verb with : 1st rd SG, OBJ AGR, SUB: 3 SG (You master…). 21 (481) : “See me (AGR) off to Nepal (?)” (FIVE), 1st SG, OBJ AGR (482) , : “Because of [my] donation, teach me (AGR) the upadeśas.” (FIVE). Verb modifier in basic stem; 1st SG, OBJ AGR, SUB: 3rd SG (You, master…). (lit: “make me the one to whom the upadeśas are taught”). (483) : “After seeing teacher for a while, he immediately transmitted the upadeśas to me, without exception.” (GAMPOPA, no Tibetan). OBJ AGR. Subject in 3rd SG, OBJ 1st SG; verb in basic stem. (484) ， ， : “These samayas should be given to me, and the mind of enlightenment should be given to me, Dharma of the Buddha and three refuges of the Great community also should be given to me.” (GALO, Tib: dam tshig de nyid bdag la stsol byang chub sems khyang bdag la stsol sangs rgyas chos dang dge ’dun te skyabs gsum dag kyang bdag la stsol). OBJ AGR, Subject 3rd SG, OBJ 1st SG, 3rd PL, OBJ marked with , hence no verb marking. This sentence is probably in passive voice. (485) , : “Sumpreme in the World, I should be introduced into the Fortress of Great Liberation.” (GALO, Tib: thob pa chen po’i grong mchog du bdag ni gzhug du gsol). SUB AGR, SUB 1st SG; verb in CAUS, derived stem, no marking. Suggest translation in optative. b. Examples with nja2 / , subject and object agreement (486) : “My friend, Grub, listen.” (FIVE) SUB AGR, SUB 2nd SG. (487) , : “Master, in which land you will attain joy of your heart?” (FIVE) SUB AGR, SUB 2nd SG, verb in derived stem. (488) : “You (AGR) are as if two vessels mutually pouring the upadeśas.” (GAMPOPA, no Tib, lit: you are as if upadeśas are pouring from one vessel to another) 2nd SG, SUB AGR (489) ， : “In this way, you, the foremost disciple, will today (AGR) obtain everything which has to be known.” (GALO, Tib: de phyir shin tu bu khyod kyis thams cad mkhyen nyid thob par mdzod) 2nd SG VM in basic stem, SUB AGR (490) : : “The king said: [I] make you the root teacher for the royal heir.” (FIVE). 2nd SG, CAUS in derived stem, OBJ AGR. (491) : “You go (AGR) to the realm of Bod to turn the Wheel of Dharma.” (FIVE) 2nd SG, VM in basic derived stem, SUB AGR. c. DUAL Agreement Since neither Chinese nor Tibetan has clearly identifiable dual number, the existence of a specific form of agreement is inferred on the basis of the exemplar sentences. DUAL with kjɨ1 , 1st DUAL (examples provided by Zhang Yongfu 张永富) 20 dźjɨ tjɨ ̣j2 spyod tshul 21 pja1 pu1 In Tibetan : bal yul, Nepal.  55 (492) ：“ ， ， 。“The king told the lady: “Do not be upset, let your heart be at peace, we go together to look search…”(GOLDEN LIGHT, Chin: 王又告夫人：“汝莫生煩惱， 且當自安慰，可共出追尋). SUB AGR, SUB 1st DUAL, verb in derived stem. (493) ：“ ， ?: “[He] asked his wife: “Me compared to Xü Junping, who is beautiful [among us]?” SUB AGR, 1st PL (12) DUAL with tsjɨ1 ,2 nd DUAL (494) ：“ : “The king said: “[You two] give [this gold] to the Indian scholars of the ultimate upadeśas, and then listen.” (FIVE) SUB AGR; Subject in 2nd DUAL; verb in derived stem. DUAL 1st PL inclusive ( … ), pronoun might be omitted. (495) ：“ , ， : “Jing gong in a dream saw two children saying to each other: “Doctor Huan of Qin is coming now, we certainly die.” (12) SUB AGR, SUB 1st DUAL, verb in derived stem. (496) , , ：“Although we two are not the great scholars, following the king’s order, [we two] will go to listen [to the teaching.] ” (FIVE), 1st PL, INCLUSIVE, verb in derived stem. Plural Agreement with nji2 ,1st and 2nd PL (497) : “You (2nd PL) can do as you please” (GALO, Tib: khyed rjes ’jug dag gar mos par nd gyis). SUB AGR, Subject 2 PL, VM in basic stem. (498) “By the power of previous vows, [we](1st PL) meet at this moment.” (FIVE), SUB AGR, st Subject in 1 PL, verb in basic stem. (499) : “All scholars, you [should] discuss what must [be done].” (FIVE), SUB AGR, subject in 2nd PL. (500) : “[We 1st PL] came here without reason (tentative)” (FIVE), SUB AGR, subject in 1st PL. (501) ， : “All the compassionate ones, keep me in your thoughts.” (BC, Tib: thugs rje cen po can rnams kyang bdag la dgongs du gsol). SUB AGR, SUB in 2nd PL. (502) ， ， ， ， : “So as to clarify [these] (AGR), the “Great” in amaniskara [means] it is great because of the substance, it is great because of its characteristic and great because of its function.” (MComm). OBJ AGR, Here the function of nji2 is agreement between 1st PL and rd 3 PL of object. 11. Verbs of existence and possession. There is a variety of verbs which are traditionally translated as “to be”, “to be present”, “to have” etc. In the Chinese texts from Xixia, these verbs are rendered through Chinese you 有, zai 在, etc. Below we list the main verbs of existence with some notes pertaining to the usage thereof. The verbs of existence follow similar distribution pattern as other verbs: change of initial consonant or syllabic vowel, tone signals distribution between two basic stems; the relationship between the stem alternation of the verbs of existence and possession and verb agreement is not very apparent; nature of examined texts allows suggesting that some verbs were more popular than others, regardless of the stem distribution. Generally, the verbs from Stem 1 agree with the 3rd SG; pair 2 maintains opposition between animated and unanimated subjects, pair 3, among other things, appears to contrast elevated and standard speech. From below one can observe that pairs 1, 2, 3 and 4 are probably cognate, whereas other four verbs appear to be isolated. Orthographically, one can observe that the verbs dju1 , dźjo2 , dźjij2 , share similar graphic components, only partially supported by phonetic correspondences or stem alternation paradigm. Generally, from the table below on can suggest that the original paradigm was based not on two but on three verb stems: e.g. dźjij2 ，dźjo2 however, these alternations need more research. Otherwise, the meanings of existence are rendered through two basic verbs ŋwu2 and nja̱2 , traditionally rendered as the Chinese shi 是 and fei 非. As long as lhe̱ w2 and w ̱ 1 are concerned, these two morphemes participate in a specific  56 nominalization pattern, involving possession or attribution of certain qualities or features, according to the Tibetan pattern based on ldan pa. For the sake of reading texts, the exact distribution of the verbs between the stems is not vital, however, it is important to note that the general pattern of the stem alternation is maintained in the verbs of existence as well as in the other cases. Careful study of the texts might reveal the pattern of distribution of these verbs, we are inclined to assume that this pattern is generally semantical, and follows similar lines as the stem and prefix alternations in other verbs. Distribution of antonymous verbs translated as Chinese wu 無, is even less certain, what is listed below is a preliminary table of correspondences; correspondence in the line 1 is purely semantical, since Tangut mjij1 etymologically derives from a negative marker mI. Provisionally, the verbs can be arranged in a table as below as minimal pairs. However, we cannot be sure that the table represents real distribution of the verbs and opposition between the stems. Stem 1 (basic stem) Stem 2 Possible antonyms (all translated as 無) 1. du1 dju1 to have, to be mjij1 The Stem 2 verb present with the widest distribution in the Tangut texts 2. dźjij2 dźjo2 To have, to exist, to mjij1 Stem 1 verb is the be present second most frequent Tangut verb of existence 3. wjij1 ·wjij2 To be, to be present mə2̣ Stem 1 “elevated speech” 4. tśhjo1 tśhju1 To possess, to exist mjij1 Also: “to include”, “to contain” 5. lhe̱ w2 To be, to have, possess, to be within 6. wə̱1 To be, to relate to, Translates through to belong to Chinese 7. ·o1 To exist, to be in, to To be at, to belong, be at to exist, translates as zai 在 a. Examples with du1 and dju1 In our texts we do not have examples with du1 , whereas dju1 is one of the most wide spread Tangut morphemes meaning “to have”, “to be” etc. Depending on context, it can represent Tibetan snag ba, and attributive ldan and can. Other equivalents include srid pa. Genenrally, this morpheme is equivalent with yin pa, and in the negative with med pa. (503) , : “Since there are some previous vows (i.e. karmic connections), I come here.” (FIVE), SUB in 3rd SG. (504) : “Although I have realized the principle of the dharma realm which is beyond verbal expression (lit: does not have words)…” (FIVE), SUB in 3rd SG. (505) …: “All existent (manifest) dharmas”. (MA, Tib: snang zhing srid pa’i chos thams cad)  57 (506) , : “Sometimes [the dharmas] are imagined as really existing, sometimes are seen as only the essence of mind.” (GAMPOPA, Tib: res ni chog ge ba lta bu yang byung res ni shes pa rang gi ngo bo cig pur song). Here dju1 translates chog ge ba, approximately meaning “straight, truthful (syn. drang po).” (507) : “In the world, there is nothing more treasured than children.” (GAMPOPA, Tib: ’jig rten na bu las gceg pa med pa ste) (508) : “All the disciplinary commandments…” (GALO, Tib: sdom pa kun dang ldan pa ni thams cad…) Duplicantion dju1 dju1 returns translation: “all there is”. b. Examples with dźjij2 and dźjo2 Generally, these morphemes can be roughly considered as equivalent with the Tibetan yod pa. Similarly, meanings equivalent with can, ldan are also possible. That is, in some cases we observe a specific pattern of nominalization, similar with Tibetan involving ldan and cas. Other Tibetan equivalents are also possible, depending on the context. Examples with dźjij2 (509) : “I dreamt of cutting the head of a child (lit: of a child that was there).” (GAMPOPA, Tib: bu cig yod pa’i mgo bcad pa cig rmis) (510) , ： “All the Great Merciful Ones, all the Knowledgable guides.” (GALO, Tib: rje chen po dang ldan pa thams cad mkhyen pa thams cad ston pa) (511) : “Make ganacakra for the ones possessing samayas.” (GALO, Tib: dam tshig can la tshogs ’khor bya). Here nominalization pattern is observable. (512) : “In there is power, then the world will be without self-nature.” (VAJRA, Tib: stobs su gyur na ’jig rten rang bzhin med) (513) , : “If [one] is endowed with faith and diligence, whatever [one] desires for will be achieved.” (FIVE) (514) , [] : “For the benefit of the sentient beings, I came seeking for the Dharma, [] to the [one] endowed with power of great compassion, [the one] who is like a wish-fulfilling gem” (FIVE). Here dźjij2 nominalizes preceeding clause. Examples with dźjo2 As a derived stem, dźjo2 accepts agreement markers. However, patterns without agreement marking also observable. Contrast between dźjij2 and dźjo2 is in the absence of nominalization patterns with dźjo2 . (515) , : “I also (AGR) have supernatural powers, should be shown.” (FIVE) (516) : “You (AGR) truly have the siddhi of speed-walking.” (FIVE) (517) : “Is there a reason not to meet them?” (FIVE), no AGR c. Examples with wjij1 and •wjij2 In our texts we have not discovered examples with wjij1 , whereas •wjij2 is widely attested. From contexts we suggest translation “to exist”. Generally, •wjij2 appears more frequently in the Sinitic texts, and its distribution tends towards interrogative sentences. E.g.: , : “If one practices daily, and does not abandon the “no-practice”, is there anything [one] will not practice?” (HUIZHONG). (518) : “The tops of all precious pagodas existing in India bowed down.” (519) ， “ ” 。 ， “ ”: “In the Northeast there is (exists) a country, called “the country of Tanguts”, in this country there is a mountain, called “Helanshan.” (MComm). (520) ， ：“It is not heard anywhere (lit: not exists for hearing), and is beyong verbal presentation.”(VAJRA, Tib: gang la thos pa mi rigs bsnyad pa kun dang bral) (521) ， : “Sprouts emerge from the ground, is there another arising [for them]?” (HONG)  58 (522) ， : “In the mountains there are no people, but there is a place from which sound emerges.” (HONG) (523) : “There is nothing among the dharmas in the world, to which it can be compared.” (HUIZHONG, lit: “if to use the dharmas of the world, then there is nothing comparable”). d. Examples with tśhjo1 and tśhju1 Morpheme tśhjo1 is unattested in our texts, whereas tśhju1 is quite frequent. Suggested translation is: “to be endowed with”, and “contain”, attested Tibetan yod pa. Morpheme tśhju1 is well attested in the composite word , endowed with life, i.e: “the sentient beings”. (524) , : “The ones, who are transmitting the Dharma, are endowed with body of stone and many eyes.” (FIVE) (525) ， : “Since [he] was endowed with the mind seeking the Way, he wanted to go to India.” (MComm) (526=104=264) , , , : “While (CONJ) inside of it [the box] there were bone śarira, bases (i.e., fundamental texts, bka’ gdams), training instructions and upadeśas of the Master Dīpaṃkara, [he] urged “Let me see.” (GAMPOPA, Tib: ’di na jo bo rje’i gdung ring bsrel dang bka’ gdams kyi man ngag yod pa yin”. Here tśhju1 translates yod pa yin. (527) , : “Listeners are on the mountain, with a great lake (lit: endowed with a great lake).” (FIVE) e. Examples with lhe̱ w2 “to possess”, “to be endowed with a quality”, participates in a nominalization pattern 1 1 (Tib ldan pa); w ̱ , ·o . lhe̱ w2 (528) translates Tib Ral pa can, “The One with Hair” (FIVE). (529) translates Tib rnal ’byor pa, lit: “possessing silence”, i.e. “the One possessing equipose”. (530) translates dpal ldan, “possessing glory”. (531) : “Endowed with six types of mudras” (GALO, Tib: phyag rgya drug dang ldan pa) w ̱ 1 .This morpheme occurs quite often, and currently seems to indicate “belonging”, although we are not able to provide a more in-depth discussion. (532) : “Tibet of the Mi”, i.e. parts of Tibet belonging to Xixia, Mdo smad. f. ·o1 , to have, to be Currently, attested examples confirm correspondence with Tibetan yin pa. This morpheme participates in interrogative patterns. (532) : “Because in this there are many faults…” (VAJRA, Tib: de yang rnam pa’i skyon yin pas) (533) , , …: “The one who transmitted the Dharma is endowed with the body of stone, and has is the one with the long iron lips.” (FIVE) (534) : “On the head of each one of them there is bright gold (tentative).” (DIAMOND) (535) “Although there is an ear, but it cannot hear the sound”(ULTIMATE, Chin: 有耳對聲而 不能聞 ). SYNTACTIC PATTERNS Tangut language with richness of various auxiliaries offers a perspective for the research of syntactic and grammatical patterns. Here we are not prepared to delve into linguistic peculiarities. Suffice it to say, that many among theTangut verbs are polivalent, and can participate in the constructions invlolving variety of complements and adjuncts. As in any other language, the phrases are distributed among the endocentric and exocentric structures. Among the exocentric structures, COMPLEMENT+VERB is the most common (see examples in 5). As Hwang- cherng Gong once observed, endocentric/ exocentric opposition is observable not only in the syntactic patterns, but in the structure of specific words, which can be understood as verb phrases, and even in the structure of the Tangut  59 logographs. As in other languages, we can specify several types of sentences in Tangut: declarative, interrogative and exclamatory sentences. Generally, Tangut sentences tend to be long and utilize a variety of conjunctions to indicate the relationship between the clauses. In some cases, Tangut sentences begin with a morpheme·iọ1 ，and complete with ljɨ1 .However, none of that is mandatory, and an indeclinable ljɨ1 ljɨ1 is often used in the beginning of a sentence or a paragraph (e.g. in the legal codes). Conjunctions are summarized in the Table below. Generally, Tangut is SOV language, however, since Tangut has verb agreement function, subject is often omitted, and its grammatical features are demonstrated with the agreement markers. As a rule, Tangut observes strict order of words in a clause and order of clauses in a sentence. Theoretically subordinate clauses preceed main clause, adjuncts preceed complemenets. So far, we are not able to observe formal patterns for voice in Tangut, therefore our selection of translation is context dependent. Tangut voice needs further research. The best way to analyze a sentence is from right to left, starting from the predicate of the main clause, and then proceeding further through the chain of subordinate clauses. Subject of a sentence (where present) normally occupies head position, in cases where it is omitted, it is expressed through verb agreement markers or lack thereof. E.g.: : “ , , ” , . In this sentence we identify following markers which allow initial punctuation of the sentence: verb suffix ·a , locative/ nominalizer tjị2 , morpheme tsjɨ1 , conjunction bju1 , causative marker phji1 , marker of quotation ·jɨ2 , auxiliary morpheme ljɨ1 , which marks the end of a sentence, locative ·u2 and finally the verb prefix ·wjɨ2 . The analysis is as follows: the abundance of markers in the sentence indicates that we are dealing with several subordinate clauses and one main clause. Since the next sentence in the text deals with a different subject, we can infer that the predicate of the main clause is expressed with ·wjɨ2 tsji̱ r2 , without agreement marker. That is, the subject is probably 3rd SG. Further to the right is the word zjɨr1 dźja2 , “scholar”, which can be interpreted as the complement, i.e. “select the scholars”. Further to the right is locative adjunct: “in his state”. Reflexive pronoun ·jij1 indicates that the subject of this clause is the king Khri srong, therefore we translate the sentence in active voice. Further to the right is marker bju1 , showing that the the material to the right from either chronologically preceeds or is the reason for the main clause to the left, i.e is a subordinate clause. The next to the right is auxiliary morpheme ljɨ1 showing that the material to the right from it is a complete 2 sentence, while preceeding ·jɨ introduces direct speech, i.e. Tibetan zhal nas or zhes bya ba. That is, the main clause of the sentence is thus: “Because the king Khri srong thought: [] he selected scholars in his kingdom”. The subordinate clause is thus located between ·wjɨ2 sji̱ j2 and bju1 . Subordinate clause can thus be analyzed independently: predicate features the verb wejr2 ljịj2 with both modal and causative markers, meaning that something should be made prosperous. The complement is not easily discernable, since is does not directly preceed the verb: next to the right is word nio̱w1 , “reason”, thus indicating the reason why something should be made prosperous. Further to the right is a clause: , “benefit or sake of the sentient beings”, which is in turn produced by the teaching of the Ultimate realization, which is thus the complement of the subordinate clause. Finally, thjɨ2 tjị2 is is adjunct indicative of the place, i.e. “here”, “in this place”. Proceeding in this manner we arrive to the actual translation: “[Because the king Khri srong thought: [[[“The teaching of causes and effects is prosperous in my country]]]; because this teaching of Ultimate Realization is beneficial for the sentient beings]], it should be made prosperous here”], in his state [he] selected the scholars [to go to India].” Here the the first clause is independent sentence; The reason for the main clause (selection of scholars) is indicated by bju1 , suffixed to the subordinate clause. Since the reason is the thoughts of the king, it is marked as a quotation with the morpheme ·jɨ2 , The subordinate clause marked by bju1 , which shows that the material to its right represents the reason of what is to the left. is again analyzed as a complex sentence: the reason why the teaching should be propagated is represented lexically with the word nio̱w1 , meaning “cause” As mentioned before, ambiguity in interpretation of the case system further indicates what we sometimes cannot discriminate between syntax and morphology: in many cases the morphemes which we identify as morphological, can in fact be applied to the whole phrase they modify. This might present problems for a linguistic analysis, but is irrelevant from the point of view of