Janhunen J. The Mongolic Languages

Подождите немного. Документ загружается.

preparing food’. Finally, this form can occur as an adverbial modifier to an adjectival noun,

as in ene chizhe uzhe-zhi saighang wo ‘this flower is beautiful to look at’.

While the imperfective converb mainly expresses the coordination between two equal

actions, the modal converb in -n ~ -ng indicates an action that describes the manner in

which another action is performed. For increased emphasis, this form often appears as

reduplicated, as in wila-ng wila-ng yawulawo ‘crying and crying, [someone] left’. In a

basically similar construction, the final converb in -le indicates an action that is under-

taken for the purpose of another action, e.g. bi nie uzhe-le irewo ‘I have come in order

to look a bit’.

Among the different-subject converbs, the conditional converb in -se primarily indicates

a supposition or condition (‘if’), e.g. zhochen ire-se ‘if a guest comes’. However, this form

can also be used temporally (‘when’), as in bi xieri-se ada anami mang uwo ‘when I woke

up, mother and father were both gone’. The conditional converb serves as the formal basis

for the concessive converb in -se-nu, which also has two functions. Primarily, the conces-

sive converb indicates a concessive relation (‘although’, ‘even if’), as in hhende baer olon

bi-se-nu chade ulie ogine ‘even if he has a lot of money, [he] won’t give [it] to you’.

However, it can also have a temporal use (‘after’), as in hhe iyende echi-se-nu, kieli ese

otuzhi wo ‘after he got to the hospital, his stomach did not hurt anymore’.

The basic function of the terminative converb in -tala (with variants in -tula ~ -dula)

is to indicate a temporal sequence. More exactly, the terminative converb indicates that

the action of its clause precedes the action of the next clause (‘until’, ‘before’), as in

khara olu-tala wiliegiewo ‘[he] worked until it became dark’. A second function of this

converb is to indicate a preference (‘rather than’), as in Lanzhu bi-tala Baza saozhi wo

‘rather than being in Lanzhou, it is better to live in Linxia’.

The completive quasiconverb in -de ~ -de-ne (with non-converbial parallels in Bonan

and Mangghuer) indicates an action that will be completed before the following action is

undertaken, e.g. chi ijie-de-ne bi ijieye ‘after you have finished eating, I will eat’.

AUXILIARY VERBS

Santa has a large number of auxiliary verbs, which indicate various shades of aspectual-

ity, directionality, or modality, often arranged in pairs of two semantically opposed

expressions. Etymologically, most of the auxiliary stems are of Common Mongolic

origin, but some are based on elements borrowed from Chinese.

Auxiliaries which indicate directional or aspectual relationships include ire-

‘to come’ (< *ire-) vs. echi- ‘to go’ (< *od.ci-), qi.yi- ‘to begin’ (from Chinese), and

kai.yi- ‘to open, to begin’(from Chinese). Auxiliaries which indicate progress, completion,

or degree of intensity, include bara- ‘to finish’ (< *bara-), dawa- ‘to cross; to complete’

(<*daba-), olu- ‘to become’ (< *ol- & bol-), sao- ‘to sit; to continue’ (< *saxu-), and ala-

‘to kill’ (< *ala-). The two auxiliaries agi- ‘to take’ (irregularly from *ab-) vs. ogi- ‘to

give’ (< *

ög-) indicate that the action has a beneficiary (‘for’). Several auxiliaries express

the presence or absence of ability, including shida- ‘to be able’ (< *cida-) vs. da- ‘to be

unable’ (< *yada-), and mejie- ‘to know’ (< *mede-). Finally, the auxiliary uzhe- ‘to see;

to try’ (< *üje-) indicates intention or attempt.

In the typical auxiliary construction, the auxiliary verb is preceded by the semantic

main verb in a converbial form. Most often, the imperfective converb is used, but the

final converb as well as the formally unmarked verbal stem (zero-marked sequential

converb) also occur, e.g. gaoda-zhi ire- ‘to become good’, wila-le qiyi- ‘to begin crying’,

360 THE MONGOLIC LANGUAGES

ijie agi- ~ ijie-zhi agi- ‘to eat up’. The two auxiliaries shida- vs. da-, which are

concerned with ability, can only be combined with the unmarked verbal stem. Moreover,

these two auxiliaries, unlike all others, can never be used as regular independent verbs.

