This paper was originally written for a publication in Australia in early 1990. The publication did not appear. The text was originally written on MacWrite and all the character formatting had been lost. Now I have converted the text character by character into Unicode and HTML. I have not changed the content in light of new information but I may do it later although there is not much to change. Obvious mistakes have been corrected. The purpose of the paper was to find where in China Central Asian Dungan originated. However, this may be futile because both Dungan and the Southern Xinjiang dialects were formed only after the Dungan uprising.
In 1861 an uprising of Moslems broke up in Northwest China. It was finally crushed by Zuǒ Zōngtáng 左 宗棠 in 1878. The rebels fled ahead of the Manchu troops all through northwestern China and some of them finally crossed over the border into Russian territory. The rebels included many nationalities, among them Dungans, Chinese speaking Moslems.
The Dungans crossed the border at different times and places. The first groups crossed the border in 1877–1878 and the last groups again in 1882–1888 after the Ili area was given back to China. (Jusupov 1977, Rimsky-Korsakoff Dyer 1979:1). Altogether nearly 10,000 moved over (Sushanlo 1979:66).
Janshansin (1938:83–84) mentions some places of origin for some of the Dungan groups. All the places mentioned are in Xīnjiāng. For instance the group settling down in Frunze came from Ili Valley.
The present Dungan settlements are mainly in the Chuy (Чуй, in Dungan Чүй) River Valley on the border between Kirghizia and Kazakhstan. One, Irdyk (Ирдык, in Dungan Эрдэх), is near Przewalsk (Пржевальск) south of Issyk Kul (Иссык Куль). The settlements elsewhere have probably lost their language.
The Dungans have their own literary language, which was first written in the Latin alphabet since 1932, then in Cyrillic after 1952 (Riedlinger 1989).
There were about 80,000 Dungans in 1989, of which nearly one half live in Kirghizia. The annual population growth is 3.8 per cent. (Shyjüәdi Chi, 7 November, 1989).
1.2 Writing and pronunciation
The most accurate description of Dungan phonetics is Zav'jalova 1979. It is the only source that uses the IPA alphabet. Imazov 1975 describes the sounds in articulatory terms, but not in great detail.
The orthography is phonemic as far as segmental features are concerned and the following treatment uses the Dungan orthography as the basis of the description.
Since Zav'jalova 1979 does not list all the finals some of the following transcriptions are tentative.
1.2.1 The finals
Trubetzkoy has discussed the vowel system of Soviet Dungan. He thinks it is different in different contexts (Trubetzkoy 1939). Because the medial has become a secondary articulation of the initial and the final nasal can be realized as nasalization, syllables are of the CV type. This is customary in Soviet analyses (Imazov 1975).
• ы [ɨ] is retroflex or alveolar depending on the previous consonant. It is the vowel in Chinese si and shi.
• The final nasal [n] varies with the nasalisation of the preceding vowel (Zav'jalova 1979: 57):
[æ̃] ~ [æn]
[ɔ̃] ~ [ɔn]
• и [i] palatalizes the initial, sometimes considerably
• ў labializes, у does not, otherwise the difference is slight (Zav'jalova 1979: 56). у is written instead of ў after labials.
• у [ʊ] is [ou] in Aleksandrovka (Александровка) in Chuy Valley.
• If the у [u] is medial, the initial consonant usually has no labialization. The labialization is occasionally a bilabial trill (Zav'jalova 1979: 44–45).
кў [kɸu], kuэ[kxuɛi]
тў [tɸu], чў [tʂɸu]
Imazov (1975:26, 126) gives the following phonetic values:
• эр is [əɹ] in Irdyk, but [ɯ] in Aleksandrovka (Zav'jalova 1979: 42).
• ки [kxɪi], бый [pɨi] (Zav'jalova 1979: 56). The first part of this glide is regarded as a transition by Nurmekund (1979) and Trubetzkoy (1939).
The final -и is written instead of -ый after labiodentals and velars, which are never palatalized. The final is pronounced [ɪi] in that context.
• In хуэй the final –уй is written with э in order to avoid resemblance with a Russian taboo word.
• Some of the vowels have diphthongal realizations (Zav'jalova 1979: 55):
[ɛ] vs. [uɛi]
[o] vs. [iou]
[ɔ] and [ɛ] are diphthongs in Aleksandrovka.
• With érhuà 儿 化, written –р, the opposition between nuclear vowels is obliterated. Only the medial is preserved. (Imazov 1975:30–31, 39–40)
•ң [ŋ] only occurs before [ɤ], where [n] does not occur and it is consequently an allophone of /n/. In Irdyk it is pronounced [ɣ] (Zav'jalova 1979: 41).
•Aspiration is velar: [px] [tʂx] etc. (Zav'jalova 1979:43).
