Janhunen J. The Mongolic Languages

Подождите немного. Документ загружается.

and Shira Yughur; (ii) the type *ene(e)n- in Dagur, Khamnigan Mongol, Buryat, Mongol

proper, Ordos, Oirat (with Kalmuck), Moghol (marginally), and Mongghul; (iii) the type

*enexün- > *enüün- in Buryat (archaic), Mongol proper, Ordos, and Oirat (with

Kalmuck); and (iv) the type *ene- (with no stem alternation) in Moghol, Mangghuer,

Bonan, and Santa. The original plurals of the type *ede (with or without an additional

plural suffix) are preserved in Dagur, Khamnigan Mongol, Buryat, Mongol proper,

Ordos, and Oirat. In Shira Yughur, Mongghul, Mangghuer, Bonan, and Santa, the plural

forms are based on the singular stems.

(54) Possessive suffixes

The third person possessive suffix *-ni and the reflexive marker *-xA/n are attested in all

Modern Mongolic languages. By contrast, the first and second person possessive suffixes,

based on the pronominal genitives, are taxonomically relevant, being present only in

Dagur, Khamnigan Mongol, Buryat, Mongol proper, Ordos, Oirat (with Kalmuck), Shira

Yughur, and Santa. A distinctive third person plural suffix is present only in Dagur. The

absence of the possessive suffixes in Mongghul, Mangghuer, and Bonan seems to be due

to a secondary areal innovation. In Moghol, the pronominal nominatives are used as

possessive suffixes, though analytic constructions are more common.

(55) Predicative personal endings

The use of the suffixed nominatives of the personal pronouns as predicative personal

endings is attested in Dagur, Khamnigan Mongol, Buryat, Oirat (with Kalmuck), and

Moghol. In Khamnigan Mongol, Buryat, and Moghol, the nominal plural marker *.d has

also been incorporated into the system of personal endings (3p. pl.), while Dagur uses the

secondary plural marker -sul. In Moghol, the imperative forms are also arranged into

a full personal paradigm.

(56) Voluntative

Apart from the basic unmarked imperative (*-Ø), the only imperative forms attested in

all Modern Mongolic languages are the voluntative (*-yA) and possibly the permissive

(*-g/V, though with some problems of documentation and interpretation). The expanded

voluntative marker *-yA-n is, however, present only in Ordos (1p. sg.) and Moghol (1p.

pl.), suggesting the possibility of a shared innovation.

(57) Optative

The simple and expanded optative markers *-sU resp. *-sU-xAi are both present in

Buryat and Oirat (with Kalmuck). Moghol has only the simple marker, while Mongol

proper and Shira Yughur have the expanded marker. The desiderative (*-xA-sU-xAi) is

present in Mongol proper and Ordos, and possibly Shira Yughur.

(58) Benedictive

The original singular benedictive marker *-gtUi is preserved only in Khamnigan Mongol

and Buryat, while the plural marker *-gtUn is present in Mongol proper, Oirat (with

380 THE MONGOLIC LANGUAGES

Kalmuck), and Moghol. The exceptional simplification of the cluster *gt > t in this

marker in both Moghol (2p. du. -tå : pl. -tu ~ -tuna) and Oirat (-tn) is likely to be due to

a common innovation.

(59) Other imperatives

The prescriptive (*-xArAi), precative (*-AA-), concessive (*-tUgai), and dubitative

(*-xUjAi ~ *-xUji/n) have a basically uniform distribution, covering Buryat, Mongol

proper, Ordos, and Oirat (with Kalmuck), as well as, probably, Khamnigan Mongol (not

fully documented). The concessive is also attested in Dagur, suggesting that it may repre-

sent a more ancient morphological innovation. The same may be true of the dubitative,

provided that it is represented by the indefinite imperative in Dagur (which is not certain).

The potential (*-mjA) seems to be restricted to Mongol proper and Oirat (with Kalmuck).