SYNTAX

Santa retains the Common Mongolic head-final syntactic behaviour. A verb appears in

final position in its clause, and sentence-internal relationships are indicated by suffixes

and postpositions. Generally, a modifier precedes its headword, as in the noun phrase gao

kun ‘good person’, or the verb phrase lalaghang yawu ‘walk slowly!’. However, the two

adverbs no (native) and hen (from Chinese), both meaning ‘very’, always appear post-

positionally, e.g. enedu kuiqien no wo ‘it is very cold today’, hhe saighang hen wo ‘it

is very beautiful’. These postpositional adverbs can be used simultaneously with regular

prepositional adverbs, e.g. yaghani saighang hen ‘extremely beautiful’ (literally: ‘very

beautiful, very’).

Exceptionally, in topic-marked sentences, an object can be moved to the beginning of

the sentence, but it has to be marked by the connective ending, e.g. Ibura-ni qiche

penjiwo ‘Ibura was hit by a car’. A kind of passive, with the semantic object function-

ing as the grammatical subject and occupying the initial position in the sentence, can also

be formed by using the borrowed passivizer ne.yi- ‘to be affected, to receive’ (from

Chinese). The agent (if present) is indicated by the ablative, e.g. gayimi mini eghiwo

‘my brother hit me’ vs. bi gayi-se-ne (idun) neyi-wo ‘I was hit (one blow) by my brother’;

(without agent) ene kewon changbang pipin neyizhi wo ‘this boy is frequently being

rebuked’.

The copular verb wo (o) is used with nominal (including adjectival and participial)

predicates, e.g. bi Santa wo ‘I am a Muslim’, hhe saighang wo ‘it is beautiful’. It can

also be used in existential (both presentative and possessive) constructions, e.g. ula jiere

nie oqin wo ‘there is a girl on the mountain’, made nie kewon wo ‘I have one son’

(literally: ‘there is one son to me’). When the existential sense needs to be emphasized,

the regular verb wai- ‘to be’ is used instead of wo, as in (dur.) made nie kewon waine

‘I do have a son’. In the copular function, the emphatic stem embai- is used, e.g. bi

Santa embai-ne ‘I am a Muslim’.

Santa (like Mangghuer) has also borrowed the Chinese copula shi, which can be used

to accompany the native copula wo. Both copulas stay in their original syntactic posi-

tions, with shi preceding and wo following the nominal predicate, e.g. hhe shi xieshen

wo ‘he is a student’. In such examples, the Chinese copula can be omitted, yielding the

regular Santa type of clause hhe xieshen wo. The Santa copula can also be omitted,

yielding the Chinese type of equational clause hhe shi xieshen. Even so, the Chinese

copula remains in its original syntactic position.

To indicate various types of emotion and attitude of the speaker, Santa uses a selection

of final particles, which usually follow a finite verb. The particles sha, ya, yo, and ang

are used with imperatives to modify the tone of the commandment, e.g. ochi sha ‘drink!’

(demand of quick action, expressing some dissatisfaction), hhetere nie baigha yo

‘please stop [the vehicle] over there!’ (strong request), xiendase ire ang ‘come when you

have time!’ (strong offer or request expecting a positive response). The particle bai

(probably from Chinese) can be used in a series of connected sentences to get the listen-

er’s attention or response. The particle ma is used to invite the listener’s agreement; it

can also function as a coordinating conjunction between clauses, expressing weak

causation, e.g. enense nie ijie ma, andatu no wo ‘try to eat one of these, it is very

SANTA 361

delicious’. The particle dai emphasizes the speaker’s opinion (especially with regard to

second-hand information), e.g. enede baer olon echiwo dai ‘a lot of money has been put

into this’.

Negation is expressed by several Common Mongolic negative particles. In the finite

conjugation, three preverbal negative particles are used: ulie (< *ülü) for the durative and

progressive forms, ese (< *ese) for the terminative, and bu (< *bUU ) for the imperatives.

e.g. bi maghashi ulie echi-ne ‘I won’t go tomorrow’, bi ese ijie-wo ‘I have not eaten

yet’, kieme-de bu kielie ‘do not tell anyone!’. In equational clauses, the particle pushi

(< *bisi) is used, e.g. bi laoshi pushi wo ‘I am not a teacher’. In possessive clauses,

negation is expressed by using the negative existential uwo ~ wiwo (< *ügei), e.g. made

kewon uwo ‘I do not have a son’. There are some indications that the negative existential

may synchronically have been reanalysed as the terminative form of a hypothetical

negative verb u- ~ wi- (term. u-wo ~ wi-wo), though the occurrence of the terminative

in this construction is certainly somewhat unexpected.