•в [v], ж [ʐ] and й [ʝ] are strongly fricated. They are originally medials with a zero initial. In Dungan they are probably best regarded as initials.
• р [r] occurs in loan words, also syllable finally.
1.2.3 The tones
It is usually stated that Soviet Dungan has three tones with no difference between the yīnpíng and yángpīng, i.e. between the first and second Pekingese tones (Kalimov 1968, Rimsky-Korsakoff 1967). In isolated syllables this is true, but in connected speech there are four tones just as in Standard Chinese. (Zav'jalova 1973, 1979, Salmi 1980).
There are differences from Pekingese in the distribution of the tones caused by the fact that the old entering tone becomes yīnpíng or yángpíng, depending on the nature of the initial consonant. In table 3 the number 1 means low and 5 high pitch.
|first||11 (21)||13 (131)|
There is one case of tone sandhi. A first tone is changed into a second tone if it is followed by another first tone. In other words, sequences of low tones are not allowed.
The neutral tone occurs after third and fourth tones. After a first tone it is replaced by a second tone (rising) and after a second tone by a third tone (falling).
(1979) describes two dialects. The above tones are those of Irdyk,
which has a simpler system. In Aleksandrovka some of the tones are
slightly different. The fourth tone is 35 and finally 342. In addition,
the first tone has a level final variant 11, which mostly occurs after
an underlying first tone.
Structurally it is most elegant to think that Soviet Dungan tones are low, rising, falling and high. Low changes into rising finally and before another low.
1.3 Dialects within Soviet Dungan
It is traditional to say that there are two main dialects of Soviet Dungan. The literary dialect is based on Gānsù dialect and the other dialect is a Shǎnxī dialect.(Kalimov 1968, Rimsky-Korsakoff 1967, 1979). The best account of dialect differences between the two main dialects of Soviet Dungan are Janshansin 1968 and 1938. The descriptions are not very systematic. They are of little use in the placement of the Shǎnxī dialect in the context of Chinese dialects.
According to Janshansin 1938 the first primer was printed in 1929. The teachers in the first Dungan school were from Frunze and spoke the Gānsù dialect and that is why it became the literary language.
Below are some of the differences between the two dialects.
seems to be widespread in the northwest
|3 tones||4 tones|
There is a long thematic vocabulary list of the two dialects in Janshansin 1938. The differences in vocabulary are often cases like the following, where both variants are formed from Chinese roots.
лыйзы 肋子 ~ лыйба 肋吧 'rib'
эрхуанзы 耳环子 ~ эрҗуйзы 耳坠子 'earring'
The chief difference between the two dialects is the number of tones, the Shǎnxī dialect having preserved the difference of the two píngshēngs. They share the other main characteristics of Soviet Dungan. Both have lost the oppositions -en vs. -eng and shu vs. fu.
1.4 Differences from Pekingese
1.4.1 Phonological features
When we compare the sounds of Soviet Dungan with Standard Chinese, there are considerable differences, but the phonological differences are surprisingly slight, especially if one compares with some dialects further to the east like Zhèngzhōu 郑州 (Zhōu Qìngshēng 1987), where rhotacism is a [u] and where there are finals like iai, where jian-tuan 尖团 are distinguished, and where ten is used for tun.
Some of the differences are such that the Dungan form can be derived from the Pekingese form. There are two main cases of this.
• First, the difference between dental and velar nasal final consonants has been obliterated so that there is no opposition between the following finals:
-in = -ing
[ĩ] ~ [in]
-en = -eng [ə̃] ~ [ən]
-un = -ong [ũ] ~ [un]
The difference between -an and-ang is reflected in the quality of the vowel: –ан [æ̃], –он [ɔ̃].
• Another difference that no longer exists in Dungan is between shu- vs. fu- and ru- vs. wu-. So фу 熟 'ripe' and фу 浮 'swim' are homophonous and so are ван 软 'soft' and ван 晚 'late'. Incidentally, the ф– and в– are never followed by a labial medial. The same constraint holds in Chinese. This means that shui 水 'water' becomes фи (fei), not фуй.
• Zero initial occurs in Soviet Dungan only in the prefix а–. Otherwise it has coalesced with н– as in нанзы 鞍子 'saddle'. Before [ɤ] it is the velar [ң] as in ңә 鹅 'goose', but Soviet Dungan has no [n] in this position, so there is no opposition, although the orthography marks the difference.
Other differences have to be derived from forms of Chinese earlier than present-day Pekingese. There are two main differences of this sort.
• The first of these is the tonal system. There are differences in the distribution of the tones caused by the fact that the old entering tone (i.e. stop final words) becomes yīnpíng or yángpīng, depending on the nature of the initial consonant. In Soviet Dungan the ancient stop final words belong to the present first tone if the initial consonant was earlier a voiceless stop or a sonorant. If it was a voiced stop the word is pronounced in the second tone. So the tones of җў– 竹 'bamboo' with an earlier voiceless stop initial and ва– 袜 'sock' with a sonorant initial have fallen together with yīnpíng, while the tone бый 白 'white' with an original voiced initial has coalesced with yángpíng.
