"Kissine offers a new theory of speech acts which is philosophically sophisticated and builds on work in cognitive science, formal semantics, and linguistic typology. This highly readable, brilliant essay is a major contribution to the field."
AUTHOR: Hans Nugteren TITLE: Mongolic Phonology and the Qinghai-Gansu Languages SERIES TITLE: LOT dissertation series PUBLISHER: Netherlands Graduate School of Linguistics / Landelijke – LOT YEAR: 2011
Benjamin Brosig, Department of Linguistics, Stockholm University
SUMMARY Beginning in the early 1980s, research into what has recently been termed the Shirongolic branch of Mongolic substantially increased. With grammars of Kangjia (Siqinchaoketu 1999) and Mangghuer (Slater 2003), reasonably detailed information on the phonologies and morphologies of all known Shirongolic languages has become available. This is complemented by dictionaries of most of these languages and even some dialects. On the other hand, most taxonomies and comparative treatments of Mongolic are based on materials produced earlier. Putting the known data into a larger context has been an important desideratum, and this is what “Mongolic phonology and the Qinghai-Gansu languages” (MPQGL) sets out to do for phonology.
The book is divided into four parts: the introduction (pp. 19-56) discusses the source material, presents the phone inventories of the languages under discussion and gives a taxonomic classification of South Mongolic (discussed below) where Baoan, Santa and Kangjia are assumed as a Baoanic group next to Monguoric consisting of Mongghul and Mangghuer. All these languages together contrast as the Shirongolic group with Shira Yugur. Next to phonetic criteria, lexical and morphological criteria are employed as well.
In the second chapter (pp. 57-84), the phonology of Proto-Mongolic is reconstructed. After short clarifications about the nature of vowel harmony, the fortis/lenis opposition in consonants and accent placement, more particular issues are addressed. These include the reconstruction of *ï, *o/*ö in non-first syllables and possible vowel sequences across and within syllables on the vowel side. As <g> is not reconstructed at all, this includes e.g. *iü in seriün ‘cool’ instead of *serixün/*serihün/*serigün where such a diphthong wouldn’t occur. On the consonant side, the placement of consonants within the syllable and their co-occurence restrictions are discussed next to the reconstruction of phonemes such as *š, *y and Pre-Proto-Mongolic *p.
The third (pp. 85-196) and fourth chapters (pp. 197-260) trace the development of reconstructed vowels (simple vowels, double vowels, complex vowels and lengthening) and consonants, respectively, in the modern languages. Finally, a Comparative Supplement (pp. 263-546) presents etymologies for 1350 words that survive in a number of modern Mongolic languages.
EVALUATION Getting started with MPQGL is a rather difficult task. For example, looking at how Proto-Mongolic *a developed in Shirongolic (94-102), we learn that *a is generally preserved unchanged in accented final syllables, while it often dropped in other syllables, unless the overall syllable structure was not conducive to elision or they were not close to any consonant prone to palatalise or labialise its environment. Final *a was also changed when it became non-final through the reflex of a formerly epenthetic *[u], e.g. gar- ‘exit’ [vs. gar(u)-gsan exit-PRF.PTCP] > Dongxiang qïri-. Moreover, some dialects of Baoan have recently started to replace a with ə even in final syllables (*sara > sarə ‘month’). Middle Mongol harmonic rounding of o - a to o - o took place, but a is retained after *oi in Dongxiang. Rounding is common close to labials and through the absorption of *b, while regressive assimilation occasionally occurs in Baoan if the following vowel was contracted (χutuŋ kuŋ ‘woman’ < *katun küün [‘queen’]). In a number of words (e.g. Eastern Yugur čüsa < *ǰasa- ‘make’), it is difficult to connect rounding to any cause. In contrast to Mongolian and Dagur, umlaut is very rare (e.g. Kangjia ǰeǰi- < ǰaǰil- ‘chew’), but a glide gets inserted in Dahejia Baoan after palatal consonants and Mongghul, in the latter even changing the quality of a to ä. Palatalization occurred under the influence of *y [j], *ǰ and *č. In cases like Ñantoq Baoan yimaŋ < *ïmaan ‘goat’, it is impossible to say whether palatal breaking took place due to the possibility of an intermediate form *yamaŋ. Loss of *a such as Shira Yugur ndaǧar < andagar ‘oath’ often results in consonant clusters, but most Qinghai-Gansu languages preserve *a here even though such clusters would be permissible. Elision of *a via devoicing may create fricatives, e.g. h in Shira Yugur (hsar < *asar ‘village’) and s, ś and ş in Mongghul, a process that may even lead to the loss of initial consonants. In contrast to previously discussed phenomena, Mittelsilbenschwund of *a is rather consistent and might be assumed for Proto-Shirongolic.