LEXICON

Although the grammatical structures and the majority of all lexical items in Santa are

clearly Mongolic in origin, the Santa lexicon has undergone massive influence from

other languages. Most notably, Chinese loanwords occur in nearly every area of Santa

vocabulary. A survey of the 10,145 items in the largest extant collection of Santa words

(Ma & Chen) shows that 61 per cent of all items are Mongolic and 34 per cent Chinese

(including partial borrowings, mostly verbal roots), while the remaining 5 per cent derive

from Arabic, Persian, Turkic, and Tibetan.

Naturally, the proportions of native and borrowed vocabulary vary depending on the

type and size of the corpus studied. Another estimate (Field) has put the proportion of

borrowed vocabulary at as high as 55 per cent for nouns, but only 29 per cent for verbs.

If we consider closed classes, the numbers will be quite different: for example, almost all

Santa pronouns (with the exception of the reflexive pronoun gojia) appear to be of

Common Mongolic origin, while the numerals are overwhelmingly being replaced by

Chinese borrowings.

The scope of the Chinese loanwords is wide, comprising all word classes, and ranging

from premodern (e.g. giegun ‘rooster’) to modern (e.g. dangyen ‘party member’), from

local (e.g. ga ‘small’) to standard (e.g. kuaiji ‘accountant’), and from everyday (e.g.

beizi ‘cup’) to specialized (e.g. jiennao ‘computer’) items. By contrast, most Arabic and

Persian words in Santa are from the domain of religion, e.g. salang+gie- (Arabic) ‘to

greet’, aghili (Arabic) ‘wisdom’, huda (Persian) ‘god’, maiza (Persian) ‘tomb’. In spite

of their relatively small number, many of these words are frequent in everyday usage.

The small number of Turkic loanwords (some of them originally from Persian) are also

commonly used, e.g. ana ‘mother’, baza ‘city (market place)’, tashi ‘stone’, buwa

‘imam’. Tibetan loanwords are mainly limited to a few local toponyms.

REFERENCES AND FURTHER READING

A Shelefu [Ma Guozhong] and A Yibulaheimai [Chen Yuanlong] (2001) Dongxiangyu Hanyu

Cidian (Dunxian Kielien Khidei Kielienni Lugveqi), Lanzhou: Gansu Minzu Chubanshe.

Buhe [Buigae] and Chingeltei [Cavggaldai] (1985) [published 1986] Duivgsiyavg Gala bae

Mungqhul Gala (Dongxiangyu he Mengguyu) [= Muvgqhul Tuirul uv Gala vAyalqhuv u Sudulul

uv Cuburil / Menggu Yuzu Yuyan Fangyan Yanjiu Congshu 007], Guigaquda.

362 THE MONGOLIC LANGUAGES

Buhe [Buigae] et al. (1983) Duivgsiyavg Galav u vUigas (Dongxiangyu Cihui) [= Muvgqhul Tuirul

uv Gala vAyalqhuv u Sudulul uv Cuburil / Menggu Yuzu Yuyan Fangyan Yanjiu Congshu 008],

Guigaquda.

Buhe [Buigae] et al. (1986) [published 1987] Duivgsiyavg Galav u vUigae Galalgae jiv Ma’teriyal

(Dongxiangyu Huayu Cailiao) [= Muvgqhul Tuirul uv Gala vAyalqhuv u Sudulul uv Cuburil /

Menggu Yuzu Yuyan Fangyan Yanjiu Congshu 009], Guigaquda.

Field, Kenneth L. (1997) ‘A Grammatical Over view of Santa Mongolian’, PhD dissertation,

University of California, Santa Barbara.

Hao Sumin (ed.) (1988) Dongxiangyu Lunji, Lanzhou: Gansu Minzu Chubanshe.

Kakudo [Kakudou], Masayoshi (1995) ‘Dunshango no On’in Taikei (Phonemic system of the

Dongxiang Language)’, Oosaka Gaikokugo Daigaku Ronshuu 13: 31–56.

Kim, Seogyeong Stephen [Jin Ruiqing] (1998) ‘Dongxiangyu Cifa Yanjiu’ [A Study of Dongxiang

Morphology], PhD dissertation, The Central University of Nationalities, Beijing.

Liu Zhaoxiong (1981) Dongxiangyu Jianzhi [Zhongguo Shaoshu Minzu Yuyan Jianzhi Congshu],

Beijing: Minzu Chubanshe.

Potanin, G. N. (1893) Tangutsko-tibetskaya okraïna Kitaya i central’naya Mongoliya, vols. 1–2,

S.-Peterburg: Izdanie Imperatorskago Russkago Geograficheskago Obshhestva.

