Abstract
Abstract
The aim of the present paper is to discuss the differentiation between direct and indirect quotations in Tangut texts. Based on the analysis of Buddhist texts and translations from Chinese classics in the Khara-khoto collection preserved in the Institute of Oriental Manuscripts, RUS, the quotations in Tangut may be divided into two categories. Direct speeches are not marked or attached by a quotative marker jɨ2妒. Indirect speeches are composed by using jɨ2妒 or a compound marker jɨ2djij2妒纓 at the end of the quote. Tangut quotation structure is closely related to the verb jɨ2妒. It is illustrated in the study of cross-linguistic typology that the semantic verb ‘say’ may usually evolve to quotative markers. The semantic and grammatical functions of the verb he semantic and grammatical functions of the verb jɨ2妒 which the quotation structure used reflected the grammaticalization from an independent verb to a grammatical marker. This phenomenon can provide a new perspective for the grammatical study of Tibetan-Burman languages.
Introduction
Quotations refer to sentences that directly quote another person’s speech, or reports that are derived from the third person’s saying or writing. Quotations in Tangut texts can be expressed by adding a semantic verb ‘say’ and by using different quotative markers. In this paper, we shall discuss three issues related to Tangut quotations. The first and main issue involves the two different categories of quotation structure; that is, so-called direct quotation and indirect quotation. The second one concerns the function of the semantic verb ‘say’ jɨ2妒 and ·jɨ1端. In quotation sentences, these two words are used both as a notional verb and a quotative marker. No evidence represented in any Tangut texts has provided the clearly functional differences between them. A textual analysis will be demonstrated in this part to explain the function of jɨ妒. The third part is focused on the co-existing phenomenon of two words and also on the grammaticalization of the speech verb ‘say’, and particularly on the change from a semantic verb to a grammatical marker in Tangut quotations.
The quotations in Tangut texts
The materials in this paper are mainly based on the Tangut version Ārya- Susthitamatidevaputra paripṛcchānāma-mahāyānasūtra (Shanzhuyi Tianzi Hui 善住意天子会), a single chapter of Mahāratnakūṭa-sūtra (Dabaoji Jing 大宝积经) for comparative study. We chose a Chinese original version of this sutra, Shanzhuyitianzi Suowenjing 善住意天子所问经, translated by Dharmagupta in the Sui Dynasty. Likewise, there are other materials used in our analysis. For reasons of convenience, the materials are listed here with their abbreviations (see: Ma (2016); Sun (2015); West (2015); Shi et al. (1983)):
DBJJ Mahāratnakūṭa-sūtra 大宝积经
DMDS DamoDashi GuanxinLun 达摩大师观心论
ZYCC Zeyao Changchuan Tongming Zazi Xu 择要常传同名杂字序
WH Wenhai 文海
For each example to be analyzed we provide four pieces of information in the following analyses:
Tangut sentences from sutra or other manuscripts;
phonological reconstruction of the Tangut characters by Hwang-cherng Gong;
Chinese words and grammatical markers used for translating Tangut words; and
English translations of the examples.
Direct quotations
Direct speech is a form of speech presentation whereby a speaker’s utterance is reported verbatim. Direct speech construction in Tangut is usually accompanied by a speech verb which occurs before the quotations, named ‘quotative verbs’. The function of quotative verbs is to indicate that an adjacent utterance is reported. According to the contexts, speech verbs such as tshjiij1铜, dạ2禑 and ·jɨ2妒 are most commonly used as the quotative verbs corresponding to shuo 说, yue 曰 and yan 言 in Tangut–Chinese translation. The compound speech verbs such as dạ2·jɨ2禑妒, dạ2·jɨr1禑瑚, kụ2dạ2框禑 and nẹ1tshjiij1模铜, which correspond to the yanwei 言谓 (to say), wenyue 问曰 (to ask), dayue 答曰 (to answer) and xuanshuo 宣说 (to declare) in Chinese, are also frequently used in the translation.
Extract (1)
Extract (2)
As is well known, most of the Tangut manuscripts are translated from Buddhist sutra; some non-speech verbs, particularly Buddhist terms such as ljɨr2簄 (to consider) and ŋwuu1贡 (to praise), are also often used instead of speech verbs, as in Extracts (3) and (4).