(60) Narrative and durative

The simple narrative in *-m is preserved in Moghol, Mongghul, and Bonan, while the

durative in *-nAi ~ *-nAm > *-nA is present in all Mongolic languages with the excep-

tion of Dagur. This distribution suggests that the durative is an innovation that has grad-

ually replaced the more original narrative as the basic indicative form of the present tense

range. Against this background, it appears likely that the formally ambiguous Dagur

present-tense form in -ng (: -n- : -m-) represents the narrative (*-m), rather than the sim-

ple durative (*-n), which is not reliably attested in finite use in any Mongolic language.

The only languages preserving a trace of the final nasal of the durative marker *-nAm are

Khamnigan Mongol, Moghol, and (marginally) Shira Yughur, suggesting that the devel-

opment *-nAm (~ *-nAi) > *-nA (~ *-nAA) is a common innovation of all the other

languages.

(61) Terminative

Although the terminative (*-bA ~ *-bAi) is formally attested in all Mongolic languages,

its temporal reference in Dagur (present-future) differs from that observed in all other

Mongolic languages (past). Since no common temporal function can be reconstructed, it

is likely that the attested functions are the result of two separate innovations (Dagur vs.

Common Mongolic), both of which were based on the original aspectual content

(perfective aspect) of the form.

(62) Confirmative

The confirmative (*-lUxA ~ *-lUxAi) is attested in all Mongolic languages with the

exception of Bonan and Santa. Its presence in Mongghul (in the composition of the ‘con-

cessive’ marker -lagi ~ -lahgi) remains, however, unconfirmed. Moreover, due to its

functional overlapping with the other forms of the past-tense range, the confirmative has

tended to be marginalized in several other languages, notably Dagur, Khamnigan

Mongol, and Buryat. In Shira Yughur it is only attested in the periphrastic construction

for the ‘impending future’.

INTRA-MONGOLIC TAXONOMY 381

(63) Resultative

The resultative (*-ji ~ *jixAi) is absent in Dagur, Buryat, Shira Yughur, and Santa. This

distribution suggests that this form may be due to a Common Mongolic innovation that

never comprised Dagur, while Buryat, Shira Yughur, and Santa may have lost it secon-

darily. The resultative seems to have originated as an inferential past tense (as still

synchronically in Khalkha), which explains its use as an objective form in the perspective

system of Mangghuer.

(64) Participles

Of the five Common Mongolic participles, only the futuritive, perfective, and agentive

participles are attested in all Mongolic languages. Of these, the agentive participle

(*-gci/n ~ *-xAci/n) is mainly attested in fully nominal functions, though verbal

functions are also possible especially in Dagur, Buryat, and Mongghul. The perfective

participle marker has two shapes: (i) the original shape *-gsAn with an initial cluster

(Written Mongol -qsav -gsav), preserved only in Moghol (-xsan) as well as, rudimenta-

rily, in Dagur (conv. abtemp. -rsAAr ~ -sAAr) and Shira Yughur (-GsAn ~ -sAn), and

(ii) the simplified shape *-sAn, present in all other languages. The imperfective participle

is absent as a verbal form in Dagur, Moghol, and the Gansu-Qinghai complex, though

traces of the derivational use of the imperfective participle marker (*.xA ~ *.xA/n) are

present in all Mongolic languages. The habitive participle (*-dAg) has a similar distrib-

ution except that it is also attested in Shira Yughur.

(65) Secondary finite forms

All Mongolic languages use participles (with or without a copula) in finite temporal-

aspectual functions, though this usage tends to be marginal in Moghol and the

Gansu-Qinghai complex. The futuritive participle (without a copula) yields a future tense

in Khamnigan Mongol, Buryat, Mongol proper, Ordos, Oirat (with Kalmuck), and (with

a copula) Shira Yughur. For the past tense, Khamnigan Mongol and Buryat use the

imperfective participle (without a copula), while Dagur, Mongol proper, Ordos, Oirat

(with Kalmuck), and Mongghul use the perfective participle in a similar function.

(66) Basic converbs

The elaboration of the synchronic converbial system varies greatly, ranging from a single

form in Moghol to close to twenty more or less actively used converbs, quasiconverbs,

and periphrastic converbial structures in Dagur, Buryat, and Shira Yughur. The only basic

converbial form attested in all Mongolic languages is the imperfective converb (*-ji).