• In some words the distribution of retroflex and dental fricatives and affricates, sh- and s- for instance, differs from Pekingese. Dentals like s- are more widespread in Soviet Dungan. The occurrence of dentals is easily stated by traditional Chinese terminology: they occur in words that belonged to either the second division or else to the zhǐ 止 rhyme class in Middle Chinese. In Dungan it is сан (san) 山 'mountain' and -сы (si) 是 'to be'.
• A third difference occurs in some words originally with a final velar stop, so there are words like ки (kēi) 客 'guest', мый- (mēi) 麦 'wheat', мый (mēi) 墨 'ink', бый (bēi) 百 'hundred', бый (béi) 白 'white'. These are words in the zēng 曾 and gěng 梗 rhyme classes. There are also other cases that have developed in a way that differs from Pekingese such as фә- (фә<shuó) 勺 'spoon' and щүә (xuē) 削 'whittle'.
1.4.2 Differences in grammar
There are a lot of differences in vocabulary, syntax and idiom, but because we have no information on these features on the Chinese side, they are not discussed in detail in this paper. However, some of the more important ones are mentioned.
In morphology the differences are of minor importance. In Soviet Dungan it is common to reduplicate nouns. In Pekingese this only occurs in children's language (Chao 1968:202).
Dungan does not form compound words as freely as Chinese. It commonly uses phrases where Chinese uses words.
Chinese often uses abbreviations that Thompson calls alluding forms (Thompson 1965:137). They are actually abbreviations like Zhōnggòng 'Middle common' for Zhōngguó gòngchǎndǎng 'The Chinese Communist Party'. There are none of these in Dungan and the reason is probably that they are of predominantly literary origin in Chinese.
Soviet Dungan has almost no words from literary Chinese. The ones it has are from folk poetry. There are a few Chinese loans like дянйинзы 'cinema' and гәмин 'revolution'. There are differences in the frequency of many words and some faux amis. A lot of Chinese morphemes do not occur in Dungan and there are also a considerable number of Dungan morphemes that do not occur in Chinese and consequently have no corresponding character.
There are a lot of unassimilated Russian words that occur in economic and political texts, but are not so common in other kinds of texts. Other loan words are religious terms from Arabic or Persian. They are mainly phonologically assimilated, and some of them seem to occur in everyday language.
1.4.4 Differences in syntax
The most obvious syntactical characteristic of Soviet Dungan is the system of marked aspects. Nearly every sentence has an aspect particle and the absence of one is grammatically determined. (Salmi 1984).
A less obvious feature is that Soviet Dungan is far less topic oriented than Chinese. Most topics are marked by a preposition, including objects and patients of passive sentences. In general, the order of constituents is freer. Prepositional phrases can rather freely occur before the subject.
In general, there is more morphological marking of syntactical relations in Soviet Dungan. For instance, sentences as subject are usually marked, and clauses as object can be preposed with a preceding ба 把 and a following –ди 的.
There are also differences in the use of complements. They rarely occur before the object and discontinuous directional complements only occur with –до 到, while simple directional complements are rare.
There are a fair number of other differences.
2 Dialects in China
There are several mutually incomprehensible dialects in China, but some 70 percent of the population speaks the Northern or Mandarin dialects. In this paper, only the northern dialects are dealt with, because Soviet Dungan clearly belongs to this group.
The Northern Chinese dialects are usually called Mandarin dialects. It has been traditional to separate Mandarin from the rest of the Chinese dialects. However, Forrest says there are no real criteria for lumping them all together in one group. The grouping is unfortunate, because it does not express that Cantonese is nearer to Northern Chinese than to the other major dialects. (Forrest 1948:198–199).
Yuán Jiāhuá et al. (1960:23) stress mutual comprehensibility as a criterion for separating the Mandarin group from the rest of the Chinese dialects, but there are also phonological features which can be used to set up the major dialect areas. These are the development of voiced initial consonants into voiceless stops, tonal development, and the development of final consonants (Yuán Jiāhuá et al. 1960:23, Zhān Bóhuì 1981:93). It is especially the present reflexes of earlier voiced initial stops that set Mandarin dialects apart from their southern neighbours. In the southwest final stops and tones are important in this respect.
There have been objections to the use of the name Mandarin. Forrest calls the Mandarin dialects Northern Chinese and so does Zhān Bóhuì. The present classification of dialects is based on work by Lǐ Róng (1985a), who again calls them Mandarin, because he makes a distinction between Mandarin and the Jìn dialects, mainly spoken in Shānxī. He regards them as entirely outside Mandarin, although they are in the North.