translation. Thus, as soon as the verb of the main clause is identified, translation becomes smooth. Verb is easily identified either through the presence of prefixes, and/ or agreement markers. E.g: , ,  60 , , , . Here ŋwu2 (INS) is indicative either of the manner of action, or shows that the action in the modified clause was completed. Morpheme kha1 marks spatial adjunct (in the mandala); śjɨ1 is a verb in a subordinate clause, and is the predicate of the main clause, which is 2 2 analyzed as VM we , preceeded by PREF dja , which serves as an adverbial marker, and actual adverb expressed with a verb phrase mji1 śja2 , “not to see”. Thus, the translation is “to become invisible”. Here it is an adverbial phrase, which might be reconstructed as Tib: mi snang bar gyur. For convenience, a comma was inserted after the indeclinable “together”. That is, this declarative sentence can be analyzed as: Main clause: “The Master Vairocana became invisible in the sky”, with two subordinate clauses: (1) “having revealed illusory representations of the Five Buddhas in his body” and (2) “together with his [physical] body, in the mandala of light ascended upwards”. Each of these clauses can be analyzed further, e.g. (1) is an ajunct of the manner of action, with its own verb, complement and spatical adjunct: (LOC) (OBJ) . This analysis demonstrates that INS , belongs to the whole phrase. (2) is analyzed as follows: (LOC), (LOC) (VERB): spatial adjunct 1 (place), spatial adjunct 2 (direction), verb. Since spatial adjunct 2 uses morpheme bju2 , which apart from its lexical meaning can probably be identified as an adverbial marker, the main verb śjɨ1 is not prefixed. In this manner we arrive to the translation: “The master Vairocana, having revealed illusory representations of the Five Buddhas in his body, together with his [physical] body, in the mandala of light ascended upwards, and became invisible in the sky.” (FIVE). That is, the main problem in reading Tangut is identifiying the verb in both main and subordinate clauses and correct interpretation of conjuntions. 1. Phrase connectors/ conjunctions Tangut sentences are normally long, which requires usage of phrase connectors between clauses of a long sentence. In many cases, the homonyms and homographs of postpositions are used in the capacity of phrase connectors between main and subordinate clauses. Marker meaning Explanation Translation kha1 Circumstantial condition or “While”, “whereas”, connection circumstance in which the “in such circumstances main clause occurs as” tśhja̱1 Temporal connection temporal connection “While”, “during”, “in between main and the duration of” subordinate clauses ljɨ1̣ ŋwu2 mjɨ2 djij2 Concessive Standard form of “Although”, “though”, concessive “despite”. ljɨ1̣ Phrase connector Between words or “and” clauses tsjɨ1 Concessive Between words and “although” clauses bju1 Instrumental Between two clauses “since”, “according”, “because” ŋwu2 Instrumental Between two clauses “because”, “through” ku1 Conditional Between two clauses “if, then” Declarative and interrogative sentences Normally, in Tangut topic is marked with tja1 (Tib ni) or unmarked. Basic structure: tja1 …ŋwu2 (Something is something else). However, application of the morpheme tja1 is context dependent. a. Topic/ Comment  61 (536) , : “Concerning the explanation of the basic principle of the self-nature of the true mind (TOP) as the original source of all the dharmas in the samsara (TOP) is the true mind.” (FIVE) (537) , : “Concerning the True Mind, which originally is inexistent and free from sensual cognition, it is the foundation of all the dharmas.” (FIVE) (538) , : “To concentrate on the understanding of this principle as of a “single” (TOP), it called “meditation”(MA, Tib: rtog pa’i don de la sems rtse gcig du ’dzin pa sgom pa) (539) , : “Concerning all the dharmas (TOP), they became like an outer skin”. (GAMPOPA, Tib: sngar gyi chos thams cad phyi shun du song) (540) , : “This moment was [the time since when] the devils could no [longer] harass and harm [me] (TOP). ” (GAMPOPA, Tib: nga yongs bdud la sogs pa’i bar chad kyis mi tshugs nyam pa la de kho na byung, Tangut and Tibetan not fully congruent) b.Subordinate clauses with kha1 (declarative and interrogative) Here we not consider cases when kha1 applies to a single noun, i.e. in its locative capacity. We belive that the adequate translation will be “while”, implying a specific circumstance in which the main clause occurred. This usage seems to be best represented in the native Tangut compositions. (541) ， : “While I abandon desires (LOC), give me refuge in the the supreme nirvana.” (GALO, Tib: ’dod chags dang bral ba rnams kyi mchog zhi ba’i chos la skyabs su mchi’o, Tangut and Tibetan not fully congruent) (542) , : “While (LOC) [it] receives hundreds, thousands, tens of thousand and million names, no matter how [one] practices, a flower will produce a fruit” (VAJRA, Tib: rnam grangs brgya stong phrag yas pa gang ltar spyad kyang me tog skye) (543) , : “While (LOC) the Buddha turned the Wheel of Dharma in the Three Worlds, [he] spoke a lot about discipline.” (TWO) (544) ， ， ， ？: lit: “Concerning (TOP) this fisrt roor master, while (LOC) Tathāgata must be called [as such], that what (NOM with , TOP) the Great person Vimalakīrti became the first root master, is there a reason for that?” (MComm, translation: “What is the reason for the Great Person Vimalakīrti to become the first root teacher, while in fact it is Tathāgata, who is the first root teacher). Here kha1 represents the circumstances for the main clause to occur. (545) ，“ ， ， : “While (LOC) there had been other st teachings, [their members] said: “We (1 PL, inclusive) composed [?] treatises, and although we exhausted our minds, but [these texts] are nevertheless useless.” (MComm) (546) ， : “While (LOC) [the master Dehui] was teaching the Middle Way and pramāṇa, he obtained supreme teaching of no-thought from the Master Brtson ’grus.” (MComm) c. Subordinate clauses with tśhja̱1 (allative/ illative, generally congruent with Tib la, na) In general, morphological and syntactic functions for this morpheme are not easy to discriminate, below are sveral examples which we deem representative of its syntactic function. (547=526=104=264) , , , ” , “ ”: “While (CONJ) inside of it [the box] there were bone śarira, bases (i.e., fundamental texts, bka’ gdams), training instructions and upadeśas of the Master Dīpaṃkara, [he] urged “Let me see.” (GAMPOPA, Tib: ’di na jo bo rje’i gdung ring bsrel dang bka’ gdams kyi man ngag yod pa yin byas pa kho nan che bas kha phye ba cig rmis te; here tśhja̱1 emerges in the context: yod pa yin byas pa la. Here la represents allative of time.)  62 (547) ，《 》 : “This was bestowed on the [master Galo], upon (LOC) seeing the face of the tutelary diety in the mandala of Śri Cakrasamvara, and thus (INS) “Brief ritual of consecration” was composed.” (GALO, Tangut colophon, no Tib) (548) 《 》 , , : “ ”: “While (LOC) teaching “Garuda in Flight” in a manner comparabale to the “inner clothing”, the master asked: “Is the Dharma satisfactory?” (FIVE) (549) , : “In (LOC) the time of meditation, there is nothing to be contemplated.” (VAJRA, Tib: sgom pa’i tshe na ci yang mi bsgom mo) (550) , : “While (LOC) everyone had no leisure in the autumn time, [they] came to the master to listen to the teaching.” (FIVE) (551) , : “Since [he] became a scholar in the world, the name of G.yu sgra Snying po was called upon.” (FIVE) d. Tangut Concessive with ljɨ1̣ ŋwu2 (nja̱2) / mjɨ2 djij2 Tangut concessive almost never occurs in the full form, so in the example sentence it is quoted as it appears in a variety of forms. Generally the structure consists of a conjunction ljɨ1̣ , copula verb (positive or negative) and negative compound mjɨ2 djij2 . Suggested translations: “although”, “despite”. (552=17=545) ， : “Although our mind might be exhausted (CON), [our treatises] are nevertheless useless.” (MComm) (553) , : “In this, although (CON) my understanding became solid, there was no increase in it.” (GAMPOPA, Tib: rtogs pa brtan du song ba las bogs ’don rgyu cher mi gda’) Here Tangut mjɨ2 djij2 corresponds with Tibetan las. (554) , , : “Despite (CON) we [two] are not scholars, following the order of st the Sacred King, we go to listen (AGR, 1 DUAL)” (FIVE) (555) ， ， ?: “Although (CON) I say that it is not similar [with anything], but [saying that] there are no merits, can it be true?” (HONG) e. Tangut concessive with tsjɨ1 As mentioned above, this morpheme generally has equivalent distribution with the Tibetan kyang, yang. (556) , : “In previous time, although I recited many mantras, was often afraid.” (GAMPOPA, Tib: sngon tsa na bzlas pa mang po byas kyang ’dre tshur gnod) (557) , : “Now, whichever scholar I might (CON) meet, I am [no longer] confused.” (GAMPOPA, Tib: da lta ni sangs rgyas byon kyang nyam mi nga) (558) : “There is no attachment even to this.” (GAMPOPA, Tib: de la yang ma chags bya ba yin) (559=170) , : “However (lit: Although this is so), what concerns me, in order to make joyful the minds of those who are beginners in learning….” (GALO, Tib. ’on kyang kho bos ni gzhan dag las dang po pa phal cher yid ’jug par bya) (560) : “Although the self nature various accumulations is non dual…” (VAJRA, Tib: sna tshogs rang bzhin mi gnyis kyang) (561) : “Although [it] does not emerge in whichever way…” (VAJRA, Tib: don rnams ji bzhin mi bskyed de) (562) , : “Even wrath, desires and stupidity, are born from the way of great enlightenment.” (VAJRA: ’dod chags zhe sdang gti mug kyang byang chub chen po’i lam las byung). Meaning “also” (563) : “Thus there is no object for attachment, and the place for the mind also does not exist”. (VAJRA: Tib: de bzhin bzung pa’i yul med chos kyi gnas kyang med)  63 (564) : “Truly self-emerging wisdom also exceeds the limits of words and names.” (VAJRA, Tib: rang ’byung ye shes nyid kyang tshig gi mtha’ dang bral) (565) , ：“Because here this Dharma of the Great Ultimate will also benefit sentient beings” (FIVE) f. Conjunction