The modal converb (*-n) is absent (i.e. lost) in Moghol and Mangghuer, and the perfec-

tive converb (*-xAd) in Moghol, Mangghuer, Bonan, and Santa. In Dagur, Shira Yughur,

and Mongghul, the perfective converb marker appears without the final consonant (*-xA >

*-AA), which makes it formally impossible to distinguish from the imperfective

participle (otherwise not attested in these languages).

(67) Conditional converb

The conditional converb is attested in three variants: (i) the primary form in *-xAsU,

preserved in Dagur, Khamnigan Mongol, and Buryat; (ii) the secondary form

382 THE MONGOLIC LANGUAGES

*-bAAsU < *-bA+a-xasu (Written Mongol -basu), attested in Ordos and Oirat; and

(iii) the ‘colloquial’ form in *-bAlA, attested in Buryat, Mongol proper, Ordos, and Oirat

(with Kalmuck). A distinct group is formed by the Gansu-Qinghai complex, in which the

conditional converb is expressed by the marker *-sA, which seems to reflect the primary

marker (i), but which may incorporate other influences (including external influence

from Turkic). The element *-sA > -sa is also present in Moghol, but here it functions as

the marker of the subjunctive (irreal) mood of the finite conjugation. An additional

unique conditional form in (-ngg-)UUn/i is present in Ordos.

(68) Concessive converb

The concessive converb is morphologically related to the conditional converb and

appears with two principal markers: (i) *-bAci(gn) (Written Mongol -bacu), attested in

Buryat, Mongol proper, Ordos, and Oirat (with Kalmuck); and (ii) *-sA+da, attested in

Shira Yughur, Mongghul, and Bonan. Additional language-specific markers are present

in Santa (-se-nu) and Dagur (-tgai-c/ig, based on the concessive form of the imperative

sphere).

(69) Other converbs

The only language lacking the terminative converb (*-tAlA ~ *-tAr) is Moghol. The final

converb marker *-ra (as attested in Middle Mongol) seems to be preserved in the

Gansu-Qinghai complex in the shape (*)-la. An important innovation also confined to the

Gansu-Qinghai complex is the zero-marked serial (sequential) converb (-Ø), attested at

least in Shira Yughur, Mangghuer, Bonan, and Santa.

(70) Quasiconverbs

Owing to the synchronic transparency of the forms, it is often difficult to establish the

degree of grammaticalization present in the various quasiconverbial constructions. The

most widespread form of this category is the abtemporal (quasi)converb in *-gsA-xAr,

which is absent only in Moghol and Santa (and possibly Khamnigan Mongol).

Mangghuer and Bonan are linked by a special innovative shape of this marker (> -ser),

which seems to confirm that the form is fully grammaticalized. The successive

(quasi)converb in *-kU-lAA ~ *-kU-lAAr/n is attested in Buryat, Mongol proper, Ordos,

Oirat (with Kalmuck), and Mongghul, while the contemporal (quasi)converb in *-msAAr

is present only in Buryat and Mongol proper, though the related form in *-mAgcV is also

attested in Ordos, Oirat, and Shira Yughur.

(71) The category of perspective

The morphological distinction between two perspectives (subjective vs. objective) has

emerged (under Tibetan influence) in Mongghul, Mangghuer, and Bonan, though the

manifestations of the feature differ in the three languages. In Mongghul and Bonan, per-

spective is mainly expressed by using different copulas (or copular suffixes), while in

Mangghuer it is more intimately integrated with the system of verbal temporal-aspectual

markers. As an incipient feature, the category of perspective is also present in Shira

Yughur (the past non-progressive forms of the finite indicative sphere).

INTRA-MONGOLIC TAXONOMY 383

(72) Medial copula

The Chinese medial copula shi has been borrowed (as it would seem, separately) into

Mangghuer and Santa. The feature is also present dialectally in (Gansu) Bonan.

(73) Negation

In Middle Mongol material, the distribution of the prepositional negative particles (*)ülü

and (*)ese depends mainly on the temporal range of the verbal form to be negated, with

(*)ülü being used for present tense forms and (*)ese for past tense forms. This tendency

is still observed in Moghol, Mangghuer, Bonan, and Santa. Both particles are also pre-

served in Dagur, which, however, uses them indifferently. Only *ülü is attested in Shira

Yughur and Mongghul, while only *ese is attested in Mongol proper, Ordos, and Oirat

(with Kalmuck). Khamnigan Mongol and Buryat seem to have lost both particles.