Sechenchogtu [Siqinchaoketu] (1999) Kangjiayu Yanjiu [Zhongguo Xin Faxian Yuyan Yanjiu

Congshu], Shanghai: Yuandong Chubanshe.

Todaeva, B. X. [H.] (1959) ‘Über die Sprache der Tung-hsiang’, Acta Orientalia Hungarica

9: 273–310.

Todaeva, B. X. (1961) Dunsyanskii yazyk, Moskva: Izdatel’stvo Vostochnoi Literatury (AN SSSR:

Institut narodov Aziï).

Todaeva, B. X. (1965) ‘Dunsyan’, in Narody mira: Narody Vostochnoi Aziï, Moskva & Leningrad:

Nauka, pp. 661–2.

Todaeva, B. X. (1997) ‘Dunsyanskii yazyk’, in Mongol’skie yazyki – Tunguso-man’chzhurskie

yazyki – Yaponskii yazyk – Koreiskii yazyk [Yazyki Mira], Moskva: Rossiiskaya Akademiya

Nauk & Izdatel’stvo Indrik, pp. 60–6.

SANTA 363

364

CHAPTER EIGHTEEN

INTRA-MONGOLIC

TAXONOMY

Volker Rybatzki

The mutual relations of the Mongolic languages can be summarized in terms of the

similarities and dissimilarities that exist between them. Since all the Mongolic languages

are, by definition, descendants of the same relatively uniform protolanguage, their

similarities are mainly due to common genetic heritage, while their dissimilarities reflect

the effect of secondary divergent developments. These divergent developments, in turn,

are connected with innovations that have taken place in the individual Mongolic branches

and languages after the breakup of the protolanguage. The internal taxonomy of the

Mongolic language family can most reliably be based on the analysis of a maximally

representative selection of innovations shared by specific groups of the Modern

Mongolic languages.

The parallel presence of any given Post-Proto-Mongolic feature (isogloss) in two or

more Mongolic languages can, in principle, be due to three different factors: (a) genetic,

(b) areal, and (c) typological. A genetic feature may be defined as an innovation that has

been inherited by all the affected languages from an intermediate protolanguage. An areal

feature, by contrast, may originally have characterized only a single language, from

which it has been borrowed into other members of the family. A typological parallel,

finally, is an innovation that has arisen independently in two or more languages owing to

their similar structural disposition. Obviously, an adequate taxonomy of the Mongolic

languages requires that the three factors are properly distinguished from each other.

Ideally, the genetic innovations present in the Mongolic languages should provide a

basis for establishing a system of genetic branches within the Mongolic family. This task

is, however, considerably complicated by the chronological shallowness of Proto-

Mongolic. Another problem is that the extant Middle Mongol corpus is not very infor-

mative for dialectological purposes. It is particularly problematic to trace the possible

dialectal origin of the Mongolic languages spoken in the peripheries of the Middle

Mongol area, such as Afghanistan and the Gansu-Qinghai region. In the absence of

concrete data we can only presume that Moghol and the languages of the Gansu-Qinghai

complex represent two primary genetic branches deriving directly from peripheral

Middle Mongol dialects.

In this situation, it is often impossible to make an unambiguous distinction between

genetic and areal features. Geographically, most of the Mongolic languages are parts of

a belt that stretches from Manchuria in the east (Dagur, Khamnigan Mongol, Eastern

Buryat), through Buryatia, Mongolia and Eastern Turkestan (Western Buryat, Mongol

proper, Ordos, Oirat), to the Caspian region in the west (Kalmuck). Even the areally

peripheral languages of the Gansu-Qinghai complex (Shira Yughur, Mongghul,

Mangghuer, Bonan, Santa) are still loosely linked with the main belt (through Shira

Yughur and Qinghai Oirat). Only Moghol in Afghanistan is entirely isolated from all

other Mongolic languages.

INTRA-MONGOLIC TAXONOMY 365

From the geographical distribution of the Mongolic languages it is easy to see what

the natural sphere of Intra-Mongolic contacts has been for each actual modern language.

On this basis it may be assumed that areal features unite (1) Dagur with Khamnigan

Mongol and Mongol proper; (2) Khamnigan Mongol with Dagur, Buryat, and Mongol

proper; (3) Buryat with Khamnigan Mongol, Oirat, and Mongol proper; (4) Mongol

proper with Dagur, Khamnigan Mongol, Buryat, Oirat, Ordos, and Shira Yughur; (5) Ordos

with Mongol proper, Oirat, and Shira Yughur; (6) Oirat with Buryat, Mongol proper,

Ordos, Shira Yughur, and possibly historically Moghol; and (7) Shira Yughur with Oirat,

Ordos, Mongol proper, and the (other) languages of the Gansu-Qinghai complex.

