Janhunen J. The Mongolic Languages

Подождите немного. Документ загружается.

‘he would eat’, though the simple ending -qu is also attested, e.g. ida-qu-bi ‘I would eat’.

The perfect subjunctive incorporates the perfective participle marker and has the ending

-xså-sa, e.g. ida-xså-sa-bi ‘I would have eaten’. The perfect passive subjunctive, finally,

incorporates additionally the passive derivative suffix .gh- and has the ending .gh-så-sa,

e.g. ida.gh-så-sa-bi ‘I would have been eaten’.

PREDICATIVE PERSONAL ENDINGS

Like several other Modern Mongolic languages, Moghol has a system of predicative per-

sonal endings, which are attached to the finite forms after the temporal-aspectual (or

modal) markers. A specific feature of Moghol is, however, that the personal endings are

differentiated for three numbers: singular, dual, and plural. It is true, the system is not

complete, in that the distinction between the dual and the plural is only made in the sec-

ond and third persons (vowel alternation å : u).

Diachronically, the personal endings are a heterogeneous collection of pronominal ele-

ments, nominal number markers, and obscured verbal suffixes. Altogether there are five

different sets (Table 12.8), the first of which (I) is combined with the narrative marker

-m : -n, the second (II) with the durative marker -na, the third (III) with the terminative

and confirmative markers -ba resp. -la (with variants), and the fourth (IV) with the resul-

tative marker -zha (with variants). The third set is also used with the subjunctive marker

-sa, while the imperative conjugation, with no modal suffix, uses a fifth set of personal

endings (V).

It is obvious from the data that the five sets of personal endings are multiply inter-

dependent both synchronically and diachronically. There is a particularly close connection

between the first and second sets, and also between the third and fourth sets. The fifth set,

consisting mainly of restructured rudiments of the original system of imperative markers,

is most different from the others. In this set, the Common Mongolic optative (1p. sg. -su

< *-sU), permissive (3p. -gha/h < *-gV), voluntative (1p. du. -ya/h : pl. -yan < *-yA/n),

and benedictive (2p. du. -tå : pl. -tu ~ -tuna < *-gtUn) can be immediately identified.

SYNTAX

Like the intonation patterns, the sentence structure of the Moghol language, both spoken

and written, has been intensively influenced by the Tajik-Persian linguistic environment.

260 THE MONGOLIC LANGUAGES

TABLE 12.8 MOGHOL PREDICATIVE PERSONAL ENDINGS

I II III IV V

sg. 1p. -m-bi [-na]-m-bi -bi -bi -su

2p. -n-ci [-na]-n-ci -ci -n-ci -h

3p. -m -Ø [-na]-h -h -h -gha/h

du. 1p. -m-da/h [-na]-m-da/h -bda/h -bda/h -ya/h

2p. -n-tå [-na]-n-tå -tå -n-tå -tå

3p. -m-nåd [-na]-nåd -nåd -nåd -gha/h

pl. 1p. -m-da/h [-na]-m-da/h -bda/h -bda/h -yan

2p. -n-tu [-na]-n-tu -tu -n-tu -tu(na)

3p. -m-nud [-na]-nud -nåd -nud -gha/h

The normal word order is subject–object–predicate (SOV), with the predicate slot

being filled by either a finite verb or a predicatively used noun. There is regular agree-

ment between the subject and the predicate with regard to number and person. The

normal sentence may, however, be altered in poetic language in accordance with the rules

of Persian metres, rhymes, and rhythms, which also had to be observed by a Moghol

poet. For instance, to achieve the correct emphasis, the finite verb can be placed at the

beginning of the sentence (VSO), as in å mida-m daidån deksh ti ku-



i kåfer bål/u-m

‘And God the Exalted knows this one who becomes an infidel.’

Another element of alien syntactic influence originates in the religious language

derived from Arabic. Arabic elements, introduced by Moghol mullahs into their native

language, have always been farz ‘obligatory’ for every Muslim in his religious life, and

they have thus become effective also in the everyday life and language of the Moghol.

The Persian-Arabic influence is responsible for the expression of hypotactic structures

by using subordinate dependent clauses introduced by subordinating conjunctions and

relative pronouns, non-existent in most other Mongolic languages. The subordinate

clause can either precede or follow the main clause. The following subordinating

conjunctions are attested:

(1) oghlah ‘that’ (functionally corresponding to Arabic anna), introduces a consecu-

tive nominal dependent clause, e.g. å ulah gah-gh/u-m oghlah yad-ni cidån-i be-h ‘and

it should not be said that his hand is his strength’;

(2) ul ‘so that’ (also ‘until’, cf. Mongolic *ulam ‘further, still more’), introduces a final

dependent clause, e.g. du maktab såu.l-gha-h ul sah

aqå’ed tah awåmer wåcugh bol-ga-

h ‘he should be placed in a school, so that he can be made aware of the articles of faith

and the orders of communication’;