ljɨ1̣ , “too, and, also” (Tib dang), complex conjunction ljɨ1̣ nio̱w1 So far, this morpheme is the most widely attested Tangut auxiliary. Generally it is analogous with the Tibetan dang (“and”, “too”), however in combination with other morphemes it participates in a variety of other patterns, including a specific version of concessive, is used to maintain meter, etc. There are several cases for which we are currently able to provide explanation. (566) : “Kun dga’ ma (CONJ) and Bdag nyid ma” (FIVE) (567) : “All sorts of causes and fruits of retribution are manifest” (FIVE) (568) , : “As soon as the seven scholars (CONJ) and the king completed their discussion, Vairocana replied…” (FIVE) (569) , : “Then the Head translator, king (CONJ) and officials and the people…” (FIVE) (570) , : “Attained the essence of the upadeśas (CONJ) and fully realized intentions of the scholars” (FIVE, connecting two clauses) (571) , : “Because is exceeds the three vehicles (CONJ), this Dharma should be explained…” (FIVE, connecting two clauses) (572) ， : “There are [those] who fully comprehend the meaning of teachings and are able to perform basic rituals” (LAW) (573) ， : “Hommage to Dpal ldan dkyes pa’i rdo rje (CONJ) and Bdag med ma” (GALO, Tib: dpal ldan dkyes pa’i rdo rje dang bdag med ma la phyag ’tshal lo) (574) ， …: “Family, merits (CONJ) and enjoyments, and wisdom of unmistaken understanding of the words” (ZHANG, Tib: rigs dang yon tan longs spyod dang ma nor tshig don shes rab dang) Concessive pattern with ljɨ1̣ (575) , , : “Although (CON) this is not a different teaching, the two Tibetan monks who stayed here before are truly the holders of upadeśas.” (FIVE) (576) : “Although (CON) [we] obtain the Dharma, but it will not be successful” (FIVE) 22 (577) , , : “After that, although the Tibetan king Khri srong has died (CON), his son Ral pa can invited Indian pandita Vimalamitra.” (FIVE) Maintaining meter (578) , : “Let the powers of Vajrapani (CON) collect on me without exception.” (ZHANG, Tib: phyag na rdo rje’i mthu stobs rnams ma lus bdag la rdzogs par shog). Here there is no Tibetan equivalent for ljɨ1̣ , thus we suggest it is used to maintain the meter. Miscellaneous cases (579) : “Liberation rises by itself, not from (CONJ) the other” (VAJRA, Tib: grol ba rang ’byung ghan las ming). Here ljɨ1̣ corresponds with ablative las. (580) : “With frown countenance in the mandala (CONJ), it is imagined in the body of a wrathful king…” (VAJRA, Tib: dkyil ’khor khro gnyer cha lugs kyis khro bdag chen po’i lus bzung nas). Here ljɨ1̣ corresponds with Tibetan INS kyis. 22 pjij1 mja1 lja1mju2 tjɨ1 rjar1 Indian pandita Vimalamitra, Tangut returns *Vimalamudra.  64 (581) , : “While it is not possible (CONJ) to go to India , returning to my own place is also unpleasant.” (FIVE, parallel usage of ljɨ1̣ and tsjɨ1 , possibly similar to English “either/ or” or “neither/ nor”) g. Clauses with bju1 (instrumental and various cases) Clauses may be connected with bju1 , with implying that the subordinate cause is the reason or condition for the main clause. When suffixed to separate words the morpheme is indicative of the reason or cause or source (“from”, “because”, “for the purpose of”, etc) of the following action. Distribution of bju1 is partially equivalent with Tib las, rten, kyis, etc. In some cases we observe its correspondence with terminatives (586). Suffixation to words (source, adverbial usage) (582) , : “From (CON) the ultimate truth only manifests non-dual self-awareness.” (TWO) (583) , : “All three places also arise from (CONJ) me” (HE, Tib: nga las gnas gsum po yang ’byung), ablative las. (584) ， : “By (CON) the benevolence of Vajraguru, this sādhana and viddhi were composed.” (GALO, Tib: rdo rje slob dpon bka’ drin gyis de yi sgrub thabs cho ga bshad) Instrumental usage. (585) : “Self nature is certainly not as one (ADV)…” (VAJRA, Tib: rang bzhin gcig du ma nges pas). Here , gcig du, adverbial usage. Suffixation to subordinate clause (reason, perspecive, mode of action, etc) (586) ， ...: “Because (CON) Saraha could play musical instruments, master Nagārajuna studied [with him]” (MComm) (587) , , …: “Contemplation from (ADV) the perspective of the mode of general meaning, [means that] all dharmas are empty and unborn…” (GAMPOPA, Tib: don spyi'i tshul du bsgom pa dang [...] chos thams cad stong nyid skye ba med). Here bju1 corresponds with terminative du. Adverbial (588) , , : “By means of (INS ) of the eye of wisdom, [which originates”] from the door (CON= ) of the supreme masters’ blessing, this will be seen.” (GAMPOPA, Tib: bla ma’i byin rlabs sgom pa yin shes rab mig gis de mthong ’gyur”.) Here bju1 represents sgom pa understood as intstrumental. (589) ， : “Because [CONJ they are] afraid of the water demons, [the sentient beings] should be delivered from the ocean of samsara (Verb in CAUS).” (GALO, Tib: chu srin la sogs kyis ’jigs pas srin pa’i mtsho las sgrol byed cing). Here = ’jigs pas, INS (590) ， : Because (CONJ) you are a vessel of Mahāyāna, you should listen to the son of the great ritual.” (GALO, Tib: khyod ni theg pa chen snod yin pas tshul chen po ’di la bu tshul shog, Tangut and Tibetan not fully congruent). Here: = yin pas (INS) (591) : “In whichever way (CONJ) it was brought into harmony by myself…” (ZHANG, Tib: rang gi gdul byar gang bab la). Here there the Tibetan equivalent for bju1 is gang bab, represents gdul byar (adverbial ). Lexical meaning (to rely) (592) ， : “Realize that the nature of original enlightenment is devoid of support.” (GAMPOPA, Tib: rigs pa’i ngo bo rten med du song nas…). Here = “support”, adverbial construction: “to rely upon”. h. Clauses with instrumental ŋwu2 There are frequent cases of homonymity (tongjia 通假) between ŋwu2 and copula verb ŋwu2 .  65 The distribution of example sentences between the categories is a matter of expedience rather than accuracy, since most cases allow alternative interpertations. Instrumetal clause (manner of action) (593) , : “In this, through (CONJ) resolute wrathful conduct, there will be no name “this is the place of abiding.” (VAJRA, Tib: ’di la brtul zhugs drag shul gyis gnas su bya ba’i ming med de). Here ŋwu2 translates kyis. (594) , …: “Imagining (CONJ) [oneself in the] the body of the Great Wrathful king, although the letters will appear clearly…” (VAJRA, Tib: khro bdag chen po’i lus bzung nas yi ge mngon du byung na yang…). Here ŋwu2 translates Tib elative nas (ablative of origin). (595) , : “The Great Ultimate joy of non-production is obtained through (CONJ) the absence of desires.” (VAJRA, Tib: bya bral zin pa’i bde ba ’dod pas med pas len)” Here standard instrumental applied to a nominalized verb phrase. (596) : “It cannot be remedied with (CONJ) what can remedy” (Nominalization of the the first clause with ; VAJRA, Tib: de la gnyen pos bcos shing dgag tu med) (597) : “Passing through (CONJ) much difficulty [they] came from afar”. (FIVE) (598) , : “With (CONJ) presenting mandala, [they] again asked for the teaching…” (FIVE) Clauses representing condition or reason (Tib na) (599) , : “Because the one who can contemplate and the the object of contemplation are non dual, [yogi] abides in the yogic manner in the realm of the supreme truth.” (Note contrasts between two usages of ŋwu2 : instrumental and adverbial; VAJRA, Tib: bsgom pa dang bsgom par bya ba gnyis med na don dam ngang la bde bar gnas pa ni) (600) ， , : “Because (CONJ) the teachings of cause and effect were translated and transmitted in this realm, they became prosperous…” (FIVE) (601) , : “Because (CONJ) the master gave out gold to the guards one by one, he left Inida.” (FIVE) (602) , : “Because (CONJ) of the accordance with the upadeśas (i.e. because of what is said in the upadeśas), I went to Tsawarong.” (FIVE) (603) , : “Because (CONJ) you took the donation of instruct me in upadeśas (object AGR).” (FIVE) (604) ， : “Because these [masters] loved what is above, and despised what is below, they practiced four contemplations.” (MComm) i. Conditional clauses with ku1 The usage of ku1 is quite straightforward, and can be equated with English “if… then” or Tibetan “gal te; na” or Chinese ruo…ze 若…則 (full form in 608: tjɨ1 tjij1… …ku1 ). However, as translated texts demonstrate, Tangut morpheme does not emerge in environments observed for Tibetan, and vice versa. In some cases, we observe its equivalence with other Tibetan patterns (607)， where this morpheme corresponds with Tibetan INS. Specific case is the occurrence of ku1 as the equivalent of Tibetan ni, i.e. where tja1 would normally be expected. (605) , : “If it directly manisfests in the inner mind, it is cognition” (GAMPOPA, Tib: nang du sems la shar na rtogs pa) (606) : “If one realizes truth of emptiness, the faith into causes and effects rises.” (GAMPOPA) (607) : “Wisdom emerges if there is reliance on the methods” (VAJRA, Tib: shes rab thabs la brten pas grub; Tib INS) (608) , : “If [one] will walk along the road, as the limit os space, [it] cannot be reached.” (VAJRA, Tib: ci ste lam la bgrod ’gyur na nam mkha’i mtha’ bzhin thob pa med)  66 (609) : “Because it does not move, it great true wisdom.” (VAJRA, Tib: mi bskyod pas ni ye shes te) (610) , : ““If one realizes that the discriminations are the dharma nature, the dharma realm should not be contemplated as something different.” (VAJRA, Tib: rtog pa de nyid chos nyid shes na chos kyi dbyings ni gzhan du bsgom mi dgos) (611) , : “If there are wisdom, zeal and supreme wisdom, then whatever one desires will be fulfilled.” (FIVE) j. Comparative clauses Tangut comparative is   ab u  ented by morphem su1 , which can generally be equated with the Tibetan comparative las, bas; indelinable tśhjɨ1 su1 directly representing Tibetan de bas. Similarity is expressed 2 either lexically, or with the help of adverb sju . This morpheme is suffixed to nouns and pronouns, and does not emerge as syntactic marker. The nature of texts at our disposal is such that comparative sentence emerge seldom, and in some contexts present difficulties for interpretation. Below