Postpositional negation of verbal forms with the negative noun *ügei is present in all

Mongolic languages with the exception of Moghol, Bonan, and Santa. Generally, *ügei

is only used to negate participles (and the modal converb), but in Khamnigan Mongol

and Buryat, as an apparently secondary innovation, it can also be attached to original

finite indicative forms.

(74) Prohibitive particle

The basic prohibitive particle *bü (> *bUU) negates imperative forms (with minor

language-specific restrictions) in Dagur, Khamnigan Mongol, Buryat, Mongol proper,

Ordos, (Written) Oirat, Moghol, Mongghul, Mangghuer, Bonan, and Santa. The particle

(conc.) *bütügei is attested in the same function in Mongol proper, Ordos, (Spoken) Oirat

(with Kalmuck), Shira Yughur, and Bonan.

LEXICAL FEATURES

Compared with phonological and morphological features, the differences that exist

between the Mongolic languages in the lexicon are much more difficult to systematize.

For one thing, owing to the prolonged influence of Written Mongol (and Written Oirat)

on several Mongolic languages, it is often impossible to tell whether a given lexical item

represents the original genetic heritage of a language or the secondary interference of the

written medium. Continuing interaction across language borders between closely related

and areally contiguous languages, such as Khamnigan Mongol, Buryat, Mongol proper,

Ordos, and Oirat, also tends to obscure lexical differences. Moreover, it is difficult to

quantify the semantic distinctions that lexical cognates in different languages may show.

Another problem is connected with phonological irregularities, which can double or

even triple the taxonomic relevance of a lexical item. In principle, each irregularity has

to be counted as a separate taxonomic feature. The numeral for ‘nine’, for instance, is

shared by all Mongolic languages with the exception of Moghol and Mangghuer, but the

shape of the item is (i) *yersü/n in Bonan, (ii) *yesü/n (> *yösü/n) in Khamnigan

Mongol, Buryat, and Santa, and (iii) *yisü/n (> *isü/n) in Dagur, Ordos, and Shira

Yughur. Mongol proper, Oirat, and Mongghul are divided between *yesü/n and *yisü/n.

This suggests two consecutive innovations: *yersü/n > *yesü/n and *yesü/n > *yisü/n,

each with a different areal coverage. It cannot be ruled out, however, that both innovations

took place several times, separately in different idioms.

384 THE MONGOLIC LANGUAGES

Phonological irregularities illustrate the fact that lexical archaisms are often preserved

on the margins of a language family. The word for ‘ice’, for instance, is attested as *mösü/n

in most Mongolic languages, but direct reflexes of Middle Mongol mölsü/n are preserved

in Khamnigan Mongol (mulihu/n) and Buryat (mülyhe/n), on the one hand, and Mongghul

(molsi), Bonan (milsung ~ mensu), and Santa (miensun), on the other, suggesting that the

development *mölsü/n > *mösü/n was due to a single innovation radiating from the centre

of the language family. It is, however, not rare to find more complicated types of distribu-

tion, as in the Common Mongolic item *nabci/n ‘leaf’, which is attested as *labci/n in

Dagur, Mongol proper (dialectally), Mongghul, Shira Yughur, Bonan, and Santa.

On the basis of a tentative survey of 452 lexical items (unpublished database of the

author), it seems that the Mongolic languages can be divided into six categories, depend-

ing on how large the proportion of Common Mongolic items in their vocabulary is. The

following calculations also comprise the dialectal level for some of the major Mongolic

languages (Mongol proper, Buryat, and Oirat):

(1) Below 50 per cent

The only language belonging to this category is Mangghuer, in which the proportion of

Common Mongolic vocabulary would seem to be as low as 39 per cent. Although the fig-

ure might turn out to be somewhat higher on the basis of a different corpus, it is obvious

that Mangghuer has suffered a massive loss of native vocabulary, making it, at least

lexically, a good candidate for a ‘mixed language’.