(3) munkeh ‘that’ (from Moghol mun ‘this’ and Persian keh ‘that’), introduces a con-

secutive dependent clause, e.g. ‘. . .qunah munkeh ghar/u-m sah munah gaiti qoluc/o-xci

‘. . .after he goes out of this world as a believer’;

(4) ul munkeh ‘until’ (from ul and munkeh), introduces a terminal dependent

clause, e.g. wa esah jiji-zha-h sah ti ul munkeh oku-ba-h qoluc/o-xci ghar/u-xci bar

oghlati tu eri-yi daidån dekh be-h ‘and he did not turn away from this until he died as

a believer committing transgressions, so really he is under the volition of God the Exalted’;

(5) kull ghåt ‘when’ (from Arabic kull ‘all, every’ and Moghol ghåt ‘time’), introduces

a temporal dependent clause, e.g. kull ghåt kela-ba-h daidån deksh muså-i kela-ba-h

te-



i du kelanni ‘when God the Exalted spoke to Moses, he spoke to him in his language’;

(6) ticur ‘as soon as’ (from Moghol ti ‘this’ + ucur ‘time’), introduces a successive

dependent clause, e.g. å mida-m oghla ti oshtin å-m ticur å-lgha-h ti-’i ‘and he knows

how it is, as soon as he creates it’;

(7) sah ‘if’ (identical with the subjunctive marker -sa, cf. similar forms in both

Mongolic and Turkic), introduces a conditional dependent clause, e.g. farz azhar sah ti

nika-yah å-la-h å sah oroxshi lah be-la-h ‘Farz Azhar [must be recited] if he has been

there himself alone, and if there has not been a precentor’;

(8) sin ‘even if, though’, introduces a concessive dependent clause, e.g. å lah ga-mdah

oghlah ti mughatti-xsan be-h ti-du sin a-gha-h ghar/u-xci ‘and we do not say that he is

one who is forever retained in it even if he becomes a transgressor’;

(9) mun agar sah ‘because’ (from Moghol mun ‘this’ and Persian agar ‘if’, to which

Moghol abl. sah ‘from’ is added), introduces a causal dependent clause, e.g. å gaiti du

kolkah joz-nud-neh moxdas be-h mun agar sah ti

ain-nud be-h ‘and the world is in all

its parts created, because it consists of substances’.

MOGHOL 261

Subordinate causal clauses can also be paraphrased in various ways. Two such para-

phrases are imah ga-zhi . . . ga-h . . . ‘saying what . . . one must say . . .’ and . . . ga-h eh

be-h imah ga-la-h oghlah . . . ‘. . . one must say it is what one had said that . . .’, e.g.

imah ga-zhi du ti ga-h gaimahku-yi åyatur be-h ‘because in this [there] is the suspension

of the attribute’ (literally: ‘. . . saying what – in this – one must say – is the suspension

of the attribute’), å mun ga-h eh be-h imah ga-la-h oghlah daidån deksh mida-m lah ådål

mida.gh mån ‘and this is, because God the Exalted knows not according to our knowl-

edge’ (literally: ‘and this is – one must say it is what one had said – that God the Exalted

knows not according to our knowledge’).

Relative words introducing subordinate relative clauses include: ti ‘which’, e.g. du

kelan ti sifat bar ti-



i be-h ‘by means of the speaking which is an attribute with him’;

imah ~ timah (ti+imah) ‘what, [the one] that, [he] who’, e.g. ti imah-du mida-na-h

daidån ‘this, by what means God will know’, å timah såyar be-h ‘and that which is

worthy of belief’, ådål timah dekr ki-ba-h ‘like he who mentioned it’, å timah ådåli-ne

be-h ‘and that which is similar to it’; ku ‘who’, e.g. har ku enni ura-sa-yah ga-ba-h mumen

sådiq be-na-h ‘everybody who recited this from his heart will be a sincere believer’.

For the expression of negation Moghol uses the following negative particles of

Common Mongolic origin, all of which are placed before, and in some cases prefixed to,

the word to be negated: ulah ~ ula- ~ ul- ~ lah ~ la- ‘not’; ugai(-) ~ gai(-) ‘not, without’;

essah ~ essa ~ esah ~ sa- ‘not’. Additionally, the Persian preposition bi ‘without’ is used.