Extract (3)
Extract (4)
Since the quotative marker ·jɨ2妒 in Extracts (3) and (4) is not showing up in the first two examples, we separate the direct speech into two types. The canonical structure of Tangut direct speech is not marked – there is a clear speech verb used as a quotative verb to report the quote. From Extracts (3) and (4), we know that the clause with the Buddhist verb ljɨr2wji1簄属 is marked. These examples are from Tangut sutra Ārya-Susthitamatidevaputra paripṛcchānāma-mahāyānasūtra. The direct speech marker ·jɨ2妒 occurs more than 10 times in this sutra, and the examples show that in terms of the quotative verb, the typical speech verb never shows up in these sentences. The ‘think’ or ‘remember’ verbs such as ljɨr2wji1簄属 and sjwɨ 1ljɨr2粙簄 are used instead of the speech verbs, which are commonly used for translating Chinese Buddhist terms such as siwei 思惟 (to think) and sinian 思念 (to think of). Siwei 思惟 is a Buddhist term that refers to thinking about or considering an object with discrimination. These sentences lack any lucid expression of the quotation. The function of the quotative marker at the end of the quote may be the complement for the absence of quotative verbs.
Indirect quotations
In Tangut, every indirect speech construction is marked with ·jɨ2妒 or jɨ2djij2妒纓, indicating to report the third person’s speech in dialogue. An interesting approach in the function of djij2纓 was described by Sun (2010). This Tangut character is considered to be an adversative conjunction used at the end of the clause with ljɨ1̣菋 in the transitional complex sentences. However, the function of this conjunction in the indirect quotations is not very clear. We can only consider it as part of the compound quotative marker of indirect quotations. When jɨ2djij2妒纓 occurs in conversation, it intends to report indirectly or quote the speaker’s words.
Extract (5)
Extract (6)
Extract (7)
The three cases presented in Extracts (5)–(7) are intercepted from the dialogue. In Extract (5), Buddha is talking about the hell where sentient beings heard the voice. After the quotation, there is a marker ·jɨ2妒 attached to the end of the statement. In Extract (6), the quotation occurs in a canonical clause, which is the dialogue between Buddha and Manjuśrī; Buddha quoted the two questions asked by Manjuśrī in the conversation, ‘what is a Bodhisattva?’ and ‘what are Bodhisattva’s doctrines?’. After ·jɨ2妒, a second single-person marker nja2萯 is used for echoing the ‘you’ in the former content. Extract (7) is a part of the conversation between Buddha and Śāriputra. The Buddha is speaking about the conferral of the prediction of the attainment of Buddhahood in the future by Buddha Dīpaṃkara. In this case, the reported words are enclosed in the quotative marker ·jɨ 2妒.
Extract (8)
Extract (9)
Extract (10)
The above three cases are enclosed in the indirect quotative marker jɨ 2djij2妒纓. The first two cases reported the words of Buddha in the dialogue. The last case comes from the DMDS. Master Huike cited the content of the sutra, which ended with the marker jɨ 2djij2妒纓. The differences between ·jɨ 2妒 and jɨ 2djij2妒纓 are not yet realized in Tangut literature and further effort is still needed. However, all cases of reported speech are characterized by the use of quotative marker ·jɨ 2妒 or jɨ 2djij2妒纓. The reporting clauses attached to these markers indicate the content and the source of the speech presentation.
Tangut speech verb ·jɨ1端 and ·jɨ2妒
Quotations of Tangut show that the speech verb ·jɨ2妒 is versatile. However, there is another word – ·jɨ1端 – which is very similar to ·jɨ2妒. Both words have the same initial and final characters, and only have subtle differences such as the tone and the stroke. It is generally agreed that these two words are both speech verbs and quotative markers. But unfortunately, there is no clear distinction between them, mainly because the usage of these two words is never separated in the published Tangut scriptures. For example, in Ārya- Susthitamatidevaputra paripṛcchānā ma-mahāyānasūtra, ·jɨ2妒 appears more than 220 times and ·jɨ1端 does not appear even once. The same is seen in Leilin 类林, Chanyuan Zhuquan Ji Duxu 禅源诸诠集都序, Saddharmāṛya-smrity-upasthāna-sūtra 妙法圣念处经, Mahāma ṇ ī-vipulavimāna-viśvasuprati ṣṭ hita-guhya-dhāra ṇ ī-sūtra 圣广大宝楼阁善住妙秘密论王总持经, Trisambara-nirdeśe 无边庄严会, etc. Compared with ·jɨ2妒, ·jɨ1端 appears much less. In Tangut Dexingji 德行集, there is no ·jɨ2妒, only jɨ1端 appears – the same can be seen in the Tangut version of the Tantric work Jixiang Bianzhi Kouhe Benxu 吉祥遍至口合本续. Since we have not yet found a single scripture that uses both words at the same time, and the function and distribution are definitely distinctive, there is therefore no way to determine the difference between the two words from the perspective of grammar and semantics.