The two Mongolic languages with the widest profiles of family-internal areal

connections are Mongol proper (including Khalkha) and Oirat (including Kalmuck). In

the case of Oirat, the number of contact partners is further increased by the possibility of

separate interaction with each member of the Gansu-Qinghai complex, including, apart

from Shira Yughur, also Mongghul, Mangghuer, Bonan, and Santa. The areal relations

within the Gansu-Qinghai complex, are organized in a local chain, in which (8) Shira

Yughur seems to have contacted mainly with Mongghul; (9) Mongghul with Shira

Yughur and Mangghuer; (10) Mangghuer with Monghul, Bonan, and Santa; (11) Bonan

with Mangghuer and Santa; and (12) Santa with Bonan and Mangghuer. The only lan-

guage without any substantial contacts with any other member of the family is Moghol.

In addition to the family-internal contacts, all Mongolic languages have also been

exposed to external influences. The main external contact partners are (i) Turkic for

Buryat, Mongol proper, Ordos, Oirat, Moghol, Shira Yughur, Mongghul, Mangghuer,

Bonan, and Santa; (ii) Tungusic for Dagur, Khamnigan Mongol, Buryat, and Mongol

proper; (iii) Bodic (Tibetan) for Oirat, Shira Yughur, Mongghul, Mangghuer, Bonan, and

Santa; (iv) Sinitic (Chinese) for Dagur, Khamnigan Mongol, Buryat, Mongol proper,

Ordos, Oirat, Shira Yughur, Mongghul, Mangghuer, Bonan, and Santa; (v) Iranian for

Moghol and possibly Oirat; (vi) Arabic for Moghol, Bonan, and Santa; (vii) Samoyedic

and (viii) Yeniseic for Buryat; and (ix) Russian for Dagur, Khamnigan Mongol, Buryat,

Mongol proper, and Oirat (especially Kalmuck).

Most of the Extra-Mongolic partners, with the notable exception of Russian, came

into contact with Mongolic already in Middle Mongol times or even earlier, the two most

intensive contact relationships being those with Turkic and Tungusic. In these cases, the

contacts have resulted in profound structural interaction between the language families

concerned. Structural influences have also been exchanged between Mongolic, Bodic,

and Sinitic in the Gansu-Qinghai region. In the rest of the cases, the contacts are mainly

confined to the lexical level.

DATA AND SOURCES

The first serious synchronic classification of the Mongolic languages was proposed by

A. D. Rudnev (1908), who made a distinction between four main groups: (1) Eastern

Mongolic (= Mongol and Ordos), subdivided into a southern (Ordos, Chakhar, etc.) and

a northeastern (Khorchin, Kharachin, Khalkha, etc.) subgroup; (2) Northern Mongolic

(= Buryat and Dagur), subdivided into Western Buryat, Eastern Buryat, Bargut, and (as

a separate marginal entity:) Dagur; (3) Western Mongolic (= Oirat and Kalmuck), including

the Torghut, Dörbet, Öelet, Khoshut, and Khoit dialects; and (4) a residual group, com-

prising Moghol, the Gansu-Qinghai complex, and a number of dialects today classified

as Oirat or Kalmuck. Rudnev also recognized the transitional nature of certain dialects,

including Khotogoit, Sartul, Jakhachin, and others.

Three years later Rudnev (1911: 231) reshaped his classification, which now

comprised three groups identified as (1) Northern Mongolic (= Dagur, Buryat, Khorchin,

etc.), (2) Central Mongolic (= Khalkha), and (3) Southern (‘Meridional’) Mongolic

(= Ordos, Chakhar, Kharachin, etc.). In this classification, Oirat, Moghol, and the Gansu-

Qinghai complex appear as loose satellites of Southern Mongolic. The obvious fault of

Rudnev’s two classifications is that they do not sufficiently reflect the actual diversity

extant beyond the core group of Mongolic (Buryat, Oirat, Ordos, and Mongol proper).

On the other hand, Rudnev had relatively detailed data on the dialects of Mongol

proper, for which more up-to-date classifications were only proposed by Louis Ligeti

(1933: 14–20), B. X. Todaeva (1960), and György Kara (1970: 13–26, 268–84).