(2) 50–64 per cent

This category comprises, not surprisingly, two other languages of the Gansu-Qinghai

complex, Bonan (50 per cent) and Santa (56 per cent), as well as Moghol (52 per cent).

The figure for Bonan would be more variable if dialects (Qinghai vs. Gansu) and subdi-

alects were considered separately, but there is no doubt that both Bonan and Santa retain

a larger proportion of native vocabulary than Mangghuer.

(3) 65–84 per cent

This category comprises the two remaining languages of the Gansu-Qinghai complex,

Mongghul (72 per cent) and Shira Yughur (77 per cent), as well as Dagur (81 per cent).

Interestingly, Western Middle Mongol (81 per cent) would also seem to belong to this

group, suggesting an early dialectal difference against Eastern Middle Mongol.

(4) 85–89 per cent

This category comprises three groups of marginal dialects: Western Buryat (87 per cent),

Mongolian Oirat (87 per cent), and the Kharachin dialect of Mongol proper (89 per cent).

Historically, Eastern Middle Mongol (87 per cent) would also seem to fall in this

category, though the evaluation of the Middle Mongol data is complicated by the

insufficiency of the extant material. The status of Mongolian Oirat is likewise somewhat

problematic, though there are strong indications that it contains a smaller proportion of

native vocabulary than other Oirat dialects.

(5) 90–94 per cent

This category comprises the bulk of the Mongolic languages and major dialects, including

Sinkiang and Alashan Oirat (90 per cent), Khamnigan Mongol (91 per cent), New Bargut

INTRA-MONGOLIC TAXONOMY 385

(91 per cent), Kalmuck (92 per cent), Qinghai Oirat (92 per cent), Ordos (93 per cent),

and Eastern Buryat (94 per cent). Of the dialects of Mongol proper, Sönit (91 per cent),

Baarin (92 per cent), and Chakhar (92 per cent), would also seem to belong here.

(6) Above 95 per cent

This category comprises the rest of the dialects of Mongol proper, notably Khalkha and

Khorchin (as well as, apparently, Modern Written Mongol), in which the proportion of

native vocabulary in the sample is as high as 98 per cent. Lexically, at least, these are the

‘most Mongolic’ of all Mongolic idioms.

More specifically, several subgroups can be established. A strong lexical connection

exists, for instance, between Buryat and Khamnigan Mongol, as is evident from many

diagnostic words, phonological shapes, and semantic developments, e.g. *jon ‘people’

(attested only in Buryat and Khamnigan Mongol), *kalaakai ‘nettle’ vs. Common

Mongolic *kalakai, *samagan ‘wife’ vs. Common Mongolic ‘old woman’. In many

cases, Buryat and Khamnigan Mongol preserve lexical archaisms, lost in Mongol proper,

e.g. *kübexü/n ‘son’ > Khamnigan Mongol kubee/n = Buryat xübüu/n vs. Mongol proper

xüü (< *keü ‘child’). On the other hand, Buryat and Khamnigan Mongol stand with

Mongol proper against Dagur, which shows either archaisms or innovations of its own,

e.g. Dagur *kein ‘wind’ vs. Common Mongolic *kei ‘air’, *salki/n ‘wind’.

The lexical differences between the dialects of Mongol proper, Oirat, and Ordos show

two patterns. On the one hand, the proportion of Chinese loanwords increases towards the

east (Ordos and the Inner Mongolian dialects of Mongol proper), cf. e.g. Oirat and Khalkha

*dere > der vs. Ordos der ~ jintü vs. Kharachin jentu (Chinese zhentou) ‘pillow’. On the

other hand, the Turkic impact increases towards the west (Oirat), cf. e.g. Ordos and Mongol

proper *kalbaga ‘spoon’ vs. Oirat *kasig (recent borrowing from Turkic). It goes without

saying that there are also lexical transitions from Mongol proper towards Buryat (through

Tsongol and Sartul), Ordos (through Chakhar), and Oirat (through Khotogoit).

The lexical situation within the Gansu-Qinghai complex is most difficult to assess. An

apparently valid generalization is that there is a some coherence between Mongghul and

Mangghuer, on the one hand, and between Bonan and Santa, on the other, cf. e.g.

Mongghul and Mangghuer bagha- ‘to hit’ vs. (Gansu) Bonan eke- = Santa eghi- id.