The negation of nominal words, including nominal forms of verbs (participles), takes

place with bi, lah, ugai ~ gai, e.g. bi ci ‘without you’, bi jonun ‘without madness’, lah

or/o.xsh-i lah quina.xsh-i ‘not the forward, not the backward’, lah nudun-du ugai cekin-

du ‘to the one without eyes and without ears’, gai ådål ‘unlike’. In the finite conjugation,

the particle ulah ~ ula- ~ ul- ~ lah ~ la- is used to negate the narrative and durative forms,

e.g. ulah bai-m ulah lula-na-h ‘he does not stay and he will not rotate’, ceni ula-uzha-

na-mbi ‘I will not see you’, ul-årå-m ‘he does not enter’, lah anqas/u-m lah ulas/u-m ‘he

is not hungry, nor thirsty’, la-mida-mbi ‘I do not know’. The negation of existence is

expressed as ugai+be-h ~ gai+be-h

‘it is not, there is not; does not exist’ (with the cop-

ula +be-h). The other finite forms are negated by bi, lah, or essah ~ essa ~ esah ~ sa-,

e.g. bi dur-pa-ci ‘you did not burn’, lah be-la-h ‘has not been’, essah ida-ba-nud ‘they

did not eat’; sa-ira-zha-nci ‘you were not coming’. The prohibitive (negative imperative)

is expressed by the (Mongolic) particle bi(-) or be(-), e.g. bi tus-gha ‘he must not touch’;

be-dur.ga-l.gha-tu ‘you must not allow him to get burned’.

LEXICON

The intensive language contact with the Iranian environment has particularly profound-

ly influenced the Moghol lexicon. More than 70 per cent of the words used in Moghol

are of Persian-Arabic origin. The Moghol lexicon comprises, therefore, (1) native

Mongolic words, (2) loanwords from Persian-Arabic, (3) ‘Mogholizations’ (Moghol

adaptations) of foreign words, and (4) loan translations. A study of the early Moghol

word lists beginning with Leech (1838) reveals relatively little of the development of the

lexicon, though the variation observable in the material may reflect some diachronic or

dialectal patterns (see the section on Dialects above).

Owing to the impact of the Islamic society, many ordinary words borrowed from

Persian are used as technical terms even in everyday life, e.g. namâz ‘devotional exer-

cise of the prescribed prayer’. Some such terms have, however, been replaced by loan

translations, e.g. usunghar for Persian âbdast ‘ablution’. Similar loan translations have

262 THE MONGOLIC LANGUAGES

been made for some special terms of Islamic theology, e.g. unang coqu-håt for Arabic

khaqâ



iq ‘truth, sincerity (as the essence of a thing)’. On the other hand, items of non-

religious idiomatic usage have also been adopted by way of loan translation, e.g. nur-

sah-i for Persian az rû-yi ‘on account of’ (literally: ‘from the face of’), as in nur-sah-i

mart ‘on account of forgetting’.

Even verbal morphological patterns have occasionally been translated from Persian.

An interesting example is offered by the documented use of the Moghol plain verbal

stem tom ‘to cut’ as a past tense form in the sense ‘he has cut’, replacing the normal

conjugated form term. sg. 3p. tom/u-ba-h ‘he has cut’. The reason is that the Persian

counterpart of Moghol tom- ‘to cut’ is ciid, which happens to function both as the

general preterite stem of the verb and as the specific personal form sg. 3p. ‘he has cut’.

Against this massive alien influence, there are examples which suggest that the

Moghol in some cases may have tried to avoid Persian loanwords by using native

paraphrases, e.g. usun-i dotana.ki-ni ‘the one being inside the water’ for Persian

morghâbi ‘duck’, usun-i nure.ki ‘the one being on the face of the water’ for Persian kashti

‘boat’, bughur-i sundun-i ‘tooth of the mouse’ for Persian berenj ‘rice’. The reasons

underlying such paraphrases may be connected with some unknown social or cultural

factors. Similar factors may underlie the ‘Mogholized’ shapes of certain loanwords,

e.g. mahtoi ‘moon’ for Persian mâh id. (the native Moghol word is caqel or ceqeldaur).

Altogether, the remarkable impact of the non-Mongolic linguistic environment on

Moghol lexicon and grammar (syntax, morphosyntax, and even morphological structures

like the ezafe construction) make Moghol look like a true Inner Asian creole language,

worthy of careful study in its areal setting.

REFERENCES AND FURTHER READING

Buhe [Buigae] (1996) Muqhul Galav u Sudulul, Guigaquda: vUibur Muvgqhul uv Yagae

Surqhaqhuli jiv Gablal uv Quriie.

Heissig, Walther (1969) ‘Der Moghol-Dichter

Abd al-Qader’, Zentralasiatische Studien 3: 431–8.

Heissig, Walther (1974) Schriftliche Quellen in Mogholi, 1. Teil: Texte in Faksimile, Opladen:

Westdeutscher Verlag.

Homam, S. S. S. (1970). ‘Sawanih-i muhtasar-i mulla

Abd al-Qadir mughul hirati’, Aryana 23:

20–26.