Rudnev’s terms (Western, Eastern, Northern, Southern, and Central Mongolic), with

varying redefinitions, have been retained by all subsequent Mongolists. The next attempt

at an overall classification of the Mongolic languages was made by B. Ya. Vladimircov

(1929: 5–18), who recognized two main entities: (1) Western Mongolic and (2) Eastern

Mongolic. In his scheme, Western Mongolic comprises (1.1) Oirat (with Kalmuck) and

(1.2) Moghol, while Eastern Mongolic comprises (2.1) Buryat, (2.2) Bargut, (2.3) Dagur,

(2.4) Khalkha, and (2.5) the other dialects of Mongol proper as well as Ordos. Shira

Yughur and the other languages of the Gansu-Qinghai complex (‘Shirongol’) were

regarded by Vladimircov as impossible to classify due to the shortage of material.

Following Rudnev, G. D. Sanzheev (1952: 40) distinguished between (1) Northern

Mongolic (= Buryat), (2) Eastern or Central Mongolic (= Mongol proper and Ordos),

(3) Western Mongolic (= Oirat), and (4) a group of isolated languages (Dagur, Moghol,

and ‘Monguor’). A very similar scheme was presented by Denis Sinor (1952), who addi-

tionally favoured the division of Eastern Mongolic into Central Mongolic (= Khalkha)

and Southern Mongolic (= other dialects of Mongol proper and Ordos). Sinor also divi-

ded Western Mongolic (= Oirat) into three subgroups, comprising (i) European (Volga)

Kalmuck, (ii) Mongolian (Kobdo) Oirat, and (iii) Jungarian (Ili) and Qinghai (Kukunor)

Oirat. Like Sanzheev, Sinor lumped the peripheral Mongolic languages (Dagur, Moghol,

and the Gansu-Qinghai complex) together as a separate ‘archaic’ group.

The issue concerning the classification of the Mongolic languages was also taken up

by Nicholas Poppe (1954: 6–7), who followed Vladimircov and operated with two main

groups: (1) Eastern Mongolic and (2) Western Mongolic, as well as (3) an additional

category of ‘insular’ languages. In this framework, Eastern Mongolic comprises three

subgroups: (1.1) northern (= Buryat, with Tsongol and Sartul), (1.2) central (= Khalkha,

with Khotogoit), and (1.3) southern (= the other dialects of Mongol proper and Ordos).

Western Mongolic is identical with Oirat (with Kalmuck), while the ‘insular’ languages

include Dagur, Moghol, and ‘Monguor’. In a slightly different formulation, Poppe (1955:

14–23) classified Dagur and ‘Monguor’ also as Eastern Mongolic, and Moghol as

Western Mongolic.

Poppe’s classification with two main groups and one additional group of ‘isolates’ was

echoed by Gerhard Doerfer (1964: 41–3). On very similar lines is also the classification

of Sh. Luwsanwandan (1959), who divided the core group of Mongolic into (1) Northern

Mongolic (= Buryat), (2) Central Mongolic (= Khalkha, Chakhar, and Ordos), (3) Eastern

Mongolic (= Khorchin, Kharachin, etc.), and (4) Western Mongolic (= Oirat and

Kalmuck). Luwsanwandan also recognized the existence of several transitional dialects

between Central and Eastern, Central and Western, as well as Central and Northern

Mongolic. The peripheral languages remained outside the scope of his classification.

On the basis of fresh field observations, Todaeva (1959, 1963) made it for the first

time clear that Bonan and Santa are distinct languages, clearly separate from ‘Monguor’.

366 THE MONGOLIC LANGUAGES

This information was gradually absorbed into the taxonomic schemes of other

Mongolists. Poppe (1965: 7–15, 1975: 151), for instance, refined his former classification

by recognizing Santa (but not Bonan) as a separate Mongolic language on par with Dagur,

Buryat, Mongol proper (with Ordos), Oirat, Moghol, and ‘Monguor’ (with Bonan).

A corresponding revision of Doerfer’s classification was made by Michael Weiers (1986:

66–9). Other classifications of later years, with minor variations in the details, include

those by Yu Shichang (1983), Chingeltei (1989), and N. S. Yaxontova (1997).