No real reason for the separation of Jìn dialects is given in Lǐ Róng (1985a). It is simply stated that the area of Mandarin has diminished in successive descriptions of Chinese dialects and now the Jìn dialects will be separated from them. The distinguishing feature is that they have preserved a final stop consonant ([ʔ]). If this were the only reason for keeping these dialects apart, the same thing should apply for the Jiāng-Huái dialects in the southeast as well, because they too have preserved the old final stop.
There might be some real grounds for the separation of the Jìn dialects, because in the colloquial pronunciation of some Shānxī dialects the initials have developed differently from the Mandarin dialects (Hóu Jīngyī 1986). Moreover, they have often quite aberrant developments of finals.
The name Jìn is the old name for the province of Shānxī and it is used instead of the present name of the province, because the area where they are spoken is not coextensive with Shānxī province. They same principle is followed in Chinese in the names of many other dialects, but it is not possible to follow it systematically in English.
2.2 The basic criterion
There are different ways of classifying Mandarin dialects. The classification of Mandarin dialects in Yuán Jiāhuá et al. (1960) can be regarded as traditional. The book divides them into four groups, Northern, Northwestern, Southwestern, and Jiāng-Huái between the Yángzǐjiāng River and the Huái 淮 River. The same classification is used by Zhān Bóhuì (1981:94–98). This classification is mainly geographic, there are no simple linguistic criteria for it.
Lǐ Róng 1985a is the first who uses a single isogloss to separate the major dialects within the Northern Mandarin area. The feature used for the classification is the development of ancient stop final words. The stop has disappeared in most Northern Chinese dialects and the earlier stop final words were apportioned between the other four tones. Table 5 describes the developments in the dialect areas proposed by Lǐ Róng. The numbers are the traditional numbers for the four tones.
2.3 Northern dialects
The development of stop final words is a very easy basis for the classification of Mandarin dialects. It is also used by Zhān Bóhuì to describe his groups, but only secondarily, after he has established areas based on geography.
On the basis of this criterion we can talk about the following groups of Northern Chinese:
•Central Plains Mandarin. Zav'jalova 1979:63 calls this group the Gānsù-Shǎnxī-Hénán dialect as opposed to Northern Gānsù (Lánzhōu-Yínchuān 兰州 银川). In Zhān Bóhuì's classification this dialect is divided between Guānzhōng (Central Shǎnxī Plains) dialect among the Northwestern dialects and Hénán and Huáiběi among the Northern Chinese dialects.
•Lánzhōu-Yínchuān Mandarin in Níngxià and northern Gānsù, including Héxī 河 西 (Gānsù Corridor). Zhān Bóhuì calls it the Níngxià-Gānsù subgroup of the Northwestern dialects. Zav'jalova (1979) calls it Northern Gānsù dialect. Further investigation might reveal that it is a development of Jìn dialects.
•Peking Mandarin, spoken mainly north of Peking (Hè Wēi, Qián Zēngyí and Chén Shūjìng 1986). Zhān Bóhuì calls it the Héběi dialect. Northeastern Mandarin also belongs to this group (Lǐ Róng 1985a, Lín Tāo 1987). Both are Northern dialects in Zhān Bóhuì's classification.
Northeastern Mandarin has also been regarded as a separate dialect area (Hè Wēi 1986b). On the northeastern side of the dividing isogloss retroflex sounds have coalesced with dentals. The tonal developments are basically the same as in Peking, except that it has more shangsheng than Peking. This is probably caused by immigration from the Jiāodōng-Liáoníng area.
•Northern Mandarin, spoken east and south of Peking until it meets Central Plains. In Zhān Bóhuì's classification it belongs to the Shāndōng group of Northern dialects.
•Jiāodōng-Liáoníng Mandarin in eastern Shāndōng (Jiāodōng 胶 东) and Liáoníng peninsulas. The group is called Jiāodōng by Zhān Bóhuì.
•Jiāng-Huái Mandarin roughly between the Yángzǐjiāng and the Huái River and Húběi northeast of Wǔhàn 武汉.
•Southwestern Mandarin, mainly spoken in Sìchuān, Yúnnán, southern Shǎnxī, and Húběi (Huáng Xuězhēn 1986).
•Jìn Dialect. The nucleus of this dialect is in Shānxī. It is also spoken in northern Shǎnxī, Inner Mongolia, and in Hénán in areas bordering on Shānxī. Forrest's Chin dialects, spoken mainly in Shānxī, (207–8) include Xī'ān 西安 in the Central Plains area, so his term does not cover the same area as the present term. In Lǐ Róng's classification this is not a Mandarin dialect.
Table 6 summarizes the differences (not the similarities) between Lǐ Róng's and Zhān Bóhuì's groupings.
|Zhān Bóhuì||Lǐ Róng|
It can be seen that Lǐ Róng's classification cuts across the geographical areas of Zhān Bóhuì.