we list several cases where we can be sure of our understanding. (612) : “In the world there is nothing more treasured than (COMP) children” (GAMPOPA, Tib: ’jig rten   ab u las gces pa med pa ste) (613) : “It is greater than (COMP) the five elements” (VAJRA, Tib: ’byung lnga bas gtso chen yin) (614) , , : “Vairocana came to Samye temple before (CONJ) them, and met the king Khri srong.” (FIVE, lit: , “before as compared to them”) (615) : “Before Heaven and Earth” (HUIZHONG) (616) : “The light of non-sign exceeds (lit: “exceeds as compared to”) the Sun and the Moon, it is bright and clear.” (ULTIMATE) k. Temporal clauses As mentioned above, in Tangut there are three morphemes equally meaning “time/ when”, “while”: zjịj1 ，dzjɨj1 , zjọ2 . Out of these, zjịj1 emerges in conditional clauses, sometimes accompanied with tjij1 , that is, its dusrtibution is partially similar with conditional conjunction . This is confirmed with the examples sentences where is equated with Tibetan conditional na. However, other examples are indicative of the relationship between zjịj1 and Tibetan ablative nas. Tangut construction djij2 sju2 wjɨ2 zjịj1 , implying verb usage of zjịj1 · remains unexplained for the moment. In most cases, however, we can observe that “temporal meaning” for zjịj1 is clearly contextual, and its classification as temporal connective is based on the arbitrary usage in the Timely Pearl, where it emerges as both shi 時 and ze 則. Calssfication as a “conditional” conjunction might have been more appropriate, given the established Tibetan equivalents. However, we locate it among the temporal conjunctions for the sake of expedience. Morpheme dzjɨj1 is a lexical word “time”, whereas zjịj1 and zjọ2 only emerge as grammaticalized morphemes, and apparently do not represent the word “time”. Examples with zjịj1 Despite this conjunction has temporal connotation, we suggest translating it as the marker of the conditional clauses. (617) , : “When [Vairocana] reached Indian border, he gave the gold to the guards”. (FIVE, Temporal) (618) , , : “If/ when an investigator (lit. “the one who asks quiestions”) comes, do not tell (2nd SG AGR) him that I came here.” (FIVE, Temporal/ conditional) (619) , , : “In the earlier times (CONJ zjọ2 ), when / if (CONJ zjịj1 ) I met a Dharma master or a scholar, I was afraid.” (GAMPOPA, Tib: sngon chos mkhan dge shes kun dang phrad tsa na  67 nyam nga ba sems byed). Here we observed contrastive usage of zjọ2 and zjịj1 : mja1 njɨ̱ 2 zjọ2 : “in the 1 time of old days” which is a translation of Tibetan sngon. Morpheme zjịj translates Tibetan tsa na, which can be either temporal or casual connective. (620) ， ， : “You, supreme disciple, will now obtain all the teachings from (CONJ) [the initial] turning of the Dharma Wheel [until] entering nirvana”. (GALO; Tib: ’khor lo bskor nas mya ngan ’das de phyir shin tu bu khyod kyis thams cad mkhyen nyid thob par mdzod). Here zjịj1 corresponds with Tibetan nas, i.e ablative usage. (621) …: “After the self-pride and others [afflictions] are abandoned…” (ZHANG, Tib: nga rgyal la sogs gtan spangs nas). Here zjịj1 translates ablative/ causative particle nas. Examples with zjọ2 This morpheme has a more pronounced temporal meaning, however it is distribution is partially equivalent with zjịj1 : e.g it also translates Tibetan tsa na (623). Generally, this conjunction modifies Tangut indeclinables, such as mja1 njɨ̱ 1 (Tib sngon, “old days”), pjɨ1 njwo2 (Tib sngon), tśhjɨ1 zjọ2 (then), etc. Apart from this, we so far cannot established a clear discstribution pattern of morphemes zjịj1 and zjọ2 , while zjịj1 appears to be the most frequent in temporal clauses in the texts which we examined. (622) ， “ ” : “When (CONJ) [he] was promoting the teaching of valid cognition, [his] name was “Direction Elephant (phyog glang).” (KU) (623) , : “In previous times (CONJ, lit: “when before”), although I was reciting many mantras, I often was afraid.” (GAMPOPA, Tib: sngon tsa na bzlas pa mang po byas kyang ’dre tsur gnod) (624) , : “Today (CONJ), whatever scholar (Tib: sangs rgyas) I might meet, I am not longer confused.” (GAMPOPA, Tib: da lta ni snags rgyas byon kyang nyam ni nga) (625) ， : “When (CONJ) I was travelling, I saw these [people]. ” (HUIZHONG, Chin: 我遊行 時，有見如此) Examples with dzjɨj1 This has partially equivalent distribution with zjọ2 , e.g. tśhjɨ1dzjɨj1 , identical with tśhjɨ1 zjọ2 (then). Other than that, unlike two other morhemes, dzjɨj1 represents substatitive “time” (e.g: “Dharma Wheel of the final times”; : “Dharma Wheel of the three time periods”; : “Because he wanted to stay [there] for a long time”). Generally, this morpheme is equivalent with Tibetan dus. Generally, we believe dzjɨj1 is not a grammatical morpheme at all, but should be translated in its nominal capacity. (626) : “When (CONJ) it is practiced, this is contemplation based on the common principle.” (GAMPOPA, Tib: nyams su len dus su don spyi’i tsul du bsgom pa) (627) : “Characteristics and aspects depend on time.” (VAJRA, Tib: rnam spyod dus la ltos pa rnams) (628) : “Yesterday in the first hours of night in the sleeping time…” (FIVE) (629) : “Because of [your] compassionate mind, we seeking time to meet [you].” (FIVE) (630) , : “Before long time passed, the master came out of the door”. (FIVE) l. Morpheme djij2 , Tangut “continious”. Morpheme djij2 is one of the most frequent among Tangut auxiliaries, however its interpretation remains ambigious. As far as we are able to establish, it has no lexical meaning, and cannot be directly identified with any particular Tibetan particle. As mentioned above, it participates in the Tangut concessives, but can also be interpreted as a nominalization marker, and as the morpheme indicative of a “prolonged action”, i.e. modifying the subordinate clause as to demonstrate that the action in the subordinate clause is proceeding simultaneously with the action in the main clause. Thus the marker is suffixed to the main verb in the subordinate clause. In all probability, this morpheme combines all the above functions, and can be seen as both morphological and syntactic marker.  68 Below are the examples which we deem appropriate for three aspects: continuous, concessive, nominalzing and semifnal. Considering the multifunctional nature of the Tibetan semifinal, we suggest that the functions of djij2 are dependent on its origin as the Tangut analog of the Tibetan semifinal particle. However, this requires more research than provided here. Continious 23 (631) ， : “ , ”: “Queen Margyen realized [this and] said: The Kings is not staying alone, he is talking (CONT, lit: “is doing talking”) with someone.” (FIVE) Here djij2 indicates that the kings’ talking to someone is simultaneous with with his not staying alone. (632) , , , , .: “Because the son of the junior queen was able to recite [the mantras] in whichever manner the master was reciting (CONT), then, while playing with [other] children, he was explaining (CONT) [mantras] to the children”. (FIVE) (632) , : “While (CONT) what was said by the masters had some slightly different phrases, the meaning was identical without contradictions.” (FIVE). That is “having no contraditions” happened simultaneously with “meaning being identical”. Concessive (633) ， ， ， : “That is, althoug he was staying (CONS) in the world for 517 years, his appearance was that of a young adolescent, and that is why he was called “Kumara”. (KU) (634) ， ， ， ， : “While (CONS) these masters had immeasurable merits, it to discuss these briefly, the masters were the “family of the Great Vehicle”, with soft original nature, without self-arrogance, evel and wrath.” (2885) Nominalization (635) :“ ” ， : “The masters who follow Āgamas in their practice, (i.e. bka’ gdam pas) say: “If one realizes the causes and effects, the faith will rise”. It is truly so”. (GAMPOPA, Tib: bka’ gdams pa kun gyi zhal nas stong pa nyid rtogs na las ’bras la ’dzem du ’gro gsung ba la de med par ’dug. Tangut and Tibetan not fully congruent.) Here djij2 nominalizes the preceeding clause, which become the topic of the sentence marked with tja1 . (636) ， ， : “Concerning the great saint Master Galo, sometimes he did according to this, but sometimes he did not follow this.” (GALO, Tib: dpal chen po rgwa lo ni, res ’ga’ ’di mdzad res ’ga’ mi mdzad par gda’o). Here djij2 corresponds with gda’o, “to be”. However, heredjij2 can be interpreted as semifinal/ final. Seminfinal (637) : “Emptiness and (SF) what is not non-empty, in the realm of emptiness.” (VAJRA, Tib: stong ste stong pa ma yin stong pa’i yul las ’das. Tangut and Tibetan not fully congruent) (638=38) , : “After the king Khri srong died his son Ral pa can…” (FIVE) Here the position of djij2 is identical to the semifinal (connective or temporal). (639) , : “Because there is such an outstanding teaching, Indians do not allow its propagation.” (FIVE) (640) , : “After [royal heir] paid homage to the masters footprints, the master met the royal heir and asked:” (FIVE) 23 tshjɨj2 sjɨ2 po1 zjɨ1̣ dzjɨ1̣ ，tshjɨj2sjɨ2po1zjɨ1̣ is a place name tshe spong Tangut zjɨ1̣ “氏”represents Tibetan བཟའ་，dzjɨ1̣ means “red”， i.e. Tib dmar.Tangut bu̱ ŋwər dzow 2 2 1 Empress. Thus: Tshe spong bza’ Dmar Rgyen, i.e. Queen Margyen.  69 SELECTED TEXTS 1. Texts with full analysis Attack on the town of Yuan 已过 在 “ 愿意，应 ” “ 【 】 ， ” 原  70 tsjĩ1 ·wẽ1 kow1 njij2 ɣjwã1 we2 tśhja̱1 lju2 śji2 (śjɨ1 ) ，sọ1·ja̱r2 tśhja̱1 nja1 djɨj2 ·wji1。 