Another pattern unites Mongghul, Mangghuer, Bonan, and Santa against Shira Yughur,

cf. e.g. Shira Yughur xoto (Common Mongolic *koto) ‘town’ vs. Mongghul badzar

(bazar) = Mangghuer badzer (bazer) = Bonan badzar = Santa badza (baza) id.

(< ‘bazaar’, via Turkic from Persian). However, the position of Shira Yughur is very intri-

cate, as it can also go together with the rest of the Gansu-Qinghai complex against all

other Mongolic languages, as in Common Mongolic *alima ‘apple’ > Gansu-Qinghai

(with Shira Yughur) ‘fruit [in general]’. Within the Gansu-Qinghai complex, Shira

Yughur would seem to be closest to Mongghul, but the diachronic background of this

connection remains to be clarified.

TENTATIVE CLASSIFICATION

In any classification of a group of related languages, the principal problem is how to

balance the taxonomic features against each other. Although it is often easy to identify

386 THE MONGOLIC LANGUAGES

INTRA-MONGOLIC TAXONOMY 387

features that are more primary and, consequently, taxonomically more relevant than

others, it is not immediately clear how these differences in relevance should be measured.

In the absence of any solution to this problem, it is often inevitable to treat all taxonomic

features as if they were equally significant (one feature = one point). Even this principle

is difficult to follow, however, for it is not always possible to determine whether a given

feature is actually present in a language or not. Problems arise, in particular, with

features that involve gradual tendencies, sporadic phenomena, or dialectal differences.

Under such circumstances, any linguistic taxonomy is bound to be to some extent

arbitrary, for the choice and interpretation of the taxonomic features is always based on

subjective judgement. Moreover, there are different ways to count taxonomic features.

The similarities among a group of related languages can often most conveniently be mea-

sured in terms of the number of innovations shared by them, while the distances between

languages are more clearly revealed by a taxonomy based on the number of innovations

separating them from each other. When measuring similarities, genetic and areal inno-

vations should be distinguished from each other and from cases of structural conver-

gence, but, again, any attempt to make such a distinction requires subjective judgement.

With these reservations, the taxonomic relationships within the Mongolic family are

surveyed here (Table 18.1) on the basis of the phonological and morphological (but not

lexical) features shared by any two or more Mongolic languages. Phenomena that are

present in only a single language are not considered, since they do not involve shared fea-

tures. Only innovations (not retentions) are counted, but, depending on the case, the loss

of a feature can also be regarded as an innovation (negative innovation). It should further

be noted that many of the phonological and morphological phenomena discussed above

(1–74) involve several innovations, each of which has to be counted separately (the actual

judgements made in the process of counting will not be elaborated here).

The languages included in the survey are: Dagur (D), Khamnigan Mongol (KM),

Eastern Buryat (EB), Khalkha (Kh), Ordos (Or), Oirat (Oi), Moghol (Mo), Shira Yughur

(SY), (Huzhu) Mongghul (HM), (Minhe) Mangghuer (MM), Bonan (B), and Santa (S).

In a more comprehensive survey, other dialects of Mongol proper and Buryat, notably

Western Buryat, would also have to be considered. The difference between Kalmuck and

(the rest of) Oirat is, however, negligible for taxonomic purposes. It may be recalled that

TABLE 18.1 INTRA-MONGOLIC TAXONOMIC RELATIONSHIPS

KM EB Kh Or Oi Mo SY HM MM B S

D121417111246109 96

KM 33 27 24 23 10 12 12 12 11 10

EB 42 32 35 11 17 15 14 13 12

Kh 43 45 11 20 18 17 16 13

Or 40 10 18 18 15 15 11

Oi 14 19 17 15 13 12

Mo 14 15 17 17 18

SY 25 23 20 16

HM 37 34 30

MM 36 33

B36

the political definitions of the Mongolic languages are often very different from the

linguistic realities, especially as far as the Gansu-Qinghai complex is concerned.