From the previous description of Soviet Dungan peculiarities it is easy to see that the subgroup of Mandarin dialects to which Soviet Dungan belongs is the Central Plains Mandarin. Consequently that group is discussed in further detail below.
2.4 The extent of Central Plains
2.4.1 Extent of Central Plains in the east.
The area of the Central Plains Dialect is extensive and it has a peculiar shape. It extends all the way from the East China Sea to the Dungan villages in the Soviet Union, but in the north-south direction the area is not very broad. It only forms a narrow belt.
The area of Central Plains Mandarin meets the sea in northern Jiāngsù. In the south it extends to the Huái River, in western Ānhuī even south of the Huái River (Hè Wēi 1985b).
The following areas belong to Central Plains Mandarin: Hénán south of the river, southern Shāndōng up to the southern tip of Héběi, northern Jiāngsù up to the border with Shāndōng , northwestern Ānhuī down to Húběi all belong to the Central Plains area, as well as the southeastern tip of Héběi (Hè Wēi 1985b, 1986a).
Dialect areas are usually separated by bundles of isoglosses. If you choose a good isogloss, it might serve well, but it is more likely that one is not enough. In the case of Chinese dialects the long and narrow shape of the Central Plains area causes some doubts. A good dialect should have historical reasons for its existence (Lǐ Róng 1985b). Geography and communications should be considered when establishing what isoglosses should be used. The area of Central Plains Mandarin in the east is very natural because it roughly follows the old course of the Yellow River (Clark 1983:46).
As was stated above, the boundaries of the dialect areas are determined by one isogloss only, the development of the stop final words. We have some information on another isogloss, a feature typical of Soviet Dungan: the use of dentals instead of retroflex initials in some words (p. 8 above). This feature also extends as a narrow zone in the Yellow River basin and even to the north of Shāndōng and to Tiānjīn 天 津, which belongs to Northern Mandarin (Hè Wēi, Qián Zēngyí and Chén Shūjìng 1986:249, Zav'jalova 1979:117). Even Chānglí 昌 黎 northeast of Peking has this feature (Chānglí fāngyánzhì 1960). In Hénán it occurs in the northern part of the province in the Yellow River basin and in a few localities with the easternmost point in Xúzhōu 徐州 (Hè Wēi 1985a, 1985b).
2.4.2 Northwestern Central Plains
126.96.36.199 Extent in the northwest
In the northwest other dialects limit the area of Central Plains Mandarin in the north. To the east it is bounded by Jìn, which is spoken both in Shānxī, except the Central Plains Mandarin speaking southwestern part, and in northern Shǎnxī North of Yán'ān 延 安.
From Níngxià to the Soviet border Central Plains Mandarin is limited by Lánzhōu-Yínchuān territory (the southern extremity of Níngxià is Central Plains). Lánzhōu-Yínchuān extends in Gānsù south to Lánzhōu and to all of the Héxī Corridor. Dūnhuáng is the first place in the west were the Central plains is again spoken. From there the area of Lánzhōu-Yínchuān continues over northern Xīnjiāng. So the Central Plains Mandarin area has a gap in Héxī.
In Xīnjiāng the dialects south of the Altai belong to the Lánzhōu-Yínchuān group while the Altai region is basically Pekingese. The reason is that the Chinese settlement here is recent, dating from the period after Liberation in 1949. (Liú Lìlǐ and Zhōu Lěi 1986).
The dialects in southern Xīnjiāng belong to the Central Plains group (Liú Lìlǐ and Zhōu Lěi 1986). Their area starts from Shànshàn 鄯善 and Ruòqiāng 若 羌 and extends all the way to the Ili River in the west. The northern boundary clearly follows the watershed.
In the south the Central Plains Mandarin area is bounded by Southwestern Mandarin and Tibetan so that Eastern Qīnghǎi is Chinese speaking. The area of Southwestern Mandarin extends to Shǎnxī to some extent. It seems to have spread from Sìchuān along the Hàn 汉 River Valley and is spoken along the southern slopes of the Qínlǐng 秦岭 Range.
It is possible that in the west Central Plains Mandarin was originally associated with the Wèi 渭 River Valley.
188.8.131.52 Subgroups and their criteria.
There are three major subgroups of Central Plains Mandarin in the Northwest (Zhāng Shèngyù and Zhāng Chéngcái 1986). The two features that have been used to set up subgroups are:
1) the opposition between dental and velar nasal finals after non-low vowels (-en vs. -eng etc.).
2) the number of tones.
In the Central Shǎnxī Plains subgroup the -en vs. -eng opposition has been preserved. This dialect is spoken in Eastern Shǎnxī with Xī'ān as a representative.