From the explanations provided by Tangut dictionaries, ·jɨ2妒 and ·jɨ1端 were two different words when the Tangut characters were found. In the Tangut dictionary of homonyms, Tongyin 同音, two words occurred simultaneously as 端妒[夷夷] (42B2). WH lists the two words as synonyms, explained as follows.
Extract (11)
Explanations in WH also give us some additional information. This dictionary was drafted in the mid-12th century. It used a special structure which is similar to the Chinese ‘… zhiweiye 之谓’ – ‘… zhiweiye 之谓也’ – to explain Tangut words. ·jɨ 2妒 and ·jɨ 1端 correspond to the verb wei 谓. Even in the official dictionary, which clearly lists these two characters, the two words are still mixed up when they are used.
Extract (12)
Extract (13)
In the newly discovered Tangut dictionary ZYCC, these two words occurred in the same sentence. However, both are used as the speech verb, and there is no functional difference between them.
Extract (14)
Based on the historical records, Tangut people created more than 6000 Tangut characters, which imitated the form of Chinese characters all at once in the 11th century. The explanation in the dictionary indicates that there are differences between the two words when they are created. Even though the differences between them are not yet known from the literature, it can be confirmed that the two words have many functions, such as the speech verb and the grammatical marker in the Tangut language. Based on the Tangut narrative Buddhist sutra Ārya- Susthitamatidevaputra parip ṛ cchānāma-mahāyānasūtra, we carried out a case study on the distribution and function of the speech verb ·jɨ2妒.
Speech verb and quotative verb
There are two conditions when ·jɨ2妒 was used to refer to the speech behavior. One is known as a speech verb. It occurs not only as an independent verb but, also a morpheme of the compound speech verb dạ2· jɨ2 禑妒. The object can be added to point out what was said. The other is known as a quotative verb, which is always followed with a quote. For example,
Extract (15)
Extract (16)
Extract (17)
The notional words ·jɨ2妒 and jɨ1端, which relate to the speech meaning, are all interpreted by the Chinese verb wei 谓 (to tell) in Tangut translation studies. However, compared with the Chinese version translated by Dharmagupta in the Sui Dynasty, we found that the reality is quite different. Table 1 presents the frequency and translation of Chinese words for ·jɨ2妒.
As it can be seen from Table 1, the most commonly used Chinese translation word for ·jɨ 2妒 is yan 言 (to say), rather than wei 谓 (to tell). Although both wei 谓 and weiwei 谓为 appeared more than 20 times, weiwei 谓为 appeared 23 times in a parallel sentence of a single paragraph in the sutra, and wei 谓 only appeared once. Obviously, it is not appropriate to translate this word to wei 谓 without careful consideration. According to the Tangut–Chinese interpretation, ·jɨ2妒 has two meanings; one is related to the speech meaning, and this is the main meaning of jɨ2妒. It is used to translate yan 言 in Chinese, as well as similar verbs such as shuo 说, yu 语, fa 发, nian 念, etc. Another meaning is ‘called’/‘named’, which is used to translate ming 名, mingwei 名为, wei 谓 and shiwei 是为 in Chinese sutras.
As a quotative verb, ·jɨ2妒 can be used alone and can also be added to the speech verb dạ2 禑 to make a compound quotative verb dạ2 ·jɨ2 禑妒. Unlike the other quotative verbs such as tshjiij1铜 and dạ2 禑, the quotative verb with jɨ2妒 is always attached with an object marker jij1 谍, composing the structure ‘AB 谍妒’ or ‘AB 谍禑妒’, where A is the speaker and B is the listener. ‘AB 谍妒’ translates as the Chinese structure ‘A gao告B’ or ‘A yan 言B’, while ‘AB 谍禑妒’ translates as ‘A gao 告B yan 言’ or ‘A bai 白B yan 言.’ Among the 83 quotative verb examples, only one case did not use the object marker. It can be concluded that in sentences with the quotative verbs ·jɨ2妒 or dạ2· jɨ2禑妒, both listener and speaker have to appear in the dialogue.
Frequency and translation of Chinese words for jɨ2妒.
Considering the frequency with which the two speech verbs occurred, the compound verb is much more common than the single use of ·jɨ2妒. dạ2 禑, itself a very common speech verb and so also functions as a quotative verb.
Extract (18)
As a quotative verb, the listeners need not be present. So, dạ2 禑 is used more widely than jɨ2妒. In the compound verb dạ2 ·jɨ2 禑妒, the main speech meaning was taken as dạ2 禑. When ·jɨ2妒 was added to with a speech verb, it became semantically redundant; the semantic function of the notional verb became weak. Thus, a more grammaticalized function, the quotative marker, emerged from the verb.