One of the most ambitious and detailed taxonomic schemes for Mongolic is that by

Marie-Lise Beffa and Roberte Hamayon (1983). Their classification of the core group of

Mongolic is essentially identical with the framework proposed by Luwsanwandan and

comprises four principal entities: (1) Northern Mongolic, with (1.1) Buryat (proper),

(1.2) Selenga Buryat (Tsongol and Sartul), (1.3) New Bargut, and (1.4) Old Bargut;

(2) Central Mongolic, with (2.1) Khalkha, (2.2) Chakhar, (2.3) Ordos, and (2.4) transitional

dialects; (3) Eastern Mongolic, with (3.1) eastern, (3.2) central-eastern, (3.3) far-eastern,

and (3.4) transitional dialects (of Mongol proper); and (4) Western Mongolic, with (4.1)

Oirat (proper) and Kalmuck of Turkestan, (4.2) Oirat of (Outer) Mongolia, and (4.3)

Oirat of Inner Mongolia and Qinghai. Additionally, there are the (5) peripheral or

‘isolated’ languages, comprising (5.1) Dagur, (5.2) Moghol, (5.3) Shira Yughur, (5.4)

‘Monguor’, (5.5) Bonan, and (5.6) Santa.

All the classifications mentioned above are based on a more or less comprehensive

analysis of several taxonomically relevant features, which may reflect either genetic or

areal innovations within the Mongolic family. Attempts have also been made to establish

a taxonomy on the basis of a smaller number of features. This was the approach of

T. A. Bertagaev (1968), who claimed that the Mongolic languages can be divided into two

main groups, termed ‘northern’ or ‘synharmonic’ and ‘southeastern’or ‘non-synharmonic’.

By ‘synharmony’ Bertagaev seems to have implied an active type of vowel harmony.

In practice, most Mongolic languages are ‘synharmonic’, while the ‘non-synharmonic’

group is confined to the peripheral languages (Dagur and the Gansu-Qinghai

complex).

Although Bertagaev’s framework involves serious misunderstandings concerning the

diachronic phonology of the Mongolic languages, it was later accepted by Sanzheev

(1977). A similar classification was also proposed by Robert Binnick (1987), who, how-

ever, has later (1992) returned to a more conventional understanding of the issue.

Jan-Olof Svantesson (2000), on the other hand, has proposed a taxonomy based on the

diachrony and typology of vowel paradigms. In his classification, the Mongolic

languages are divided into three groups: (1) Eastern Mongolic, (2) Western Mongolic,

and (3) Oirat. Under Western Mongolic Svantesson includes Moghol, Santa, Bonan, and

‘Monguor’, while Eastern Mongolic comprises Shira Yughur, Mongol proper, Buryat,

Khamnigan Mongol, and Dagur.

A pioneering attempt to approach the taxonomy of the Mongolic languages with the

method of glottochronology was made by Shirô Hattori (1959). Unfortunately, the value

of this method for Mongolic studies is considerably reduced by the absence of sharp

boundaries between most of the Mongolic languages. Some of the most clear-cut

linguistic boundaries in the family are present within the Gansu-Qinghai complex, but

Hattori did not have access to sufficiently detailed information on this group. However,

his calculations do reveal the ‘isolated’ position of Dagur and ‘Monguor’ (Mongghul).

The proportion of Common Mongolic basic vocabulary was established to be 72–6

per cent in Dagur and 76–80 per cent in ‘Monguor’. By contrast, the proportion of the

basic vocabulary shared by Buryat, Khalkha, and Chakhar was as high as 89–97 per cent.

INTRA-MONGOLIC TAXONOMY 367

The general conclusion from past research on Intra-Mongolic taxonomy is that there

is no full consensus concerning the principles and methods of classification.

Nevertheless, frameworks going back to Rudnev still seem to prevail. The most promising

new area of research today concerns the Gansu-Qinghai complex. According to the

preliminary results of Hans Nugteren (1997) this complex involves an actual genetic

branch, which comprises two sub-branches: (i) Bonan-Santa, and (ii) Mongghul-

Mangghuer (‘Monguor’). The position of Shira Yughur remains unclear, though it seems

to be transitional towards Mongol proper.

In the discussion below, the taxonomically relevant phonological, morphological

(morphosyntactic), and lexical properties of the Mongolic languages and dialects, as

described in this volume, are systematically surveyed in order to assess the linguistic

relationships between the individual idioms. On the basis of this information, a tentative

classificatory scheme will be presented. The principal entities included in the compara-

tive database are: Dagur, Khamnigan Mongol, Buryat, Khalkha, (other) Mongol dialects,

Ordos, Oirat, Kalmuck, Moghol, Shira Yughur, Mongghul, Mangghuer, Bonan, and

Santa. When necessary and possible, information from additional dialects and

sub-dialects, as well as from Middle Mongol and Written Mongol, are also considered.

PHONOLOGICAL FEATURES

In the following, the phonological innovations uniting and separating the Modern

Mongolic languages are listed in a running order (1–32), starting with the features affect-

ing the vowels. The ordering is not supposed to reflect any immediate chronological or

geographical circumstances, though mutually related features are grouped with each other.