Homam, S. S. S. (1972). ‘A Brief Criticism on the So-Called “Zirni Manuscript” ’,

Zentralasiatische Studien 6: 561–74.

Iwamura, Shinobu (1961) The Zirni Manuscript: A Persian-Mongolian Glossary and Grammar

[= Results of the Kyoto University Expeditions to the Karakoram and the Hindukush, vol. 6],

Kyoto: Naigai Printing Company.

Iwamura, Shinobu and Schurmann, H. F. (1954) ‘Notes on Mongolian Groups in Afghanistan’, in

Silver Jubilee of the Zinbun Kagaku Kenkyusyo Kyoto University, Kyoto: Kyoto University.

Kuz’menkov, Ye. A. (1997) ‘Mogol’skii yazyk’, in Mongol’skie yazyki – Tunguso-man’chzhurskie

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

Nauk & Izdatel’stvo Indrik, pp. 91–7.

Ligeti, Louis [Laiosh] (1955a) ‘O mongol’skix i tyurkskix yazykax i dialektax Afganistana’, Acta

Orientalia Hungarica 4: 93–117.

Ligeti, Louis (1955b) ‘Le lexique moghol de R. Leech’, Acta Orientalia Hungarica 4: 119–58.

Ligeti, Louis (1964) ‘Les voyelles longues en moghol’, Acta Orientalia Hungarica 17: 1–48.

Motamedi, A. A. (1956) ‘Vestige des mongoles en Afghanistan’, Afghanistan 1: 17–24; 2: 15–20.

Poucha, Pavel (1961) ‘Die Sprache der Mogholen in Afghanistan und die Theorie der

Mischsprachen’, Studia Mongolica 2: 251–70.

MOGHOL 263

Pritsak, Omeljan (1964) ‘Das Mogholische’, in Mongolistik [= Handbuch der Orientalistik I: V, 2],

pp. 159–84.

Ramstedt, G. J. (1905) ‘Mogholica: Beiträge zur Kenntnis der Moghol-Sprache in Afghanistan’,

Journal de la Société Finno-Ougrienne 22 (4): i-iii, 1–60.

Schurmann, H. F. (1962) The Mongols of Afghanistan: An Ethnography of the Moghols and related

peoples of Afghanistan [= Central Asiatic Studies 4], ’s-Gravenhage: Mouton.

Shagdarov, L. D. and Kazanceva, A. M. (1968) ‘O yazyke mogolov Afganistana (po materialam Sh.

Iwamura i X. F. Shurmana)’, in O zarubezhnyx mongolovednyx issledovaniyax po yazyku, Ulan-

Udè: Buryatskoe knizhnoe izdatel’stvo.

Weiers, Michael (1970) ‘Zu den langen Vokalen in der Moghol-Sprache’, Zentralasiatische Studien

4: 467–73.

Weiers, Michael (1971) ‘Das Moghol-Vokabular von W. R. H. Merk’, Zentralasiatische Studien 5:

157–89.

Weiers, Michael (1972) Die Sprache der Moghol der Provinz Herat in Afghanistan

(Sprachmaterial, Grammatik, Wortliste), Opladen: Westdeutscher Verlag.

Weiers, Michael (1973) ‘Eine fünfsprachige Wörtersammlung aus dem Gebiet der Moghol von

Herat in Afghanistan’, Zentralasiatische Studien 7: 503–23.

Weiers, Michael (1974) ‘Grundzüge der Sprache der Moghol von Afghanistan’, in W. Voigt (ed.)

Supplement 2, XVIII. Deutscher Orientalistentag vom 1. bis 5. Oktober 1972 in Lübeck:

Vorträge, Wiesbaden: Franz Steiner Verlag.

Weiers, Michael (1975a) Schriftliche Quellen in Mogholi, 2. Teil: Bearbeitung der Texte, Opladen:

Westdeutscher Verlag.

Weiers, Michael (1975b) ‘Die Sprache der Hazara und der Mongolen von Afghanistan in lexiko-

statistischer Sicht’, Afghanistan Journal 3: 98–102.

Weiers, Michael (1976a) ‘Hypotaktische Modellsprachen und parataktische Replikasprachen – Ein

Beitrag zur Linguistik des sprachlichen Kontakts’, Zentralasiatische Studien 10: 411–29.

Weiers, Michael (1976b) ‘Beispiele für Sprachmischung in der geschriebenen Moghol-Sprache und

deren Bezug auf Forschungsziele in der Altaistik’, in W. Heissig (ed.) Altaica Collecta,

Wiesbaden: Harrassowitz, pp. 361–8.

Weiers, Michael (1977a) Schriftliche Quellen in Mogholi, 3. Teil: Poesie der Mogholen, Opladen:

Westdeutscher Verlag.