Quotative markers
Since some quotative marker examples have already been listed in the previous section ‘Indirect quotations’, no more examples will be repeated, as it is well known that ·jɨ2妒 is definitely a quotative marker. But there is no systematic discussion on the conditions and usages of it. From the text of Buddhist sutra Ārya-Susthitamatidevaputra paripṛcchānāma-mahāyānasūtra, 29 marker cases were collected, including 23 ·jɨ2妒 and 6 jɨ2 djij2 妒纓. In combination with other literature interpretations, we summarized the conclusive characteristics of this quotative marker ·jɨ2妒.
jɨ2妒 can be used as a direct quotative marker and as an indirect quotative marker.
Some regularity between jɨ2妒 and the quotative verb before the quote can be noticed: in direct speech, sentences with the quotative verb tshjiij1 铜, dạ2 禑, ·jɨ2妒 and dạ2 ·jɨ2 禑妒 don’t use markers, because these verbs have definite speech meaning. If some ‘think’ or ‘remember’ meaning verbs such as ljɨr2wji1 簄属 and sjwɨ1ljɨr2 粙簄 are used instead of the regular quotative verbs, the marker ·jɨ2妒 usually occurrs at the end of the quotation.
jɨ2妒 and ·jɨ2djij2妒纓 are both indirect quotative markers. However, the differences between them are not yet clear.
Besides the above cases, there is also another special use of jɨ2妒 that is worth our attention.
Extract (19)
This case is shown in the gathā of the sutra. In the Tangut–Chinese gathā translation, the translator usually follows the original Chinese version; that is, unrhymed words are allowed, whereas the number of words per sentence should be consistent. In a sentence of insufficient words, the translators often use some function words or the notional words with weakened meaning to compensate for the consistency of the whole sentence. In Extract (19), there is no clear semantic word to translate ·jɨ2妒 in Chinese gathā. Therefore, it is put at the end of the sentence to make up the number of the characters in the sentence. This is consistent with the observation we mentioned earlier that the lexical meaning of jɨ2妒 is weaker than the common notional verb.
Conclusion
The distinction between jɨ2妒 and ·jɨ1端 in the Tangut dictionary Wenhai 文海 and Tongyin 同音 may exist because the Tangut people already noticed the functional differences between the speech verb and the grammatical marker. Therefore, it is possible to distinguish the two words with different tones, although it is not strictly enforced in practice.
According to the record of the Chinese history Songshi 宋史, the 485th volume Xiaguo Zhuan 夏国传, Tangut characters were created by a minister named Yeli Renrong 野利仁荣 in the Yuanhao 元昊 period. Yuanhao 元昊 wanted to create Tangut characters and ordered Yeli Renrong 野利仁荣 to sum up the characters. Thus, 12 volumes were formed. These characters followed the Chinese bafenti 八分体 style with their square shape and intricate strokes. 16
It is possible that the creator Yeli Renrong 野利仁荣 is Khitan because of his surname. 17 Since Khitan characters were also created by imitating Chinese characters, it is easy to think that the Tangut characters should have a close relationship with Chinese scripts. Yeli 野利 followed the ancient Chinese lishu 隶书 script when creating the Tangut script.
In addition to the creation of characters, most of the existing Tangut literature are translated or created from Chinese original texts. For example, Tangut people wrote Wenhai 文海, modeled after the Chinese rhyme book Guangyun 广韵, just as Suijin 碎金 was modeled after Qianziwen 千字文, and Tongyi 同义 after Jijiuzhang 急就章. However, although the Tangut people accepted and promoted the culture of the Central Kingdom, due to the differences in language types, some of them do not fit well. The low level of Chinese also limited their understanding of the Chinese classics.
Since the Han Dynasty, Chinese people have used different tones of a character to distinguish the meaning and word class, which is called sishengbieyi 四声别义. This method flourished in the Tang Dynasty and examples can be found in many Chinese classics. Since the only difference between ·jɨ2妒 and ·jɨ1端 is the tone, and this is very similar to the principle of Chinese sishengbieyi, we can speculate that the original intention of the Tangut people was to distinguish the words by using different tones. However, the four-tone Chinese pattern is not applicable to the Tangut language, which has only two tones. Moreover, the glyph of the two characters is too similar. That is the reason why Tangut people did not make a distinction in practice.