(1) Palatal harmony

The basic type of vowel harmony in Mongolic, palatal harmony is well attested in

Middle Mongol and synchronically preserved in all Modern Mongolic languages with

the exception of Moghol and the Gansu-Qinghai complex. Even within the Gansu-

Qinghai complex palatal harmony is present, though in a reduced shape, in Shira Yughur.

To some extent, the rules of palatal harmony have also been simplified in Dagur and

Khamnigan Mongol.

(2) Labial harmony

Although secondary to palatal harmony, labial harmony was emerging already in Proto-

Mongolic and Middle Mongol, and its modern distribution resembles that of palatal

harmony. Labial harmony is, however, synchronically absent not only in Moghol and the

Gansu-Qinghai complex, but also in Spoken Oirat (with Kalmuck) and Western Buryat,

cf. e.g. Oirat zowa- vs. Mongol proper dzowo- ‘to torment’. Since labial harmony is

attested in Written Oirat, its absence in the modern language is likely to be secondary.

Disturbances in labial harmony are also present in Dagur and Khamnigan Mongol.

(3) Rotation

The rotation of the phonetic values of some or all of the vowels *e *o *ö *u *ü is

observed in all Mongolic languages with the exception of Oirat (with Kalmuck) and

368 THE MONGOLIC LANGUAGES

Western Buryat. The phenomenon is weak in Ordos, while it is most strongly present in

Dagur and Khamnigan Mongol. Originally, rotation did not interfere with the number of

vowel distinctions, but the stronger its impact in a given idiom is, the more often it has

conditioned other changes in the vowel system.

(4) Prebreaking

This feature, which involves the apparently sporadic regressive assimilation of the vowel

*i < *ï in the initial syllable, may be regarded as Common Mongolic, but it is almost

absent in Dagur, Khamnigan Mongol, Buryat, Moghol, Bonan, and Santa. The word

*mïka/n ‘meat’ (Written Mongol miqav), for instance, is attested as *mika/n ~ *miqa/n

(without prebreaking) > Khamnigan Mongol mika/n, Dagur myag, Buryat myaxa,

Moghol miqån, Bonan megha, Santa migha vs. *maka/n (with prebreaking) > Mongol

proper max, Ordos maxa, Oirat maxn, Shira Yughur maghan, Mongghul maxa,

Mangghuer maqa. However, the distribution of some items shows deviations from the

general pattern, cf. e.g. *jïrguxa/n ‘six’> Khamnigan Mongol jurgaa/n (with prebreaking)

vs. Mongghul jirghoon (without prebreaking).

(5) Palatal breaking

In difference from prebreaking, palatal breaking is systematically present only in Dagur,

Buryat, and Mongol proper. However, word-initially palatal breaking is also attested in

Ordos and Oirat, cf. e.g. *imaxa/n ‘goat’ (Written Mongol vimaqhav) > *yamaa/n >

Ordos yamaa, Oirat yaman (with palatal breaking) vs. Khamnigan Mongol imaa/n (with-

out palatal breaking). Moghol examples like nudun ‘eye’ (< *nidü/n) and su[n]dun

‘tooth’ (< *sidü/n) are problematic, but they are probably best explained in the context

of prebreaking.

(6) Labial breaking

This is a feature restricted uniquely to Dagur, e.g. *gurba/n > Dagur gwareb ‘three’. It is

true, it has been claimed that a similar phenomenon is dialectally present in Mongol

proper (Kharachin), but this information remains unconfirmed.

(7) Palatal umlaut

This feature unites Oirat (with Kalmuck) with the eastern dialects of Mongol proper

(Khorchin, Kharachin, and Juu Uda). There is a paradigmatic difference, however, for

while Oirat has merged the umlaut vowels ö ü with the original vowels *ö *ü, the umlaut

dialects of Mongol proper retain the distinction owing to the rotation of the vowel

system. The vowel ä

is a new phoneme also in Oirat.

(8) Regressive assimilation of *e-ü > *ö-ü

This development was already becoming completed during the Middle Mongol period,

and it is observed in all Modern Mongolic languages, as in *edür ‘day’ (Written Mongol

vdur) > Proto-Mongolic *ödür. An exception seems to be present in the item *ebesü/n ~

*ebüsü/n ‘grass’ (Written Mongol vbasuv ~ vbusuv) > Common Mongolic *öbüsü/n,

INTRA-MONGOLIC TAXONOMY 369