Because of the subjective element inherent in any such calculations, the actual figures

expressing the mutual similarities of the Mongolic languages should not be taken too lit-

erally. However, even if a considerable margin of error is permitted, the general picture

is fairly clear. Not unexpectedly, the survey confirms that the Mongolic family comprises

two relatively compact subgroups: Buryat–Khalkha–Ordos–Oirat (32 to 45 shared

innovations) and Mongghul–Mangghuer–Bonan–Santa (30 to 37 shared innovations).

The two most closely related Mongolic languages are Khalkha and Oirat (45 shared inno-

vations), a conclusion that would not be altered even if other dialects of Mongol proper,

as well as Kalmuck, were considered separately. Khalkha is also very close to Ordos (43

shared innovations) and Buryat (42 shared innovations), while Buryat and Ordos are

somewhat more distant (32 shared innovations).

In this taxonomy, Khamnigan Mongol turns out to stand relatively close to Buryat (33

shared innovations). This is due to the large number of morphological similarities

between the two languages, which to some extent obscure the special position of

Khamnigan Mongol as the phonologically least innovative (most archaic) Mongolic lan-

guage. To be accurate, it is not always clear whether the morphological similarities

between Buryat and Khamnigan Mongol are really innovations, for in some cases they

might also represent original features lost in all other Mongolic languages. On the other

hand, the relatively small number of similarities between Dagur and the Buryat–

Khalkha–Ordos–Oirat group (11 to 17 shared innovations) tends to obscure the fact that

Dagur is actually a rather innovative language, though many of its innovations are

language-specific.

Within the Mongghul–Mangghuer–Bonan–Santa group, the strongest bonds seem to

link Mongghul with Mangghuer (37 shared innovations), and Bonan with both

Mangghuer and Santa (36 shared innovations), suggesting an areal continuum of close-

ly related languages with only slight polarization between two subgroups:

Mongghul–Mangghuer and Bonan–Santa. The smallest (although still significant) num-

ber of similarities is present between Mongghul and Santa (30 shared innovations). It

may be noted that the relatively large number of similarities between Mongghul and

Bonan, on the one hand (34 shared innovations), and Mangghuer and Santa, on the other

(33 shared innovations), is partly due to the parallel impact of secondary non-Mongolic

influences (Tibetan on Mongghul and Bonan, and Chinese on Mangghuer and Santa).

Of the two remaining languages, Shira Yughur appears to have more or less equally

strong bonds with the Buryat–Khalkha–Ordos–Oirat group (17 to 20 shared innovations)

and the Mongghul–Mangghuer–Bonan–Santa group (16 to 25 shared innovations).

Moghol, on the other hand, shows fewer similarities with the former group (10 to 14

shared innovations) than with the latter (15 to 18 shared innovations). Again, these figures

should not be taken at face value, for the innovations shared by Moghol and the

Gansu-Qinghai complex involve almost solely cases in which original features have been

lost under the structural impact of neighbouring non-Mongolic languages. It may be noted

that the smallest number of similarities between any two languages in the Mongolic fam-

ily is shown by Moghol and Dagur (4 shared innovations), suggesting that these are the

two extremities of the family not only geographically, but also linguistically.

To correlate the taxonomic data more closely with the geographical realities, it appears

possible to establish the following six areal groups of Mongolic languages: (i) Northeastern

Mongolic (NE) = Dagur; (ii) Northern Mongolic (N) = Khamnigan Mongol–Buryat;

(iii) Central Mongolic (C) = Mongol proper–Ordos–Oirat; (iv) South-Central Mongolic

388 THE MONGOLIC LANGUAGES

(SC) = Shira Yughur; (v) Southeastern Mongolic (SE)= Mongghul–Mangghuer–Bonan–

Santa; and (vi) Southwestern Mongolic (SW) = Moghol. Finally the cycle closes, for this

taxonomy (NE–N–C–SC–SE–SW) brings us back to the scheme (N–C–S) proposed by

Rudnev already in 1911.

REFERENCES AND FURTHER READING

Beffa, Marie-Lise and Roberte Hamayon (1983) ‘Les langues mongoles’, Études mongoles 14:

121–69.

Bertagaev, T. A. (1968) ‘Mongol’skie yazyki’, in Mongol’skie, tunguso-man’chzhurskie i paleoazi-

atskie yazyki [= Yazyki narodov SSSR 5], Leningrad: Nauka, pp. 7–12.

