Mathieu Beaudouin

Over the past few years, the documentation of Gyalrongic languages has shed light on grammatical phenomena which were poorly understood in Tangut, a language of critical importance in the field of Sino-Tibetan comparative linguistics. This paper provides an explanation for the last remaining unelucidated verbal alternation in Tangut (-ɨ¹/-i² ), which, as I will demonstrate, encodes a non-past/past distinction. By doing so, it also gives fresh arguments for placing Tangut and the Horpa languages together within one clade (Beaudouin 2023). Finally, methodologically speaking, it offers an example of grammatical reconstruction from above, i.e., employing sister languages to better understand the grammar of an extinct language through their common ancestor, revealing a rare example of complex distributional retention uncorrelated with phonetic correspondences

The past decade has seen an increase in attention to Tangut, an extinct Sino-Tibetan language spoken by the rulers of the Western Xia xīxià 西夏 empire. The question of its classification especially has become a subject of discussion since the documentation of its closest relatives came to light recently. The present paper builds on the study of the Tangut verb template by Jacques (2011) to question the place of Tangut with regard to the Horpa languages (Beaudouin 2023b). By doing so, it identifies a phenomenon of attraction encompassing synchronic and diachronic analysis and shows that verbal templatic morphology in West Gyalrongic languages is nuclear.

Fieldwork from the past decade has yielded new data from a cluster of languages in Western Sichuan (China), resulting in new observations relevant for the understanding of Tangut grammar. In this paper, I intend to present morphosyntactic evidence pointing to the Tangut language’s membership within the Horpa taxon, located within the larger Gyalrongic group of the Qiangic branch of Sino-Tibetan. Tangut exclusively shares with Horpa languages cognates that are far too peculiar to be the result of mere chance. By successively considering the verbal, nominal, and postpositional domains, the present paper highlights evidence that links Tangut to Horpa, while proposing new paths to the understanding of grammatical categories of Tangut proper, such as orientational/aspectual preverbs.

The Tangut language is of particular importance to the field of Sino-Tibetan Studies, notably because of its morphological conservatism, unexpectedly correlated with a simplification of its syllable structure, a consequence of a process Miyake (2012) called “compression.” Such conservatism is evident in the number of syllable qualities reconstructed that sometimes reflect proto-Tangut’s ancient derivational processes. Verbs also exhibit various flectional phenomena, mainly due to conversion of agreement rules and referential hierarchy rules (Silverstein 1976), in a manner reminiscent of the indexation system of languages of the rGyalrongic taxon within the Qiangic family. The present paper attempts to explain the absence of indexation in the Tangut verb, a key phenomenon in the history of verb agreement analysis. First, I recall the main rules of the Tangut verb’s agreement system, as shown by Kepping (1975) and Gong (2001). Second, cases of non-agreement are analyzed. Apart from the case of non-local contexts, we see that the absence of agreement occurs in non-finite forms resulting from a dependency pattern, such as clause chaining, topic/comment context, and semantically dependent modality.

Le tangoute possède des morphèmes postposés revêtant un sens spatial, qualifiés dans la littérature de "mots à sens spatial", de "postpositions", voire d'"auxiliaires". La présente étude est la première à considérer ces morphèmes d'un point de vue indépendant et systémique. Si la bipartition de Kepping (1985), que nous reprenons en la requalifiant, est valide au regard de certains critères, elle n'est pas la seule au regard d'autres ; par ailleurs, elle n'est pas suffisante pour rendre compte de la variété des catégories, qui s'inscrivent au sein d'un continuum locatif tenant au degré de grammaticalisation de ses membres. En privilégiant successivement les grilles analytiques de la bipartition et du continuum, nous montrons par ailleurs qu'une conjonction harmonieuse de ces deux approches est particulièrement indiquée pour la description d'une langue éteinte telle que le tangoute.

There are in Tangut three locative cases giving the ability to localise the process of a verb inside a constituent: ·u², =kha¹, and gu². These particles are often glossed 中 zhōng in Chinese, without any further development about the distinctions these morphemes operate regarding, for instance, the semantic nature of the noun they refer to. We propose here an analysis of the distribution of two of these three cases, showing that the presence of one or another obey semantic constraints of the noun the clitic is attached to.

Le tangoute, langue médiévale au régime documentaire s’étalant principalement du XIème au XIIIème siècle, revêt en linguistique sino-tibétaine une importance toute particulière, du fait de sa profondeur historique et de son appartenance au taxon gyalronguique, un des plus conservateurs du sino-tibétain. Nombre de travaux, principalement de nature philologique et phonologique ont été menés sur cette langue, d’abord en Union soviétique, puis au Japon, en Chine continentale et à Taiwan. Cependant, les études, bien que systématiques, de Nishida (1964/66) et Kepping (1985) ne répondaient pas complètement à la définition d’une grammaire descriptive, laquelle restait un manque à combler au sein du champ de la documentation des langues sino-tibétaines. La thèse se propose de répondre à ce manque, en fournissant aux typologues et aux comparatistes un outil complet et cohérent, dont l’originalité de l’approche réside dans la mise en regard du tangoute avec les langues gyalronguiques les plus proches jamais documentées, du taxon horpa.

Après une introduction historique, culturelle, et linguistique (classification, typologie, parties de discours) à la langue tangoute, la thèse considère successivement la phonologie (proposant de nouvelles reconstructions), le syntagme nominal, les constituants de nature adverbiale et les verbes.

---- ---- ---- ---- ----

Tangut is a medieval language whose existence is attested across a period from the 11th to the 13th century. Its importance arises not only because it is the third oldest of all the historically widely attested languages within Sino-Tibetan, but also because it belongs to the Gyalrongic subgroup, one of the most conservative taxons of the family. While research has been carried out on this language, it has mainly been conducted with a philological or phonological approach, first in the Soviet Union, then in Japan, continental China, and Taiwan. The work done by Nishida (1964/66) and Kepping (1985), although systematic, does not meet the requirements of a description grammar; the doctoral thesis addresses this gap, thus making a noticeable contribution to the field of Sino-Tibetan linguistics. The research provides typologists and comparative linguists with a comprehensive and coherent tool, the originality of which lies in the use of comparison with the Gyalrongic languages of the closest taxon ever documented (Horpa).

After a historical, cultural, and linguistic (classification, typology, parts of speech) overview of the language, the thesis considers successively the phonology (providing new reconstructions), the nominal phrase, the adverbial clauses, and finally the verb.

It is well known since Kepping (1985) that the distinctions encoded by the two series of directional preverbs in Tangut do not exclusively pertain to orientation but also modality. These series are today commonly understood as encoding indicative and optative mood, respectively. However, new data from fieldworks conducted in the rgyalrongophonic area of Sichuan allows today a refined system, where the unique derivation of the second series from the first series obeys a binary semantic pattern similar to a configuration found in Geshiza Horpa (Honkasalo 2019). Even if only one unique form exists for each element in the two series, series 2 preverbs should be understood as displaying perfective interrogative mood when derived from perfective indicative series 1 preverbs and displaying optative/jussive mood when derived from imperative series 1 preverbs.

Cette table restitue la reconstruction du tangoute telle que proposée dans Beaudouin (2023). Elle se fonde sur les trois grades de la reconstruction de Gong Hwang-cherng (GHC) et prend en compte la proximité des langues horpa. Il s’agit d’un premier jet et le système est appelé à évoluer. Une future version inclura les équivalences avec le chinois, le pré-tangoute de Jacques (2014) et des correspondances avec les langues horpa.