Weiers, Michael (1977b) ‘Prinzipien für Notation und Notierung im Wörterbuch einer

Mischsprache (Vorbemerkungen zu einem Mogholisch-Deutschen Wörterbuch)’,

Zentralasiatische Studien 11: 389–409.

Weiers, Michael (1978) ‘Voraussetzungen für Sprachwandel bei Sprachen im Kontakt’, Central

Asiatic Journal 22: 286–319.

Weiers, Michael (1982) ‘Aus der Poesie der Mogholen’, Acta Orientalia Hungarica 36: 563–74.

Weiers, Michael (1992) ‘Das Schriftmogholi der Mongolen von Afghanistan: Ein Abriß’,

Zentralasiatische Studien 23: 60–6.

Weiers, Michael (1992–3) ‘Argun aller alten Herren, Čingiz aller Könige, und die nächtliche Reise

des Propheten: Mongolische Tradition und muslimische Inkulturation in der Dichtung der

Mongolen von Afghanistan’, Acta Orientalia Hungarica 46: 61–79.

264 THE MONGOLIC LANGUAGES

265

CHAPTER THIRTEEN

SHIRA YUGHUR

Hans Nugteren

The Shira Yughur belong to the linguistically heterogeneous Yughur (Yugur, Yugu)

nationality of China, which consists of Mongolic, Turkic, Chinese, and perhaps Amdo

Tibetan speakers. The Shira Yughur are not listed separately in the census data. Junast

(1981) estimated their number at one third of the (then) total population of 9,000 Yughur,

which would make them the second-smallest Mongolic-speaking people in China (after

the Khamnigan). The Shira Yughur language is spoken in the central part of Gansu

Province, northwestern China. Most speakers are found in Sunan Yughur Autonomous

County, in the townships (communes) of Kangle, Hongshiwo, and Qinglong of Kangle

District, the townships of Beitan and Dongtan of Huangcheng District, and in Dahe

District. Some further speakers live in the adjoining Jiuquan County.

The ethnonym Yughur, or Yoghor, as used by the Yughur themselves, denotes the

entire Yughur nationality. This also applies to the extended term Shira Yughur or Shira

Yoghor ‘Yellow Yughur’, which corresponds to the term Sarygh Yughur of the Turkic-

speaking Yughur. In the official classification the Mongolic-speaking Yughur are identi-

fied as the Eastern Yughur (Dongbu Yugu in Chinese), while the Turkic-speaking Yughur

are referred to as the Western Yughur (Xibu Yugu). These terms have no equivalents in

the Yughur languages. Diachronically, the name Yughur or Yoghor is a development of

the ancient Turkic ethnonym Uighur, and, in view of this, the Yughur are in Western lit-

erature traditionally known as the ‘Yellow Uighur’. There is, however, no evidence sup-

porting the assumption that the Shira Yughur are Mongolized Turkic-speakers.

Another term, which specifically denotes the Shira Yughur language, is Nggar (also

Anggar, Inggar), or Nggar lar (with lar ‘speech, language’). Taxonomically, this lan-

guage seems to hold an intermediate position between the central group of Mongolic

(including Mongol proper, Oirat, and others) and the languages of the Gansu-Qinghai

complex (including Mongghul, Mangghuer, Bonan, and Santa). A number of innovations

shared with the latter suggest that these are the closest relatives of Shira Yughur. At the

same time, Shira Yughur retains several phonological and morphological features that

have been lost in the other languages of the Gansu-Qinghai complex.

DATA AND SOURCES

The first known Shira Yughur materials, about 200 words, were collected and published

by the Russian explorer G. N. Potanin (edited by A. O. Ivanovskii) more than a century

ago (Potanin 1893). Other short wordlists were compiled by C. G. E. Mannerheim (1911)

and M. Hermanns (1951). The first actual linguist to visit the Shira Yughur was

S. Ye. Malov, but his materials, edited by W)adys)aw Kotwicz (1950), are also of a very

limited scope. Malov (1957) worked more extensively on Sarygh Yughur, a language

subsequently studied by E. R. Tenishev (1976) and also discussed by Kaare Thomsen

(1959) and Reinhard F. Hahn (1998).

Information on Shira Yughur began to increase only with the Sino-Soviet expedition

in the mid-1950s. The first major publication resulting from this work was a sketch of

both Eastern and Western Yughur, written jointly by E. R. Tenishev and B. X. Todaeva

(1966). Todaeva (1997) later published another brief synopsis of Shira Yughur. Chinese

scholars also worked on the language, but because of political circumstances they had to

postpone their publications until the 1980s. Junast (1981) then published a Shira Yughur

grammar with vocabulary, while Bulchulu (1984) prepared a more elaborate vocabulary

with etymological indications. Bulchulu and Jalsan also published a volume of Shira

Yughur texts and everyday phrases (1988) as well as a comparative grammar (1990).