The dialects in which this opposition is not preserved are divided into two groups. In the Shǎnxī-Gānsù subgroup there are four tones in isolated words. This dialect is spoken in western and southern Shǎnxī and southeastern Gānsù with outliers in northeastern Gānsù west of Níngxià and in the utmost west of the Chinese speaking part of Qīnghǎi. Dūnhuáng in Northern Gānsù also belongs to this group.
The two píngshēngs are not kept apart in the Central Gānsù subgroup. It is spoken in a narrow zone extending from eastern Gānsù and southwestern Níngxià to eastern Qīnghǎi.
The Southern Xīnjiāng subgroup shares the features of the Central Gānsù subgroup, but the two areas are not adjacent. (Liú Lìlǐ and Zhōu Lěi 1986). It is not clear why a separate group has to be established. These dialects essentially resemble Soviet Dungan, but not all dialects share the loss of the opposition shu vs. fu and the same reflex of zero initial. The distribution of these features seems to be rather erratic. They do not form contiguous territories.
3 The placement of Dungan
The Soviet Dungan literary language belongs to the Central Gānsù or the Southern Xīnjiāng subgroup. The Shǎnxī dialect of Soviet Dungan seems to belong to the Shǎnxī-Gānsù subgroup.
The area is still rather extensive and it is impossible to pinpoint the exact original location of Soviet Dungan. We ought to have information on further features. The only one there is information available on is the labialization of retroflex initials. This is not a feature that can be used as a criterion for dialect subgroups, because it does not define continuous territories. To the east of Xīnjiāng the feature occurs in two places. The first place is in southern Shǎnxī south of the Qínlǐng Range in two small areas separated by Southwestern Mandarin and the second in the Wèi River and Yellow River valleys north of the Wèi.
In these localities [pf] and [pfʻ] are also used instead of [tʂ] and [tʂx] before [u], that is, the affricates are also labialized. This may not be so different from дв and тф given by Imazov (1975) for Dungan, but these areas belong to the Shǎnxī-Gānsù subgroup.
In Qīnghǎi and adjacent areas of Gānsù only the fricatives have been labialized while the affricates are represented by [tʂ] and [tʂx] so those areas are closest to Soviet Dungan, especially those in eastern Qīnghǎi, because the dialects there belong to the Central Gānsù subgroup.
In Xīnjiāng there is information on some localities only. Apparently the areas with labialization are not contiguous there, either. It might be significant that Huòchéng 霍城 in Ili Valley agrees with Soviet Dungan in this respect.
If dialect features are not contiguous it is usually indicative of the fact that the feature is a relict. However, in this case this is not likely. [f] and [pf] are used in parts of Shāndōng as well, and it cannot be a feature that has earlier been used all over northern China and then disappeared from most of it.
There are some descriptions of the dialects of the general area where Soviet Dungan can be thought of coming. One of them is Xīníng 西宁 of the Shǎnxī-Gānsù subgroup. It can be regarded as a descendant of Soviet Dungan. By this I mean that Xīníng can be derived from Soviet Dungan, it has fewer segmental oppositions and only one that Soviet Dungan does not (between /v/ and /w/). However, the phonetical realization seems to be rather different. (Zhāng Chéngcái 1980, Zhāng Chéngcái and Zhū Shìkuí 1987).
The dialect of Dūnhuáng 敦 煌 belongs to the Shǎnxī-Gānsù subgroup (Zhāng Shèngyù 1985), but west of the Dǎnghé 党河 the dialect is of Lánzhōu-Yínchuān type. It gives a general impression of being rather similar to Dungan.
•Most similar is Yānqí 焉 耆 in Central Xīnjiāng of the Southern Xīnjiāng subgroup (Liú Lìlǐ 1988). In Yānqí the tone sandhi seems to differ somewhat from Soviet Dungan. The retroflex initials have been preserved before a labial medial, but otherwise the phonology is very similar to Soviet Dungan.
To conclude, Soviet Dungan belongs to the Central Gānsù or Southern Xīnjiāng subgroup of Central Plains Mandarin. It is further set apart from some dialects of the subgroup by the non-occurrence of retroflex fricatives before -u. and the use of n- for zero initial. This kind of dialect is spoken in Southern Xīnjiāng, in Qīnghǎi and in Southern and Central Gānsù. When such details as tonal realization and tone sandhi are taken into account, Soviet Dungan is most closely related to Southern Xīnjiāng dialects. It is possible that the basis of standard Soviet Dungan was actually the dialect of the group of Dungans that crossed over from Ili Valley. Soviet Dungan can thus be regarded as westernmost extention of a continuum of dialects, not an island separated from its place of origin.
Bào Hòuxīng and Yán Sēn 1986 鲍厚星 颜森: Húnán fāngyánde fēnqū 湖南方言的分区 (The grouping of the dialects of Húnán province). Fāngyán 方 言 , 4, 273–276.