sọ1·ja̱r2·wjɨ2 dzjịj1，ɣjwã1 we2 mji1 ɣiow1, ·wẽ1 kow1 dzju1 bji2 “gja1·wjij2 lhjwo1·jɨ2”, gja1 ɣu2 kạ1pjwɨ̱r1 dạ2 “njij2 śji1 ɣjwã1 we2 khwej2 ·jij1 dźjị2 ·o2 nio̱ w1, lju2 dźjịj1 lja1 na2, wa2 nio̱w1 lhjwo1 ni?” ·wẽ1 kow1 njij2 dạ2 “ɣjwã1 we2 ljɨ̣1 rjir1 ni, djij2 dźiej2 lhjo1 mji1 dzju1 (dzu1 ) ŋa2 ·jɨ2” ɣjwã1 we2 dzjwo2 mji1 (mjo1 ) nio̱w1 ·jij1 ni kjɨ1 ɣiow1 tśhjɨ1 zjịj1 rjur1 njij2 tsəj1, to2 zji2 bju1 tjị1 thja1 ɣjwã1 we2 tśhja̱1 mjij2 lju2, dźiej2 zow2 nio̱ w1 ljɨ1 Wen gong of Jin City of Yuan to attack three days establish, determine, followed by a verb modifier pass, exceed NEG + Verb root, not to surrender to command army to return army supervisor, general utter, convince in the beginning big crime conflict to have, to be because to go, to do to come why get, attain to lose trust to like people hear then, after one after another surrender then all the zhuhou everyone rely on, adhere to; keep maintain; “empty word”, marks the end of a sentence  71 Perfection of the Self 修身品 ， ， “ ” “ “ Comments Personal and place names: go̱r no 君夫子,i.e. Confucius; tśjiw lhjij, i.e. Zhou guo 周國; ，lit.: 太 祖後稷之影殿. Translates as: “The Temple (lit. Hall of Shadow) of the Great Ancestor Houji”; xew tsji is a phonetic transcription. Grammar patterns in the text: Negatives in the text: mjij Incomplete action 尚未, 未 mjɨ Used with modal verbs 不 nja Not to be 非 mji Not, universal negative 不 tji Prohibitive 勿 24 Following Nie Hongyin 聶鴻音, Xixia wen Dexing ji yanjiu 西夏文德行集研究, (Lanzhou: Gansu wenhua 2002)  72 Other patterns. Commented graphs are in squares in the main text: zjịj Temporal conjunction: “when”; conditional: “if, 時，則 then” ljɨ “empty word”, marks the end of a sentence; “once” 也，一時 tśhjwo Causative connector “because”, “then” 故，便 mja nio̱w Connector: “if so, then” 然而 tja Topical marker 者 mjɨ djij Concessive conjunction 雖然 mjij djij “Never before” 未嘗 ku Connector: “if, then” 則 ·jij Here: Genitive case marker 之 śjij nominalizer, turns preceding verb phrase into a noun tjij Conjunction: “if” 若 ·wji Verbalizer, turns preceding noun into a verb 為 tsjɨ Concessive marker 雖然 … djij… tsjɨ Even if, then (verb prefix+ verb root+ concessive) do Proposition: “towards” 向，處 ɣa Proposition: “in” 于，於 rjir Comitative, ablative marker 與 kha Proposition “within” 中 tjị·wjij Marker of possibility 可以 rjar ŋwu Conjunction “while” kjɨ Verb prefix tśhja̱ Proposition: “on” 上 lew Here: modal verb 應該 su Here: “concerning” 至於 ·wjɨ Verb prefix ŋwu Instrumental marker 以 ·jij Here: indirect case marker ljọ Interrogative 豈 phji Causative marker 使 ·wo Modal verb 應 Word list in the order of appearance: Tangut Phonetic reconstruction English Chinese njwo śjij previous rules 先朝 ŋwər khju under Heaven 天下 tśhja n. virtue 德 dźju v. make clear, manifest 顯明 kiẹj v. wish 欲  73 śji adv. at first 先 lhjịj n. state 國 mji̱ n. rectify 治 nji n. home 家 ljụ n. body 身 djọ v. practice 修 nji̱ j n. heart, mind 心 tśjɨj v. rectify 正 tśhio̱w pr. someone 或 dzjwo n. man 人 ·jɨr v. ask 問 ŋwu̱ v. speak 曰 mji v. hear 聞 ·jɨ v. speak (quotation) 言 dzjwɨ n. King, lord 君,帝 ŋwu v. to be, is 是 rər n. shadow 影 bã n. vessel 盤 ·iọ n. round, opening word 圓,蓋 zjɨ̱ r n. water 水 gju n. vessel 器 dzjɨj n. square 方 ɣjow n. source 源 sej adj. pure 淨 ljụ n. stream 流 niəj adj. muddy 濁 ne̱ w n. good 善 dźjɨ v. action 行 mər n. root 本 śjwo v. use (modal) 用 dzjọ adj. compare 譬 ljụ n. head 頭 gu pp. middle 中 pjụ n. hat 冠 rewr n. feet 足 zjị n. shoes 履 dzjɨj n. time 時 ka v. abandon 離 sju v. similar 如 śja v. clear, manifest 顯 bju n. place 邊 dźji̱ j v. dwell 住 lwu v. conceal 隱  74 wji̱ v. conceal 藏 niow n. evil 惡 mjijr actor suffix 者 nja̱ v. not to be 非 thjɨ pr. this 此 nio̱w conj. cause, because 故，因為 go̱r gji noble person 君子 mjɨ n. other 他 ljij v. see 見 tśjɨ̱ r v. fear 恐 tśhjwɨj v. be aware 戒 ŋewr v. fear 懼 le v. fear 恐 mə n. sky 天 so adj. high 高 nji v. listen 聽 zji adj. most 最 bji adj. low 底 njɨ̱ n. sun 日 khwa adj. far 遠 swew v. shine 照 njij adj. close 近 njạ n. spirits 神 dwu̱ n. secret 密 se̱ w v. discern 測 ·ju n. devil 鬼 ·jij pr. self 自 ·ju adj. permanent 常 ·u pp. inside 內 sjwɨ v. worry 優 sji̱ j v. think 思 djɨr adv. outside 外 kjạ v. fear, distress 恐，患 le v. fear 懼 tjịj n. sole, alone 獨 me v. sleep 寢 gjwɨr v. lie 臥 gjwi n. clothing 衣 zar v. shame, regret 懺 phju n. up 上 mjijr n. spiritual power 靈通 thwu̱ v. penetrate 通 bji n. below 下  75 tjɨj̣ n. ritual 禮 ljɨ̱ adv. heavy 重 ɣiej adj. truth 真 we v. act, become 作為 zji pr. all 皆 njɨ v. reach 普，及 gju̱ n. benevolent 吉祥 ljwu n. joy 慶 thja pr. that 彼 lja v. come 來 kjɨ̱ r n. room 屋 na n. dark 暗 dzew v. deceive 詐騙 dzjɨ v. beware 敬 bio̱ v. observe 觀 śji v. go 往 tha adj. big 大 tjụ n. ancestor 祖 pjụ n. hall 殿 ·o v. enter 入 tśier adj. right 右 djịj n. platform 台階 rewr n. side 岸 ·ju adv. front 前 kiẹ n. gold 金 gjɨ someone 一，某 ·jar v. establish 立 ŋwu n. mouth 口 sọ num. three 三 gu m.w. 把 kjwi n. lock 鎖 tjị v. put 置 ·wər zjir n. back 脊背 la v. inscription, 記，銘 remember ·jwɨr n. text 文 dạ v. speak 言 ŋwu̱ n. speech, language 語 lji̱ j v. harm 壞 rejr adv. many 多 dạ n. thing 事 ɣiẹ n. obstacle 礙 wa int. what 何  76 ŋjạ n. evil 惡 lhu n. increase 增 ljịj v. increase 增，盛 gie adj. difficult 難 ·ji adv. many 多眾 kụ adv. behind 後 do v. read 讀 bjịj lhjo̱r v. look back 顧 dzji̱ j gji n. disciples 弟子 zjɨ̱ r zji n. small people 小人 bju v. follow 隨 dźjịj v. proceed 行 dźjị n. misdoing 罪 dju n. have 有 lhə v. complete 滿足 dźju̱ v. humiliate 辱 djɨj v. stop 止 bjịj v. danger 殆 dźjo zjir adv. tranquility 常平 khwẹj n. arrogance 傲 ·wo v. must 應該 śja·wjạ v. let loose 放縱 kjur n. intention 志 sə lho̱ v. satisfy 滿足 rejr n. joy 樂 dza̱ adj. exceed 過 tshji n. crucial 要 Textual commentaries 1. Cf. 古之欲明明德於天下者，先治其國；欲治其國者，先齊其家；欲齊其家者，先修其身； 欲修其身者，先正其心；【欲正其心者，先誠其意；欲誠其意者，先致其知，致知在格物。物格 而後知至，知至而後意誠，意誠而後心正，】心正而後身修，身修而後家齊，家齊而後國治，國 治而後天下平。（《禮記》，《大學》） Tangut lit.: 前朝天下明德欲時，先國治也；國治欲時，先家治也；家治欲時，先身修也；身 修欲時，先心正為也。 [] 心正而後身修，身修而後家治，家治而後國治，國治而後天下治能也。 2. Cf.: 請問為國？曰聞修身，未嘗聞為國也。君者儀也，民者景也，儀正而景正。君者槃 也，民者水也，槃圓而水圓。君者盂也，盂方而水方。【君射則臣決。楚莊王好細腰，故朝有餓 人。故曰：聞修身，未嘗聞為國也】。君者，民之原也；原清則流清，原濁則流濁 (《荀子》， 《君道》) Tangut lit.: 或人國治問，答曰： “身修者，雖然聞，國治者，聞未曾；帝者，身是；身正則 景正；帝者槃是，槃圓則水圓；帝者器是，器方則水方；帝者源是，源清則流清，源濁則流濁”。  77 3. Cf.: 善者，行之總，不可斯須離, 可離非善也。人之須善，猶首之須冠，足之待履。【首 不加冠，是越類也，行不躡履，是夷民也】。今處顯而修善，在隱而為非，是【漬旦冠履而昏夜 倮跣也。荃蓀孤植，不以巖隱而歇其芳；石泉潛流，不以澗幽而不清；人在暗密，豈以隱翳而廻 操？】是以戒慎目所不覩，恐懼耳所不聞。【居室如見賓，入虛如有人。故蘧瑗不以昏行變節， 顏回不以夜浴改容，勾踐拘於石室，君臣之禮不替，冀缺耕於埛野，夫婦之敬不虧。斯皆慎乎隱 微，枕善而居。不以視之不見而移其心，聽之不聞而變其情也。】謂天蓋高而聽甚卑，謂日蓋遠 而照甚近，謂神蓋幽而察甚明。【《詩》云：「相在爾室，尚不媿於屋漏。無曰不顯，莫予云 覯。」暗昧之事，未有幽而不顯；昏惑之行，無有隱而不彰。修操於明，行悖幽，以人不知。】 若人不知，則鬼神知之，鬼神不知，己心知之，【而云不知，是盜鐘掩耳之智也。孔徒晨起，為 善孜孜；東平居室，以善為樂】。故身恆居善，則內無憂慮，外無畏懼。獨立不慚影，獨寢不愧 衾，上可以接神明，下可以固人倫。德被幽明，慶祥臻矣。（《劉子新論》，《慎獨》） Tangut lit.: 善者，行之本是；人之善用順者，譬頭中冠用，足上履用，一時離不可也。若明 顯邊住時，修善；隱藏邊住時，為惡者，善修人非也。是故君子，他不見處雖然戒慎，他不聞處 雖然恐懼。天令高雖然，聽順最下；日令遠雖然，照順最近；神令密雖然，知順最顯。若人令不 知雖然，鬼神知也；鬼神令不知雖然，自心智也。故自身常善于住，則內憂慮無，外恐懼無。獨 住時，自影處不懺，獨寢時，臥衣處不愧。此者君子屋暗住雖然，詐不為，自獨戒慎順行是。上 時靈神與通可，下時人禮厚，真為。德者，神人普及，慶祥此順來也。 4. Cf.: 孔子觀周，遂入太祖后稷之廟，廟堂右階之前，有金人焉。參緘其口，而銘其背曰： 「古之慎言人也，戒之哉！無多言，多言多敗；無多事，多事多患。【安樂必戒，無所行悔】。 勿謂何傷，其禍將長；勿謂何害，其禍將大；【勿謂不聞，神將伺人。焰焰不滅，炎炎若何；涓 涓不壅，終為江河；綿綿不絕，或成網羅，毫末不札，將尋斧柯。誠能慎之，福之根也。口是何 傷，禍之門也。強梁者不得其死，好勝者必遇其敵。盜憎主人，民怨其上。】君子知天下之不可 上也，故下之；知眾人之不可先也，故後之。【溫恭慎德，使人慕之；執雌持下，人莫踰之；人 皆趨彼，我獨守此；人皆或之，我獨不徙；內藏我智，不示人技；我雖尊高，人弗我害；誰能於 此？江海雖左，長於百川，以其卑也；天道無親，而能下人。戒之哉！」】孔子既讀斯文也，顧 謂弟子曰：「小人識之！【此言實而中，情而信。《詩》云：『戰戰兢兢，如臨深淵，如履薄 冰。』】行身如此，豈以口過患哉！」(《孔子家語》，《觀周》) Tangut lit.: 昔夫子周國觀往，而以太祖后稷之影典已入，堂右階之前，金人一立，口上三把 鎖置，背脊于銘曰：“昔朝慎言人是，戒應也。語勿多，語多則敗多；事勿多，事多則障多。何 害勿言，此害增將；何礙勿言，此礙盛將。君子天下於，上難順知，故下為求；眾人於，先為難 順知，故後為求。”夫子銘已讀，後視以弟子之言：“小子記應。身隨此如行，則口罪豈有？” 5. Cf.：知足不辱，知止不殆，可以長久 (《道德經》44) Tangut lit.: 蓋足知則不辱，止知則不殆，長久可 6. Cf.: 敖不可長，欲不可從，志不可滿，樂不可極 (《禮記》，《曲禮上》) Tangut lit.: 敖者增令不應，欲者放縱不應，志者滿足不應，樂者過令不應  78 Reading Materials Atiśa goes to Tibet 《 》 , , 1 , , , 2 , . “ , , , .” : “ , , ? , ”. , 3 4 , , , , , . 5 6 7 , , , 8 , □ , “ ” , , 9 10 , , , 11 , , “ , ”. : “ , , ? ?” , . , , , : “ , ; ”, 12 . : “ , , ?” : “ , , , 13 . , , , , , , , , ”. . Commentary. This is famous story found in various versions in numerous rnam thar of Atiśa, the story generally revolves around master’s encounters with Maitripa, whom he expels from the monastery (Vikramaśilā, though not mentioned in the Tangut text) and following encounter with Tara, his tutelar deity, who sends Atiśa to Tibet to redeem his bad karma. The source text for the Tangut version, which is different from the popular Tibetan accounts is so far uncertain. Place names: mja1ɣa1tja1 Magadha  79 Personal names: dji1 pja1 kja̱1 rjar1 Dīpaṃkara, i.e. Atiśa Dīpaṃkaraśrijñāna; me̱2 mja1 lit. “Celestial Mother”, i.e. Tara ( ); gju1 mja1 Vajravārāhī ( ); śia1mji1 śramaṇera, from Chinese 彌. ljɨ̱ r1ɣạ2 ŋa1dźjow1tshji1 ŋwu̱1《 》, book title, Upadeśa “Forty Banners of Emptiness”. Note: in the text kha1 and bju1 and tśhja̱1 are used as both phrase connectors between main and subordinate clauses (temporal and spatial meanings), and in nominal meanings. The text discriminates between ·wji1 and wjo1 , i.e. verb stem vowel alternation. 1. dzjị2 here: temple 2. gja1mji2 nwə1 ljɨ1̣ nwə1 nji2 suggested translation: “Despite both you and I know, but…” 3. ɣa1 nio̱w1 khew2 here: “pass way outside the gate” 4. suggested translation of the whole sentence: [she] was picking food out of the grease leftovers after the dishes were washed. 5. tjɨ1njɨ̱ 2 njɨ̱ 2 khju1 “one day at sunset” 2 2 1 6. tsew rjɨr ·wji suggested translation: “made arrangements”; tsew2 might also imply a Buddhist ritual (krama) 7. du1 rjɨj2 translation of the Chinese 樓閣, “tower” 2 1 8. zjur rjur “broom”. 9. Suggested translation of the paragraph: “[while she] was sweeping, she said “I will catch it [garbage] below as it drops”, and observed a place far from the bottom of the ladder [to the tower, where Dīpaṃkara was], and, having made sure that no one was nearby, [she] collected everything just by stretching her arms.” (Apparently, she did not want to climb) 10. Suggested translation: “Dīpaṃkara saw that this woman is special.” 11. nja1 in the capacity of verb agreement marker. 12. Suggested translation: sọ1mə2 thju̱1 three types of discriminations, which Tara conferred on Dīpaṃkara. 13. ɣjɨ1̣ njij2 djị1 “Diamond Hell”  80 Bodhidharma’s Four Practices [] , : “ , . “ ”1, “ ”2, “ ”3, “ ”4 . “ ” ? , 5 , , : “ , , , 6 , . , , , 7 8 , , , .” : “ , ? . , , . “ ” , , , , . , 9 10 , , , , . , . , , , , . “ ” , , , . 11 , , , . , , 12 13 , . , , , ? , . : “ , ”, , . . “ ” , 14 , , , . : “ , ; , .” ,  81 , , , , , , , , . , , , , , . Commentary: tha2 mo2 ，，達 , Bodhidharma. In the text lew2 emerges in its modal capacity. 1. ljwịj1tshja1dźjɨ , Chinese: 報怨行, “retribution for evil” 2 2 2. nio̱w śjwɨ ̣ dźjɨ , Chinese 隨緣行, “adjustment to causes and conditions” 1 2 1 3. kjụ lew mjij dźjɨ , Chinese: 無所求行, “absence of craving” 1 2 4. tsji̱ r śjwi̱ dźjɨ , Chinese: 稱法行, “harmony with the Dharma” 5. Eight characters correspond to: 棄本逐末, 流浪諸趣 6. tsew2 mjij1 , “limitless” 7. ŋwe ze̱ w lhjịj·wə̱1 1 1 ，, lit.: “gladly accept and follow”, Chinese: 甘心忍受 1 8. tśhiow unclear, probably forms a disyllabic word with ljwịj1 。