Little is known about the dialectal subdivision of Shira Yughur, but it is clear from the

sources that there is some variety within the language, at least as far as phonetics is con-

cerned. Bulchulu and Jalsan mention the Qinglong and Hongshiwo dialects and present

a small list of undramatic phonetic differences, which mainly concern the vocalism.

Materials collected by other authors cannot simply be attributed to either one of these

dialects, and, therefore, it may be expected that a more careful investigation of other

localities will yield further variants of the language. This will also make it possible to

work on the diachrony of Shira Yughur. So far, most diachronic work on Shira Yughur

has concerned the layers of the lexicon (Róna-Tas 1962, Nugteren and Roos 1996, 1998).

The Shira Yughur language is not written, nor is it used in any media or taught in

schools. In the present chapter the language is nevertheless presented in a slightly

standardized form. Most examples derive from the recent Chinese (Inner Mongolian)

publications mentioned above. The materials of Junast have been partly reinterpreted and

retranscribed to match the generally more detailed analysis of Bulchulu and Jalsan.

Variant spellings have been left intact in so far as they seem to reflect actual dialectal or

idiolectal differences (rather than merely different views or transcriptions of the same

underlying pronunciation).

SEGMENTAL PHONEMES

Shira Yughur has seven phonemic vowel qualities (Table 13.1), all of which also have

long counterparts: aa ee ii oo öö uu üü. The long vowels are possibly best analysed as

monophonemic. All vowels can occur in both the initial and non-initial syllables. There

is no subsystem of reduced vowels. The number of diphthongs is hard to determine but

it is possible that the only independent ones are ai oi üi. By contrast, both [øi] and [ui]

seem to be variants of oi, while [i] is a variant of ai üi üü. There are also variations of

the type [ai ~ ei ~ ii] and [yi ~ y].

Many monophthongs appear to be unstable as well, both in quality and in quantity.

Long ee oo öö are often raised so that they merge with ii uu üü, respectively. Short i is

typically pronounced as a mid-central vowel [], but it occasionally appears as [i] when

following a palatal consonant. Similarly, short ü is usually pronounced as a central vowel

[] except when following a palatal consonant. As a consequence, long ii and üü tend to

266 THE MONGOLIC LANGUAGES

TABLE 13.1 SHIRA YUGHUR VOWELS

uü i

oö e

be shortened into [i] and [y], since the qualitative difference is sufficient to keep them

separate from their short counterparts, e.g. ciig [ti ~ ti] ‘moisture’ vs. cig [t]

‘achnatherum’, shiira [ira ~ ira] ‘leg’ vs. shira [ra ~ ra] ‘yellow’. Many other long

vowels have developed short forms or variants: sala (< *salaa) ‘branch’, derme

(< *deerme) ‘thief’.

The scarcity of oppositions which could really cause confusion may eventually con-

tribute to the loss of distinctive vowel quantity. There are, nevertheless, some minimal

pairs which can still illustrate the potential phonological importance of the quantitative

correlation, cf. e.g. tawin ‘fifty’ vs. taawin ‘five’, xana ‘where’ vs. xanaa- ‘to cough’, ula

‘sole’ vs. uula ‘mountain’, jun ‘summer’ vs. juun ‘hundred’.

The consonant system has twenty-five members (Table 13.2). Words recently adopted from

Mandarin Chinese may additionally contain ‘loan phonemes’ corresponding to the Chinese

retroflex affricates and fricatives (Pinyin ch zh sh r) and their alveopalatal counterparts

(q j x). Chinese words which retain their original pronunciation will be cited in boldface.

As in the other Mongolic languages of the region, aspiration is the actual distinguishing

feature between the strong obstruents p t ts c k q and the corresponding weak segments

b d dz j g gh. This opposition mainly functions in initial position, e.g. buu- ‘to descend’

vs. puu- ‘to tie up’, dere ‘pillow’ vs. tere ‘that’, gii- ‘to lose’ vs. kii ‘wind’, jiluu ‘rein’

vs. ciluu ‘stone’. The analysis proposed by Junast even suggests that most non-

initial aspirates have lost their aspiration (as in Mongghul). The strong stops k q

can, however, in medial position be preceded by a phase of preaspiration, while the

corresponding weak stops g gh can occur with a fricative pronunciation, which is probably

allophonic.

Speech sounds absent from the table include the following initial sequences, historically

produced by the loss of a vowel: h + n is realized as a voiceless n, e.g. hnii- ‘to laugh’ <

*hinie-; h + l is realized as voiceless

l or as a voiceless lateral fricative, e.g. hlaan ‘red’ <

*hulaan; h + y is realized as the ‘ich-laut’ [ç], e.g. hyaa- ‘to tie up’ < hiyaa- < *huya-.