Chānglí fāngyánzhì 1960 昌黎方言志. Kēxué chūbǎnshè 科学出版社, Peking.
Chao, Yuen Ren 1968. A Grammar of Spoken Chinese. University of California Press, Berkeley and Los Angeles.
Clark, Champ. 1983. Flood. Time-Life Books, Amsterdam.
Forrest, R.A.D. 1948. The Chinese Language. Faber and Faber, London.
Hashimoto, Mantaro 1978. Tookango monji no onpyooka. Writing and language reference materials 2.
Hè Wēi 1985a 贺 崴. Hénánshěng xīnánbù fāngyánde yǔyīn yì-tóng 河南省西南部方言的语音异同 (Phonetic differences and similarities among the dialects of southwestern Hénán). Fāngyán 方言, 2, 119–123.
Hè Wēi 1985b 贺 崴. Hénán Shāndōng Wǎnběi Sūběide guānhuà 河 南山东 苏北皖北的官话 (The grouping of the dialects of Hénán, Shāndōng, northern Ānhuī and northern Jiāngsù). Fāngyán 方言, 3, 163–170.
Hè Wēi 1986a 贺 崴. Jì-Lǔ-Yù sān shěng pílián dìqūde fāngyán fēnjiè 冀鲁豫三省毗连地区的方言分界 (A note on the dialects in the border region of Héběi, Shāndōng and Hénán Provinces). Fāngyán 方言, 1, 46–49.
Hè Wēi 1986b 贺 崴. Dōngběi guānhuàde fēnqū (gǎo) 东北官话的分区(稿) (The Grouping of Mandarin dialects of the Northeast). Fāngyán 方 言, 3, 172–181.
Hè Wēi, Qián Zēngyí and Chén Shūjìng 1986 贺 崴 钱曾怡 陈淑静. Héběishěng Běijīngshì Tiānjīnshì fāngyánde fēnqū (gǎo) 河北省北京市天津市方 言 的分区(稿) (The grouping of the dialects of Héběi, Běijīng and Tiānjīn). Fāngyán 方言, 4, 241–252.
Hóu Jīngyī 1986 侯 精一. Jìnyǔde fēnqū (gǎo) 晋语的分区(稿) (The grouping of Jìn dialects). Fāngyán 方 言, 4, 253–261.
Hóu Jīngyī, Wēn Duānzhèng and Tián Xīchéng 1986 侯精一 温端政 田希诚. Shānxī fāngyánde fēnqū (gǎo) 山西方言的分区(稿) (The grouping of Shānxī dialects). Fāngyán 方言, 2, 81–92.
Huáng Xuězhēn 1986 黄 雪贞. Xīnán guānhuàde fēnqū (gǎo) 西南官话的分区(稿) (The grouping of Southwestern Mandarin dialects). Fāngyán 方言, 4, 262–272.
Imazov 1975 = Имазов, М. Х. 1975. Фонетика дунганского языка. Илим, Фрунзе
Janshansin 1938 = Janşjansьn, Jusup 1938. Gansьn de Şanşi Fonjan. In Hashimoto 1978, 81–102.
Janshansin 1965 = Янщянсын 1965. Xуэйзў йүянди Тохма фон–ян. Ilim, Frunze.
Jusupov 1977 = Юсупов, И.И. 1977. Советские дунгане в период строительства социализма. Илим, Фрунзе.
Kalimov 1968 = Калимов, А. 1968. Дунганский язык. Языки народов СССР, том 5, 475–488.
Lǐ Róng 1985a 李 荣. Guānhuà fāngyánde fēnqū 官 话方言的分区 (The grouping of the Mandarin dialects). Fāngyán 方言, 1, 2–5.
Lǐ Róng 1985b 李 荣. Hànyǔ fāngyán fēnqūde jǐdiǎn yìjiàn 汉语方言分区 的几点意见 (Remarks on the grouping of Chinese dialects). Fāngyán 方言, 3, 81–88.
Lǐ Róng 1985c 李 荣. Hànyǔ fāngyán fēnqūde jǐdiǎn yìjiàn (2) 汉语方言分区 的几点意见(二) (Remarks on the grouping of Chinese dialects (II)). Fāngyán 方言, 3, 161–162.
Lín Tāo 林焘 1987. Běijīng guānhuàqūde huàfēn 北京官话区的 划 分 (The Demarkation of the Běijīng Mandarin Group). Fāngyán 方言, 3, 166–172.
Liú Lìlǐ 刘俐李1988. Yānqí yīnxì jìlüè 焉耆 音 系记略 (A Sketch of the Yānqí dialect, Xīnjiāng A. R.). Fāngyán 方 言, 1, 31–41.