Although it is generally believed that Tangut characters were a one-time creation in the 11th century and that Tangut literature represents the language of the 11th and 12th centuries, the function of the verb ·jɨ2妒, from a speech verb and quotative verb to quotative marker, still reflects the diachronic change of grammaticalization through a synchronic phenomenon. It is seen from the study of cross-linguistic typology that the semantic verb ‘say’ may usually evolve to quotative markers. Although the paths and motivation of grammaticalization are various in different languages, the change from the semantic verb to the grammatical marker is the most common characteristic among the languages. In the minority languages in China, for instance, the speech verbs in Amdo Tibetan are grammaticalized to report evidential marker, hearsay evidential marker, topic marker and conditional clauses marker (Shao, 2015). The speech verb in Yi language has evolved into a quotative marker and hearsay evidential marker (Liu and Gu, 2008). Another example of ancient language is the Turkic. The documents of the 7th century show that the speech verb had already been grammaticalized. In addition to the function as a lexical speech verb, complement marker, quotative marker and onomatopoeia, further explanation of some mental activities are also shown in the language (Zhao, 2013). The grammaticalization of the speech verb has attracted considerable attention. Fang (2006) has summarized two grammaticalization paths of the speech verb. The first path is:
speech verb > quotative marker > semi-complementizer >complementizer
We thought that this path would fit the evolution of Tangut. However, only the first step – the evolution from semantic verb to quotative marker – has been finished. When the quotative marker brings in the ‘reported content’, the generalization of the function has not yet been sufficiently analyzed. The function and distribution of ·jɨ2妒 and ·jɨ1端 in quotations in Tangut are worth further exploration. And this research can provide a new perspective for the grammatical study of Tibetan-Burman languages.
Footnotes
Declaration of conflicting interests
The author has declared that there is no conflict of interest.
Funding
This work is supported by the National Social Science Fund of China (grant numbers 17ZDA264, 18BYY202 and 16BMZ026) and the China Postdoctoral Science Foundation (grant number 2017M621012).
Chinese text: 大迦叶言: ‘善哉! 世尊, 愿乐欲闻, 唯垂敷演.’
Chinese text: (佛)告诸魔曰: ‘众仁者, 实为厌患此身仪耶?’
Chinese text: 尊者舍利弗作如是念: ‘…作何事业. 若得睹见, 不亦善乎?’
Chinese text: 文殊师利复作是念: ‘我今应当大集十方诸菩萨众, 令皆得闻如来所说是妙法门, 证深法忍.’
Chinese text: 佛言: ‘大药, 地狱众生, 食无少乐, 惶惧驰走, 遥见镕铜赤汁, 意谓是血, 众奔趣之. 又有声呼:“诸有饥者可速来食.”便走向彼.’
Chinese text:佛告文殊师利: ‘汝问云何为菩萨, 菩萨有何义者, 以能觉了一切法故, 名为菩萨摩诃萨也.’
Chinese text:佛告舍利弗: ‘如是, 如是, 如汝所说. 舍利弗, 昔燃灯世尊授我记言:“摩那婆, 汝于未来过阿僧祇劫,当得成佛, 号释迦牟尼如来应供正遍觉.”’
Chinese text: 文殊师利言: ‘如是, 天子, 若使幻人及幻能有超行转入者, 则吾等亦当如是超越转入. 何以故? 如世尊说:“一切诸法皆如幻化, 故无转入.”’
Chinese text: 尔时, 彼会大众多有无量百千众生, 咸有疑心: ‘今此文殊师利, 所说如是者, 云何得与圣说相应? 所以者何? 世尊恒说:“若人能入三解脱门, 故名为涅盘.”又佛说:“若有修行三十七种助菩提法, 便证涅盘.”’
Chinese text:又问: 如经所说: 三聚净戒者, 誓断一切恶, 誓修一切善, 誓度一切众生.
Chinese text: 迦叶复言: ‘世尊, 我于是中, 乃至不见彼一菩萨. 云何世尊更言十方诸菩萨等?’
Chinese text: 佛告迦叶: ‘如是, 如是, 如汝所说.’
Chinese text: 时, 大迦叶复请彼佛言: ‘世尊, 惟愿大慈.’
Chinese text: 诸侍者言: ‘唯然, 世尊, 愿乐听闻.’
Chinese text: 皆说如幻化, 虚伪不真实.
Chinese text: 元昊自制蕃书, 命野利仁荣演绎之, 成十二卷, 字形体方整类八分, 而画颇重复. Bafenti 八分体 is an ancient style of Chinese calligraphy current in the Han Dynasty. It is a type of Li Shu 隶书.
Devéria (1898) believed that Yeli 野利 is equivalent to Yelyu 耶律.