WORD STRUCTURE

The syllable structure of Shira Yughur involves a regular alternation of vowels and con-

sonants, whereby the vowels can be short, long, or diphthongoid. The minimal syllable

consists of a single vowel (V), which can be preceded and/or followed by a consonant

(CV, VC, CVC). In initial position, clusters of two consonant segments (CCV, CCVC)

SHIRA YUGHUR 267

TABLE 13.2 SHIRA YUGHUR CONSONANTS

ptckq

bdjggh

ssh x

wzzh

mn ng

are also permitted. Not all consonants are allowed in all positions. For instance, strong

(aspirated) obstruents and fricatives other than s are not allowed in final position. In most

initial clusters, the first segment is a nasal or a fricative and the second an obstruent,

e.g. mba- ‘to swim’ (< *xumba-), skii ‘felt’ (< *isigei), hkü- ‘to die’ (< *xükü-). Other

clusters usually have variants which preserve the original vowel, e.g. cluu ~ ciluu ‘stone’,

cna- ~ cina- ‘to cook’, tyagh ~ tiyagh ‘cane’.

Accent is probably non-distinctive and falls normally on the final syllable of a word.

When suffixes are attached to a stem, the last syllable of the last suffix bears the accent.

However, long vowels can attract the accent away from the final syllable, and accent also

tends to flee the final syllable if it contains i, as in xuu

cin ‘old’, belwisin ‘widow’.

Most morphophonological phenomena at the boundary between a stem and a suffix

have to do with maintaining the ideal syllable structure of alternating vowels and conso-

nants (CVCV). This is achieved by means of connective segments, which might also be

termed buffer vowels and consonants. Examples are offered by the plural suffix /i-s and

the instrumental case ending /G-AAr. Some other suffixes assimilate the stem vowel, e.g.

xalda- ‘to look’ + causative .Ul > xaldu.ul-. Some suffixes have different consonantal vari-

ants depending on whether the stem ends in a sonorant or a (morphophonological) obstruent,

cf. e.g. the dative in -Di (-di ~ -ti) and the imperfective converb in -Ji (-ji ~ -ci).

Vowel harmony, which is conventionally assumed to have been based on an original

opposition between front vowels and back vowels, has considerably receded in Shira

Yughur. Many stems still reveal the effect of former harmonic constraints, but owing to

other developments, such as the palatalization of ‘back’ vowels by adjacent sounds, some

stems have acquired an ‘unharmonic’ look, e.g. jüra- ‘to mix’ < *juura-, neiman ‘eight’

< *naiman. Synchronically vowel harmony, including labial harmony, does not put any

restrictions on the vocalism of stems, but it continues to be active in some suffixes con-

taining the low vowel A, which is represented by the set a e o ö. In many cases, how-

ever, the harmonic variant ö is replaced by e, and some suffixes only feature the variants

a e, while yet others are completely invariable. Other deviations from the rules of vowel

harmony are also common in the data, cf. e.g. refl. cge-yaan ‘one’s father’ vs. com. refl.

xani-leen ‘with one’s friend’. A number of suffixes contain a harmonically variable long

UU (uu üü). The short high vowels of non-initial syllables have become harmonically

marginalized, since they have largely merged into i. Only a few suffixes, such as the

converb marker /G-AAdU, contain a harmonically variable short U.

There are also sandhi phenomena active at word boundaries. Some of these have only

a phonetic effect. For instance, initial and final g gh can be fricativized when occurring

intervocalically. In the case of b, however, a similar fricativization results in a phonemic

alternation with w. Also, some grammatical endings, notably -Gi of the futuritive

participle and -Ji of the imperfective converb, can lose their final vowel before a vowel

or a semivowel. In the data quoted below, these and other similar sandhi variants will be

indicated by the apostrophe (’).

WORD FORMATION

The Shira Yughur derivative suffixes are listed below under four basic categories. Within

each category, the suffixes are arranged approximately in an order of decreasing frequency.

Denominal nouns: [possessive adjectival nouns] .ti (< *.tU) ‘with, having’, e.g. argha

‘ruse’ : argha.ti ‘crafty’; .tii (< *.tAi) id., e.g.

jirghal ‘happiness’ : jirghal.tii ‘happy’; [the

268 THE MONGOLIC LANGUAGES

privative construction] .güi (< *+ügüi) ‘without, -less’, e.g. nere ‘name’: nere.güi ‘name-

less’; .ci [professions and occupations], e.g. cimig ‘theft’ : cimig.ci ‘thief’; .gA [affec-

tive], e.g. düü ‘younger brother’ : düü.ge id. [affective], .rKAG ‘abundant in’, e.g. qusu/n

‘water’ : qusu.rqagh ‘watery’.