; its position corresponds with the quantifier 都 in the original. 9. bju̱1 bjịj1 , lit.: “respect and elevation”, glory, 榮譽 10. ·wjɨ2 rar2 , lit.: “already gone”, i.e. the past 1 2 11. mur lhej , “to transform the ordinary people”, original: 俗反 1 12. dźjwɨ , marker of “mutuality” 13. gjɨ̱ 1 dju1 , nine abodes of sentient beings. 1 14. śiee here: “essential meaning”, 旨 1 1 15. lju̱ lju̱ “greedy” 慳惜 Introduction to the Sea of Meanings Established by the Saints  82 Commentary 1. ·jwɨr1 “forms”, i.e. all things 2. ·jiw2 suggested translation “became established”, “separated” 3. tsə̣1 ，, Chinese 色, rūpa, outside physical form 4. rjur1 lhe̱ w1 , suggested translation “in the world” 2 1 5. rjir pjụ , poetic numerical, probable meaning “countless billions” 6. i.e. sentient and insentient beings 7. phju1 gji1 tśhja2 du1 , “above are the pure ones with established virtue.” 2 8. pjụ ɣie 2 , suggested translation: “the first to obtain benefit” 9. bji niəj ·wə̱ low1 2 1 1 , suggested translation: “below are the muddy ones with broad piety”; note the contrast between “pure” and “muddy” known from other Tangut texts. 10. dzja1 wju̱1 la2 , “growth and decay are clear” 11. bji2·jiw1 , “previous causes” 12. gjɨ̱ 1tjij1 , Tangut division of people into “nine classes” 九品 13. 13 and 14 are parallel in structure and opposite in meaning. Suggested translation: “when wise language encounters profane wrath, the disorder rises; compassion of Sage (i.e. the Buddha) and establishment of Emperor brings about the correctness of the State rituals.” 14. See above 15. gju̱2 tsji̱ r1 and mur1 dzji̱ j2 : lit.: 救教 and 俗教，, Daoism and Confucianism 16. kjụ1 mji̱ 2 njijr1 dźjịj1 . Literal translation: “control over treasures and travel to the wild beasts”. The first compound probably means “financial control”; the second is treated as a derivative: njijr1 in the Tiansheng Law Code implies foreigners, suggested translation is thus “foreign travel.” 17. Buddhist sūtras, worldly canons (Confucian classics) and “collections on virtuous behavior”, Dexing ji 德行集, genres of Tangut literature. 18. The line of the Tangut royal ritual njij2 dzjo2 , and the ritual songs (odes): lwər2 kjạ2 and dzjo2 gjij2 。. 19. These two lines in all probability discuss Tangut procedure of keju 科舉 examination 20. ·wo2 dzwə1 suggested translation wenzhang 文章  83 Judge Li Li 1 2 3 4 5 , , , 6 . : “ , , , 7 ”. : “ , , , 8 , . , , 10 . ”, . Commentary: Personal names: tsjĩ1·wẽ1kow1 晉文公; lji2 lji1 李離. Hereafter, one should discriminate between bji2 and njij2 (Chinese chen 臣 and wang 王 ) in their nominal and pronominal capacities. ·jɨ2 quotation mark, not to be translated. 1. dźjị2 jir1 mjijr2 khwej2, lit. Chinese: 問罪大者，i.e. dali 大理. Chinese li 理 corresponds to Tangut dźjị2 jir1 (“to ask about crimes”), khwej1 literally translates da 大, meaning “head”, “supervisor”, etc. 2. gjɨ2 “indefinite article”. The word associated with Chinese mou 某, “one of”. 3. rjɨr2·wji1 Verb suffix + verb “to do” (Stem 1); here with the previous noun forms construction “to be in a capacity”, “to act as” 4. phji1, causative verb (Stem 2), “to make become”, with additional function of verb agreement with the subject in 3rd SG. 5. śjij1 “nominalizer”, with the preceding verb stem forms a “gerund”, here: “questioning” (investigation), for the Chinese ting 聽 6. rjɨr2 khju1 tśju1 ljịj1 dzjwo2: lit. “people serving with inferior ranks”; Chinese: xiali 下吏. tśju1 ljịj1 translates yousi 有司 7. ŋa . Verb agreement, subject 1st SG. 2 djo̱1 here probably “to share”. 8. tja1topic marker. 9. mjɨ1 tśhjɨ1 dji2. Combination mjɨ1 tśhjɨ1 implies impossibility of an action. 2 Combined with dji probably acquires “modal” meaning. Here translates Chinese fei suowen 非所聞. Such is rarely attested. 10. ·jij1·jij1. The first character is a conjunction similar to the Chinese ji 即, the second one is in its standard capacity of the reflexive pronoun. Nāgārjuna and Savari  84 Commentary. Personal names: ·we1 phu2 from Chinese 龍樹, Nāgārjuna; ŋər1 la2 , lit.: “mountain grave”, 2 2 Śavaripa; njij dzji̱ j , “Compassionate Master”, 慈師, here Maitripa; pho lo1mẽ1 1 based on Chinese 婆羅門, brahman. Place names: lji mə1 from Chinese 西天; ·jĩ2 thu1 based on Chinese 印度, Sindhu. 1. ŋwə1 swew1 “five learnings”, Pañcavidyā, rig gnas che ba lnga. 1 1 2. Śjạ ·ji , seven groups of Buddhist followers 3. xiwã1 śjɨ2 ljɨ̱ r1njij2 , Brahma, Indra and the four dharmarājas. 1 1 1 8 4. tshjwu zjiw ·jar ŋjow lit.: “eight seas of heavenly creatures”, i.e. 天龍八部 2 2 5. ·ju śja , lit.: 現前, the eighth bodhisattva bhūmi, mngon gyur pa’i sa, abhimukhībhūmi 1 1 6. njwo kạ , recollection of previous lives, jātismara, tshe rabs dran pa, Tangut based on Chinese 宿命 7. mər2 dạ2 , based on Chinese 本事, avadāna, rtogs pa brjod pa’i sde. 8. mə1 , deva. 9. tśhjɨ1̣ zjɨr1 , the fourth bodhisattva bhūmi, 焰慧, arcismatībhūmi, ’od ’phro ba’i sa. The following discussion is based on the idea of the “radiant light” issued by a bodhisattva at this stage. Preface to the Five Sūtras ： ， ， ， 。 ， ， 。 ， ， ， 。 ， ， ？ ， ； ， ， ， ， 。 ， 。 ， ， 。 ，  85 《 》 ： ， ， 。 □ ， ， ， ， ， 。 。《 》 ， 《 》 《 》 《 》 ， ， ， ， 。 Commentary: Texts: tha2 tụ1 lhjịj·wejr2 lwər2 lhejr2 , Chinese: 佛 說 守 護 大 千 國 土 經 , i.e. 1 Mahāsāhasrapramardanī; wor le lwər lhejr2 2 2 , Chinese: 佛 說 大 孔 雀 明 王 經 , i.e. 2 1 1 1 Mahāmayūrividyārājñī; tha gji śju̱ bo lwər lhejr 2 2 , Chinese: 大寒林聖難拏陀羅尼經, 1 1 1 1 i.e. Mahāsītavanī sūtra; kjụ bju zji rjir lwər lhejr2 2 ，, Chinese: 大隨求陀羅尼經, i.e. Mahāpratisarāvidyarājñi; dwu̱2 ŋwə̱1ɣiwej1·ji̱ j1 , Mahāmantrayanudhārāṇi, no Chinese translation. Personal names: xã2 njij2 , lit.: The King of Han, 漢王, from the context it implies Han Mingdi; mo2 thẽ2 , i.e. 摩騰, Kāśyapa Mātanga; ·a śja2 śjɨ2 , i.e. 阿闍世, Ajātaśatru; swa1 war1 tji2 , Swāti, Chinese: 莎底; bə1 rar1mẽ1 , Brahman 1 Place names: tśjiw lhjịj , i.e. the State of Zhou, 周國; dźjwow1 njijr1 tśiow1ŋər1 , “The mountain where the vultures (lit. “wild birds”) gather”, i.e. Gṛdhrakūṭa, Tangut based on the Tibetan bya rgod phung po’i ri; śjɨ1 tow2 Chinese: 逝多, Jeta forest; ma2 gja2 thjij1 , i.e. Magadha; śjɨ1 lo1·wa1 , i.e. Śrāvastī, Chinese: 舍衛城 1. mər dwewr2 2 “original enlightenment”, 本覺” 2. gjɨ mji̱ j rjir thwu̱1 2 1 2 “as seen in the dream”, i.e. in the dream of Han Mingdi. 3. pej1 tow1 , 貝多字, texts on the palm leaves 4. phi1 nej2 ya1 , i.e. vinaya 5. 1 śjwo ɣu 1 , Tib.: nam phyed, midnight  86 6. ɣiẹ2 mji̱ 1 lit.: 施惡, “evil doers” gnod sbyin, yakṣas 夜叉. 7. lji2 dzji1 lit.: 食香, “smell eaters”, dri za, gandharvas, 緊捺囉 8. nji ljɨj2 , lit.: 囊瓶, “sacks and bottles”, i.e. grul bum, kumbhāṇḍā, 供畔拏 9. 1 gjɨ̱ ɣju 1 , lit.: “nine channels”, unknown term, probably epithet of the Tangut state 1 10. tshji khjiw 2 , name of the author Vairocana searches for the Teaching 1 2 , , , 3 , , , , , “ ”, , : “ , , , . , . , ”. : 4 “ , , ” , 5 . , “ , : “ ” . , , , , .” : “ 6 7 , .” , , , 8 9 . : “ .” , . , , .   , 10 . , , , , . : “ , 11 , , , . , , , , , ”. Commentary. Place names: phə1 lhjịj 蕃國, Bod yul, Tibet; phə1 nji2 gu2 lhjịj lit.: 蕃人中國, i.e.西番 中 國 (Tibet) ， etymologically: Madhyadeśa of the Tibetan people. sja mo ·ja2 1 2 , i.e. Bsam yas. ·a·jar1pja1 lo1 lhjo̱r1 , i.e. Arya Pa-lo’i gling; xa sjɨ1 po1 ŋər1 仟脖亡珊 i.e. Has po ri, mountain near Samye.  87 Personal names: kie1 twụ1 , lit.: “faithful to the precepts” 戒忠 (both words are Chinese borrowings): according to the context Tibetan king Khri srong lde’u btsan; dzja̱2 kə1rjɨr2 tja1 , according to the context, this obviously is translator Vairocana. Sanskritized name, probably reads as *Cakrita; tsar1 ljij2 ku1 rjur2 , obviously Legs drug. Tangut tsar1 ljij2 is a place name Tsari, in this case it reproduces Leg drugs’ father’s name Rtsang. ku1 rjur2 can be rendered as Sanskrit “guru” (i.e. “guru from Tsari”), or as a Tangut transcription of the Tibetan “grub” instead of “drug” (*Tsari grub). ·a tji2·jo2 ɣa2 i.e. “ati yoga”. 1. dziej2 de̱ j1 njij2 輪轉王, cakravartin 2. gjɨ2 throughout the text is used in the position of the Tibetan “gcig”, in many cases not to be translated. 3. ljɨ1̣ , phrase connector, Tibetan “dang” 4. gjịj1ɣie2·jiw1 : translation of the Tibetan “gyi don” (for the sake of, in order to) 1 1 5. zjɨr dźja : Vajra of Wisdom, Tangut term for Buddhist sages. 2 2 6. phju ɣu here: first, superior. 2 7. dụ unit of measure: dou 斗 8. kiẹ2 mẹ2 , i.e. 金粉, “golden sand” 1 1 1 9. pja tji rjar “patra”, “golden plate” 2 2 2 1 10. phia̱ bọ phia̱ phjɨ , meaning unclear. Probably meaning that part of wine for Legs drug was consumed, and part thrown away. 11. This saying implies that it would be difficult for them to obtain Dharma. Fada asks for the teaching 《 》 ， “ ： “《 》 ” “ “ ， ” “ ” “ ” “ ： “ ， ”  88 ： “ ” “ ” “ ” “ ” “ ” “ “ ” ” “ ” “ “ ” ” 《 》 “ ” 《 》 “ ” “ 《 》 “ ” “ ” Commentary. Personal names: tsji̱ r1njɨ2 i.e.: 法達; place names: xjow2 śjɨj1 i.e. 豐城; xu1̃ ̣ tśjiw1 i.e. 洪州; text titles: wjạ1 sej1 lwər2 lhejr2 lit.: 清華經, in fact: 蓮華經 （法華經） 1. tśhjɨ1 su1 “is it not better”, original: 何如不  89 2. rjɨr2 adverbial 所 (possibly verb prefix) 3. this Tangut sentence is different from the available Chinese versions, thus should be translated directly. 4. sji2 verb auxiliary. 5. probably mistaken graph, read 慢 6. incomplete sentence 7. nominalizer 8. nwə1 ljij2 知見 “views”, “understanding” 1 9. dźjɨ·wji Tangut ergative, not translated. 10. phie nur tsjij2·o2 2 1 “開示悟入” (explanation, demonstration, understanding, entry), fundamental concepts of the Lotus sūtra. 11. The meaning is not exactly clear. Original probably 甘受驅馳 12. phji1thjɨ1tạ1phji1 “with an intention to make [the sentient beings] refute”; “in order to…” 13. zju 1 “difference” 14. njij1 here: “needed” 15. thjij2 ŋewr2 here: “inasmuch” 盡 16. corresponds to the Chinese: 饒伊盡思共推,轉加懸遠 “the more you think of it and discuss it, the further away it gets.” Painting an extra leg to the snake , , . , 1 . . , . 2 . : “ .” . “ , , , ?3” : “ ”. : , . . 4 . . :“ 5 6 ,  . , ,   .” . , “ ”.   , “ ”. . , . . “ , . .” , , , . , , 7 , , , ,  90 ?8 , ? .” : “ ”, . Commentary. Place names: tśhjụ1 lhjịj1 楚國; ·wji2 lhjịj1 魏國; tshjĩ1 秦. Personal names: tśja̱1·jow1 i.e. 昭陽; tśhjĩ1 tśjĩ1 i.e. 陳軫; titles: śjow1 tśhjụ1 kjwɨ1 ljij2· jwĩ1 , i. e. 上柱國令尹 1. sjwɨ1 zjị1 here: 患 1 2 2. phjo ŋa verb agreement 1st SG “I will make them return” 3. tsji̱ r2·a bjij2 here translates: “何以貴之”. “How will your position be elevated?” 2 2 4. kjɨ̱ r dzjwo here: 舍人 2 5. sji here “causative”: “it so happens that, …” 6. gji̱ here “遍”, “for everyone”, original: “不足以遍” 7. tśhjɨ su12 , here: “moreover that” (standard translation: “would it not be better?”) 2 2 8. thjij lew here: “then what is it”? Recommending the Wise 1 , .“ ”. :“ , ?” :“ 2 3 4 .” :“ ?” :“ , .”5 6 . . : “ ?” : “ ”. : “ ?” : “ , .” . : “ .  , , .” Commentary Personal names: khji1 xjã1 *Qi Xian, in fact: ; xiəj2 xu1 *Xing Hu, in fact: ; ɣu1 khji1 , *Wu Qi, in fact:. 1. tśju1 Chinese: si 司 “office”, further in the text: tśju1zow2 有司 “official” 2. note contrast between njwi1 and dźjo̱w2 and ljɨ̱ 1 3. njɨ2 njɨ̱ 1 probably a mistake for nji2 njɨ2 “both of you” (probably jiajie 假借) 1 2 1 2 4. mji ŋwu tsjɨ sja Rhetoric question with final exclamation: “is it not that you are?” 1 2 1 5. mjɨ djij … ljɨ ̣ … full form of the Tangut concessive: “although…but”; mjɨ1tśhjɨ1 “impossibility”. 6. ·wjɨ2 tjɨ1 “in the same manner”  91 Fu Ziqian , , 1 : “ , , , 2 3 4 .” , . , 5 6 , . : “ , 7 . , , 8 “ ”. , .” 9 . “ . .” Commentary: Place names: tshji1 lhjij 齊國，śja1 xu1 , i.e. 單父，亶父；personal names: xjɨ1 tsə1 tshja1 宓子賤; kjwi1 swẽ1 季孫 1. dźji̱ j1dźji̱ j1 ; lit.: 住住, “all the population” ， “everywhere” 1 2. tji prohibitive; ·wjij2 zjɨ1̣ 敵寇 1 3. tśhja̱ , normally postposition “上”, here conjunction with causative meaning, “since” 4. gja rjijr1 1 lit.: 軍馬， “army”; śio̱w1 sji2 lit: 谷稻, in the original 麥 5. źier1 njɨ̱ 2 “to criticize”, “讓” in the original 6. djij2 ŋwu2 tsjɨ1 , here we treat the combination: verb prefix + copula + concessive marker as a form of concessive phrase: “although it will become thus”. 7. nji2 plural marker; śjow1 in this context probably should be treated as affirmative 可 8. nji2 verb agreement, 1st PL. 9. See 戲水 Preface to the Recorded Sayings of Nanyang Huizhong 1 , ; , . , ; 2 , . , ; , . 3 4 : “ , ; , , , 5 . , , , , , ? , ; . 6 , . , , , , , , , , , . Commentary:  92 1. phji1 here: irrealis, “if it were so” 2. ·jij1dzju2 自在, Īśvara 1 1 3. thja bju here “thus” 1 1 4. śio̱ sã 集散 2 2 5. ·jɨr mjijr 進人，達人 “practitioners endowed with vīrya” 1 6. tsjɨ “although” Preface to the Wuyin qieyun 1 2 3 , . , , 4 5 6 7 8 9 10 , , . , 11 , . , , . , 12 , , . , , , 13 14 15 16 17 18 19 , , , , , , , 20 21 , . , , . , ; . . , , , 《 》22 , , . . Commentary 1. ŋwə1 ɣiẹ2 division of “five sounds” according to the place of articulation: labial, dental, alveolar, velar, glottal 2. ·we2 bju1 Tangut translation of Chinese qieyun 切韻 “cutting and rhyming” 3. Four tones: “even, rising, departing and entering” 4. dji2 mja1 translation of Chinese 字母, header of a consonantal group in the system of 36 initials. 5. gji1 niəj1 translation of Chinese 清濁, i.e. unvoiced and voiced consonants 6. ·ji̱ j1·ju2 probably Chinese 平仄 7. ljɨ̱ 1·jij1 “light and heavy sounds”, e.g. labials and labiodentals. 8. khie̱j1 mja̱1 果 meaning uncertain 9. bji2 bjij1 高低 high and low intonation 10. dji2 ·we2 切字 fanqie transcription characters 11. bju1 mja1 Chinese 韻母，“rhyme letters” arranged into rhyme groups 攝 （ɣjiw1 ） 2 1 12. njij tjɨ ̣j royal ritual 13. kie1dzju1 written law 14. rjijr2 dzjo2 lit.: 儒詩, “poetry of scholars”, i.e. Tangut “Odes”, epic and ritual poetry 15. gji1 niəj1 “pure and muddy”. Tangut word lists, such as “Synonyms” which otherwise follow pattern of the Chinese Erya, are divided into sections of “pure and muddy”. 16. twe̱1 dẹ1 so2 “pair of yin and yang” (characters with or without final consonant) 17. gju̱1 lji̱ j2 lu2 la1 lit.: 吉凶季記, “seasonal records on good and evil”  93 18. gju̱2 tsji̱ r1 lit.: 救法, Daoism 19. dzjwo2 djị2 tsji̱ r1 lit.: 治人法, medicine 20. sej1 lwər2 lit.: 算曆, “divination and calendar” “” 21. kjạ2 tśhji2 lit.: 歌典, collections of Tangut poetry, “songs” 22. ·jwɨr2 ŋjow2 ljɨ1̣ bju Tangut rhyme dictionary Wenhai Baoyun 文海寶韻 Tian Tan as a commander . , . , 1 , , , , . . , . , , . , . , . , . , 2 , , , . “ , , , .” , 3 . , . 4 . , . , . , , , . : “ 5 6 , , .” , 7 , , . , , . 8 9 . “ , , , .” 10 , , . , , . , . , , , “ ” . , “ ” . 11 , , , 12 13 , , , . , 14 , , ,  , . . , . .15 , 16 . . , , , . , . Commentary. Place names: ·jã2 燕；ɣã1 phjij1 安平；kju1 莒；tsji2 be̱ 1 i.e. 即墨; tśhjiw1 趙  94 Personal names: thjɨj1 tã1 田單; mjĩ1njij2 湣王; gio̱1·ji1 i.e. 樂毅; khji2 kja2 i.e. 騎 劫; tśja̱1 njij2 昭王; xjwi2 njij2 惠王; sjo2 njij2 襄王 1. zjịj2 tjạ1 here: “enforced taxes”, in fact: 市掾 1 2. sa lhjwa here: 宣言 2 1 1 3. rjir kjɨ ber “it so happened that” 1 2 4. ɣu twụ “to substitute” 2 2 2 5. ·wjij ·wjij zji here: “all and every one”; tśhiew1 here: 掘 6. bə1 dźjwo2 “to throw away” 扔棄，擲棄, here: 僇 1 7. gjij njij “extremely, very much” 8. ljị1 ljị1 “the captured ones”, in the original: 所得，降者 9. nju1 nji̱ 2 khjwɨ1 “cut off ears and noses”, in the original: 刵劓 1 1 10. śji su “more than before” 11. ·jɨr ŋwu la̱1 ·wiọ1 2 2 here translates “為絳繒衣”. Here Tangut translates as “wrapped in color”. 12. cf. Chinese: 灌脂束葦於尾 13. Probably mistake for dạ1 “cave” 14. tśja1 tśhjwo1 ·wjɨ2 thu2 lit.: “let [the cows] freely on the road”. Cf. original: 縱牛 1 2 1 15. śjwi rjɨr śjwo verb stem 1+Pref/ ADV+verb stem 2. Antipassive: “was expelled” 16. kjwi1 bju1 “as before” Playing in water , . : “ 1 2 3 ” . , , : “ , 4 ?” : “ , 5 ?” : “ , . , , ? .” . Commentary: The Chinese version found in Xin xu 新序 is attributed to Ping gong 平公 of Jin (suggested by Nie Hongyin;). In other respects, Tangut story differs substantially from the Chinese texts. Personal names: tśjiw1 śjow1 , * 周 尚 (Gu Sang 固 桑 in Xinxu; He Xu 盍胥 in Han Shi waizhuan). The story involves “six-winged crane.” 1. The exact meaning of this sentence is not clear. Multiple use of gjɨ2 is redundant. Currently 2 2 we suggest thjij sjo as interrogative structure modified with ·ja̱1 , with “irrealis” meaning (“How it is not possible to share this joy with the people of virtue”). Chinese: 安得賢 士與之樂此也 or 安得賢士與共此樂乎. tśhja2 dzjwo2 : xian shi 賢士 1 1 2. dzjwɨ ̣ ·o lit.: chuanzhu 船主, original chuanren 船人 1 2 3. mjɨ djij . Concessive, similar with “although.”  95 4. zjịj1 marker of enumeration. st 5. 1 P L, verb agreement Talking with an immortal 1 2 : “ ?” : “ ?” : “ , . 3 , , . ” : “ . . , . . , , 4 , . , . , . , , . , ”. : “ ”. Commentary: 1. śji1 dzjwo2 lit.: 仙人, Daoist practitioner 2. ɣiẹ2 祘 lit.: breath, here: qi 氣 3. le2 dzji1tsə̣1 gja̱2， gji1? niəj1 thjɨ1, lwụ1sjij2·jur1 dzja1 . Lit.: 食 露吞藥，吞清絕濁，養育虛情. “eat dew, swallow medicine, consume the pure and abandon the polluted, nourish empty mind.” Apparently, a quotation from a Daoist scripture. 4. na1 kha1zji2 na1, na1·ji1thjo̱1 ɣa1 .Lit.: “玄之又玄，眾妙之門” Ye Gong loves dragons , . : “ ” , , , . , 1 . . , , . 2 , . , .” Commentary. Personal and place names: lu1 lhjij : 魯國；tśhjụ1 lhjij : 楚國； ɣew1kow1njij2 魯哀公; tsə1 tśjow 子張；śja2 kow1 葉公，i.e.葉公子高 1. mə2 mjijr2 dja2 lhjo1 , Chinese original: 失其魂魄  96 2. thjɨ2·wjɨ2 “Now, at this moment” Yue Shifu 1 , , . , “ . ”. : “ ?” : “ 2 , .” , , . , ⃝, , “ ”. : “ 3 , , ? , ?” : “ , , . , , . , .” : 4 , ,” , , 5 . : “ , , . 6 , , .” Commentary Personal names: ŋia̱2 tsə1 晏子；·jwar1 śji1 xu1 越石父; ⃝ indicates missing graph. 1. djij2 ljij2 “progressive marker” + stem verb (main clause), indicative of simultaneity of two actions of two different subjects of main and subordinate clauses. 2. dźjɨ1 ·wji1 agent marker 3rd SG., so called Tangut “ergative”, translates as passive voice. 3. nja2 object agreement, 2nd SG. 4. ·o1 nji̱ j1 lit.: 腹心，i.e 内心 2 5. lew here modal verb. 6. Sentence translates 晏子有功，免人于厄， 1 1 而反詘下之，其去俗亦遠矣. Thus zju sjij as nominalizing construction should be translated as “difference.” Yanzi “was great in what made him different from other people.” Zhicheng askes for instruction 1 . , . 2 3 , : “ , 4 , ?” , : “ ,  97 . , . , 5 6 , . □ , ? , □ 7 , ”. □ □ □, □□□□□□□□□□□ . : “ ， 8 .” , , . : “ ?” : “ , 9 ”. : “ , . , ? : “ , . □, ?” : “ 10 ?” : “ , 11 . , . : “ , , , ”. , 12 , : “ , ? , ”. : “ ”. Commentary: □ represents missing graphs Personal names: tśji2 śjɨj2 , i.e. 志誠; śjĩ1 sew2 i.e. 神秀；zjɨr2 njwi2 i.e. 慧能. Place names: thej1 xow1 i.e. 太和; tow1·jow1 i.e. 當陽; ɣjiw2 mə2 i.e. 玉泉; tshew1 khji2 i.e. 曹溪 1. ·ji1 mji̱ 1 lit.: 眾宮 i.e. saṃghārama 1 1 2. śjwɨ śjwɨ “all the time” 3. ·jij1 object marker 4. kha1 marker of subordinate clause, indicating the circumstances/ conditions of the main clause. 5. njij1? 親近,verb 6. lew2 here: modal verb 7. tsji̱ r1tshji̱ j1 ljwu2 lit.: 說法會, original: 參請 8. lja1śjij1 nominalized verb, “coming” 9. ·ju2 dzuu2 tji1 gjwɨr1 lit.: 常坐不臥 10. lew2 here: nominalized verb, adverbial “what I give” 11. sã1 mej2 transcription: 三昧 12. unclear: tsiow de̱ j2 bju1 gjị2 obviously translates “誓依歸”. Tangut de̱ j2 probably is a mistake for ·ju2 , “permanent”. Youcai Guoshen 1 , . , 2 3 4 ,   , , , .  98 5 6 , , , : “   , . , , ?” : “ ,  , 7 , , , . , 8 9 , . , .” , . , 10 , . . , . : “ , , , 11 . , . , , . Commentary: Place names: we2 lhjij 衛國; Personal names: khji2 śjɨj1 tsə1 季成子, in fact: 郈成子; tsej1 kwo1śjĩ1 *Zai Guoshen in Tangut, in fact: 右宰穀臣; ·u2 xjã1 *Wu Xian, in fact 甯喜 1. djɨ2 verb prefix. 2. kha1 temporal marker/ conjunction: “while” 3. ·wji2·o1 lit.: 主主 “master”, “host” 4. ·jij1 future tense marker 5. thja1śjij1 Indicative pronoun with substantivator, “summarizes” previous content: “thus”, “in this manner” 6. zjọ2 temporal clause: “when” 7. bjij2 temporal clause: “while”, “as soon as” 8. bju1 here causative, “because” (“because [he] was thinking of leaving it with me”) 9. sji2 has multiple meanings, here substantivator, indicating that preceding content is an accomplished fact (“it will so happen that…”). 10. lhjị2 lit.: 下, here: “to settle” 11. tja1… tjị2 we2 … . Complement: “X … is the one who can become” (可), important: stem alternation between ·wji1 and we2