Binnick, R. I. (1987) ‘On the Classification of the Mongolian Languages’, Central Asiatic Journal

31: 178–95.

Binnick, R. I. (1992) ‘Mongolian languages’, in William Bright (ed.) International Encyclopedia

of Linguistics 2, Oxford: Oxford University Press, pp. 434–8.

Chingeltei [Cavggaldai] (1989) ‘Muvgqhul Tuirul Galat uv Qaricaqhav u Tuqai’, vUibur Muvgqhul

uv Yagae Surqhaqhuli yiv vXrdam Sivczilagav u Satgul (4): 6–19.

Doerfer, Gerhard (1964) ‘Klassifikation und Verbreitung der mongolischen Sprachen’, Mongolistik

[= Handbuch der Orientalistik I: V, 2], pp. 35–50.

Hattori, Shirô (1959 [1993]) ‘A Glottochronological Study of the Mongol Languages’, reprinted in

Selected papers of Shirô Hattori, vol. 4: Studies in Altaic Languages, pp. 87–8, Tokyo: Sanseido.

Kara, G[yörgy] (1970) Chants d’un barde mongol [= Bibliotheca Orientalis Hungarica 12],

Budapest: Akadémiai Kiadó.

Ligeti, Lajos (1933) Rapport préliminaire d’un voyage d’exploration fait en Mongolie chinoise,

1928–1931, Budapest: La Société Ko´´rösi-Csoma & Otto Harrassowitz.

Luwsanwandan, Sh. (1959) ‘Mongol xel ayalguunï ucir’, Studia Mongolica 1: 1–16.

Nugteren, Hans (1997) ‘On the Classification of the “Peripheral Mongolic Languages”’, in Árpád

Berta (ed.) Historical and Linguistic Interaction Between Inner-Asia and Europe [= Studia

Uralo-Altaica 39], Szeged, pp. 207–16.

Poppe, Nicholas (1954) Grammar of Written Mongolian [= Porta Linguarum Orientalium, Neue

Serie 1], Wiesbaden: Otto Harrassowitz.

Poppe, Nicholas (1955) Introduction to Mongolian Comparative Studies [= Mémoires de la Société

Finno-Ougrienne 110], Helsinki.

Poppe, Nicholas [Nikolaus] (1960) Vergleichende Grammatik der altaischen Sprachen, Teil 1:

Vergleichende Lautlehre [= Porta Linguarum Orientalium, Neue Serie 4], Wiesbaden: Otto

Harrassowitz.

Poppe, Nicholas (1965) Introduction to Altaic Linguistics [= Ural-Altaische Bibliothek 14],

Wiesbaden: Otto Harrassowitz.

Poppe, Nicholas (1975) ‘Altaic Linguistics – An Overview’, Sciences of Language 6, Tokyo

Institute for Advanced Studies of Language, pp. 130–86.

Rudnev, A. [D.] (1908) ‘Opyt klassifikaciï mongolov po narechiyam’, in G. I. Ramstedt,

Sravnitel’naya fonetika mongol’skago pis’mennago yazyka i xalxa’sko-urginskago govora,

S.-Peterburg: Imperatorskaya Akademiya Nauk, pp. iv–vi.

Rudnev, A. D. (1911) Materialy po govoram Vostochnoi Mongoliï, S.-Peterburg: Fakul’tet

Vostochnyx yazykov Imperatorskago S.-Peterburgskago universiteta.

Sanzheev, G. D. (1952) ‘Mongol’skie yazyki i dialekty’, Uchenie zapiski instituta vostokovedeniya

4: 30–125.

Sanzheev, G. D. (1977) ‘Mongolian Languages’, Great Soviet Encyclopaedia 16: 495, New York:

Macmillan.

Sinor, D[enis] (1952) ‘Langues mongoles’, in Antoine Meillet and Marcel Cohen (eds) Les langues

du monde, Paris: Centre national de la recherche scientifique, pp. 369–84.

Svantesson, Jan-Olaf (2000) ‘Mongolic vowel shifts and the classification of the Mongolic

languages’, Altai Hakpo 10: 193–207.

INTRA-MONGOLIC TAXONOMY 389