Denominal verbs: A very frequent, and now perhaps the only productive, verbalizer

is the suffix .lA- [from substantival nouns], e.g. gilid ‘lock’ : gilid.le- ‘to lock’. Other

suffixes include: .dA- [from substantival nouns], e.g. shiwge ‘awl’ : shiwge.de- ‘to pierce

with an awl’; .d- [from adjectival nouns], e.g. olon ‘much’ : olo.d- ‘to become numer-

ous’; .ti- [from adjectival nouns], e.g. ghaltagh ‘filthy’ : ghaltagh.ti- ‘to become filthy’;

.rA- [mainly from adjectival nouns], e.g. hkö ‘blue’ : hkö.re- ‘to become blue’. Some

nouns are made into verbs by means of the element +gi- which is strictly speaking not a

suffix but the stem of the verb ‘to do’, e.g. lar ‘speech’ : lar + gi- ‘to speak’, cimig ‘theft’ :

cimig +gi- ‘to steal’. This compound construction is, however, much less common than

in some other languages of the Gansu-Qinghai complex.

Deverbal nouns: .mA [adjectives], e.g. sur- ‘to learn’ : sur.ma ‘experienced’; .mA

[instruments], e.g. algha- ‘to clean’ : algha.ma ‘mop’; .mA [professions], e.g. döngge- ‘to

support’ : döngge.me ‘midwife’; .mA [places], e.g.

uu- ‘to eat’ : xuula uu.ma ‘restaurant’;

.KAi [adjectives], e.g. hqara- ‘to break’ : hqara.qai ‘broken’; .rAKAi ‘apt to’, e.g.

honghi- ‘break wind’ : hongho.raqai ‘flatulent’; .Ur [instruments], e.g. solo- ‘to close’ :

solu.ur (< *solo.ur) ‘stop (of a bottle)’. Common Mongolic suffixes for deverbal nouns

that have become non-productive in Shira Yughur include *.xA (the imperfective partici-

ple), as in sana.a ‘thought’; *.xA/n (id.), as in tele.en ‘firewood’; *.xA.sU/n as in

xala.asin ‘patch’; *.cA as in gebte.ce ‘lair’, nuu.ca ‘secret’; *.dAl as in hkü.del ‘death’;

*.dU/n as in xanaa.din ‘cough’; *.gAn as in yaw.ghan ‘on foot’; *.l as in tani.l ‘acquain-

tance’; *.lAng as in jirgha.lang ‘joy’; *.lGA as in ere.lge ‘income’, *.mAl as in gür.mel

‘braid’; *.Un as in xalu.un ‘warm’; and .wir ~ .wur (< *.bUri) as in hele.wir ‘remainder’.

Deverbal verbs: There is no morphological means for creating repetitives and itera-

tives. Only the so-called genera verbi listed below are formed with suffixes, which

are: .Ul- [causative], e.g. hani- ‘to go’: hanu.ul- (< *hani.ul-) ‘to make go’; .lGA [causative],

e.g.

joqoi- ‘to sit down’: joqoi.lgho- ‘to cause to sit’; .GA- [causative], e.g. ghar- ‘to

come out’ : ghar.gha- ‘to bring out’; .GdA- ~ .GtA- [passive], e.g. eje- ‘to see’ : eje.gte-

‘to be seen’; .ldi [reciprocal], e.g. xög- ‘to hit’ : xög/i.ldi- ‘to hit each other, to fight’; .lcA

[cooperative], e.g. ghadi- ‘to harvest’ : ghadi.lca- ‘to harvest together’. Of the several

suffixes for the causative (transitive) category, only .Ul seems to be productive today.

Altogether, the number of productive derivative suffixes is not large. The two most

commonly used suffixes are the versatile nominalizer -mA and the general verbalizer -lA,

which are also used to adapt loanwords to the Shira Yughur inflexional system. Turkic

verbal stems are borrowed as they are, but Chinese and Tibetan verbs have to receive a

verbalizing suffix before they can take verbal suffixes, e.g. Chinese kai tuolaji ‘to drive

a tractor’ → Shira Yughur tuolaji kai.la- id., which can then be nominalized into tuolaji

kai.la.ma ‘tractor driver’.

Although there is no morphological difference between substantival and adjectival

nouns, there are some derivational phenomena that may be regarded as characteristic of

adjectival nouns. Thus, they can be intensified by means of partial reduplication, e.g. a.b

aruun ‘very clean’, and moderated by special (lexicalized and synchronically unpre-

dictable) suffixes, e.g. moder. hkö.btir ‘bluish’. The meaning of ‘rather’ can be expressed

by means of the suffix .shig, as in ene misgi muni qoqodi hdei.shig bai ‘this garment is

rather small for my body (size)’.

SHIRA YUGHUR 269