Liú Lìlǐ and Zhōu Lěi 1986 刘 俐李 周磊. Xīnjiāng Hànyǔ fāngyánde fēnqū (gǎo) 新疆汉语方言的分区(稿) (The grouping of the dialects of Xīnjiāng). Fāngyán 方言, 3, 161–171.
Nurmekund, Pent 1979. Ausgewählte Kapitel der dunganischen Sprachgesichte. Tartu riikliku ülikooli toimetised 507. Oriental Studies V, 5–21, Tartu.
Qián Zēngyí, Cáo Zhìyūn and Luó Fúténg 钱曾怡 曹志耘 罗福腾 1987. Héběishěng dōngnánbù sānshíjiǔ xiàn-shì fāngyīn gàiyào 河 北 省东南部三十九县市方音概要 (The phonology of thirty-nine cities and counties in the south-east of Héběi province). Fāngyán 方言, 3, 173–178.
Riedlinger, Heinz 1989. LIKBEZ – Alphabetisierung bei den sowjetischen Dunganen seit 1927 und ihr Zusammenhang mit den Latinisierungsbestrebungen in China. Chinathemen band 37. Brockmeyer, Bochum.
Rimsky-Korsakoff, Svetlana 1967. Soviet Dungan: the Chinese Language of Central Asia: Alphabet, phonology, and morphology. Monumenta Serica 26, 352–421.
Rimsky-Korsakoff Dyer, Svetlana 1979. Soviet Dungan Kolkhozes in the Kirghiz SSR and the Kazakh SSR. Oriental Monograph Series No. 25. Faculty of Asian studies in association with Australian National University Press, Canberra.
Salmi, Olli 1980. Tone and Stress in Soviet Dungan. Hurme, P. ed. Papers in Speech Research. Voice, Speech and Language: Reports and Reviews. Publications from the Institute of Finnish Language and Communication. University of Jyväskylä. Jyväskylä.
Salmi, Olli 1984. The Aspectual System of Soviet Dungan. Folia Fennistica & Linguistica. Vuosiartikkelit 1984. Tampereen yliopiston Suomen kielen ja yleisen kielitieteen laitoksen julkaisuja 11. University of Tampere, Tampere.
chi=Шыйүәди Чи. Biweekly. Frunze.
Sushanlo=Сушанло М. 1979. Борьба дунган против чинской династии и преселение их в Киргизию и Казахстан. [Sushanlo, M. The fight of Dungans against the Ch'ing Dynasty and the Resettlement of One Part of them in Kirghizia and Kazakhstan. Summary.] Tartu riikliku ülikooli toimetised 507. Oriental Studies V, 22–66. Tartu.
Thompson, Laurence C. 1965. A Vietnamese Grammar. University of Washington Press, Seattle.
Trubetzkoy, Nicholas S. 1939. Aus meiner phonologischen Kartothek I: Das phonologische System der dunganischen Sprache. Travaux du Cercle Linguistique de Prague 8, 22–26.
Yuán Jiāhuá et al 1960 袁家骅等. Hànyǔ fāngyán gàiyào 汉语方言概要. Wénzì Gǎigé Chūbǎnshè 文字改革出版 社.
Zav'jalova 1973 = Завьялова, О. И. 1973. Тоны в дунганском языке. Народы Азии и Африки 3, 109–119.
Zav'jalova 1979 = Завьялова, О. И. 1979. Диалекты Ганьсу. Наука: Москва.
Zhān Bóhuì 1981 詹伯慧. Xiàndài Hànyǔ Fāngyán 现代汉语方言. Húběi Rénmín Chūbǎnshè 湖北人民出版社.
Zhāng Chéngcái 1980 张成材. Xīníng fāngyán jìlüè 西宁方言记略 (A sketch of the Xining dialect). Fāngyán 方言 1980, 4, 282–302.
Zhāng Chéngcái and Zhū Shìkuí 1987 张成材 朱世奎. Xīníng fāngyánzhì 西宁方言志. Qīnghǎi rénmín chūbǎnshè 青海人民出版社.
Zhāng Shèngyù and Zhāng Chéngcái 1986 张盛裕 张成材. Shǎn-Gān-Níng-Qīng sì shěngqū Hànyǔ fāngyánde fēnqū 陕 甘宁青四省 区汉语方言的分区 (The grouping of the dialects of Shǎnxī, Gānsù, Níngxià and Qīnghǎi). Fāngyán 方 言, 2, 93–105.
Zhāng Shèngyù 1985 张盛裕. Dūnhuáng yīnxì jìlüè 敦煌音系记略 (An outline of Dūnhuáng phonology). Fāngyán 方言, 2, 134–139.
Zhōu Qìngshēng 1987 周庆生. Zhèngzhōu fāngyánde shēng-yùn-diào 郑州方言的声韵调 (The phonology of Zhèngzhōu dialect). Fāngyán 方言, 3, 190–199.