The nature of the data seems to be such that few absolute sound laws can be postulated even for any single South Mongolic language. However, to get an idea of what developments took place in any given language, a presentation ordered by languages might have been easier to follow. As it is, MPQGL is more suitable for reference or as a lexicon than for being read from beginning to end, as no general ideas emerge.
As one might expect, the apparently chaotic data makes it difficult to present a taxonomy of QG languages. Moreover, while all QG languages are in contact with each other and have strong areal features of the overall Gansu-Qinghai area, there are also a number of smaller Sprachbünde involved, the most obvious being the one formed by Turkic Shira Yugur and Mongolic Shira Yugur. Most of the changes discussed show broader areal distributions. For example, among the phonological “Features uniting all of QG Mongolic” (35-38) the only convincing innovation (shared only by Turkic, which is not areally dominant) is the assimilation of *k to g when followed by g or ǰ. While a split of the palatal affricates separates Shirongolic from Shira Yugur, there are indications that this change was quite recent (218) and thus took place separately in each language after the split of Shirongolic had already taken place. Baoanic differs from Monguoric in preserving the difference between /ki/ and /kï/ (as /ki/ vs. /qi/), but this is only the retention of a difference lost elsewhere and can thus hardly be used to postulate Baoanic in the first place. Overall, the phonological evidence is suggestive but not conclusive. Presumably for that reason, Nugteren tries to support his classification with lexical evidence.
MPQGL does its part to contribute to the terminological confusion concerning the oldest and reconstructed stages of Mongolic. While the misleading traditional term Middle Mongol(ian) is retained for the oldest attested stage of Mongolic, “Common Mongolic” is defined as “the reconstructed language from which all medieval and modern Mongolic languages are supposed to have evolved. It is obtained by comparing the modern languages and the Middle Mongol sources in various scripts”, whereas “[m]ore speculative reconstructions, generally involving developments that must have taken place in the prehistoric period, are labeled Proto Mongolic” (57). As Janhunen (forthcoming) uses “Common Mongolic” as a label for Buriat, Khamnigan, Mongolian, Ordos and Oirat as a group, the absence of any established and valid terminology will be felt even more direly. Given the current stage of research, most taxonomic labels must be preliminary, but a more descriptive stance might help. Even while “Common Mongolic” would be the language spoken in the late part of the 12th century and thereby just a few decades earlier than the first surviving texts (from the 1220s), this stage might better be labeled “Proto-Mongolic” and the preceding stage “Pre-Proto-Mongolic”. And the immediate subdivision of modern Mongolic might run “South Mongolic” (Shirongolic plus Shira Yugur), Dagur, Moghol (both one-language branches) and e.g. “Central Mongolic”.
But what do we learn about late 12th century Proto-Mongolic? The evidence from Baoan, Kangjia and Dongxiang mentioned above supports the distinction occasionally made in early Mongolian writing between <ki> and <qi>, thus lending support to the reconstruction of an eight-vowel system for Proto-Mongolic. Nugteren does not reconstruct any sound for the intervocalic letter <g> in Mongolian script, which (largely) lacks reflexes in modern Mongolian. He thus seems to leave aside a number of problematic cases such as *adugu > Mongolian aduu ‘horses’, *adugu-s (pl) > adguus ‘beasts’. On the basis of Shirongolic evidence, *h might be reconstructed for a few additional words with voiced medial consonant, e.g. *hamu(ra)- ‘rest’ in spite of ancient Sino-Mongolian amu- and Dagur amər- due to h/χ in all Shirongolic forms (254, 348-9). Other details of Nugteren’s reconstruction also tend to be very solid, reliable and in line with previous research. Occasional intrusive sounds are identified as such and excluded, pointing out where QG forms do not tell us much about Proto-Mongolic, and competing reconstructions are indicated as such in the complement, usually directly in the headwords, e.g. *čolban/*čolman ‘Venus’, or in the text according to which the second variant of Ordos čolmon/čulmun might also point to *u for the second syllable (307). (This implicitly presupposes that the reader knows that northern Ordos has progressive assimilation in such contexts, while southern Ordos has regressive assimilation.)
At different points in the book, the question of accent is touched upon. While Nugteren does not reconstruct any accent for Middle Mongol, he consistently assumes Mongolian, Buriat and Oirat to have word-initial accent, even though newer studies (Walker 1997 for Khalkha Mongolian and Buriat, Harnud 2003 for Chakhar Mongolian and probably also applicable to Khalkha, Indjieva 2009 for Xinjiang Oirat) show that matters are not that simple for any of these languages.
Spelling mistakes are rare, as are obvious factual mistakes or contradictions. The only puzzling statement I came across was the claim that “[t]he fact that this M[iddle ]Mo[ngol] -h- never corresponds to W[ritten ]M[ongol] -g/γ- shows that it is not inspired by WM spelling.” (78), which doesn’t apply to the examples that precede or follow the sentence which include WM <deger-e> ‘above’. As Nugteren (p.c., my translation) points out, the idea is rather that “in cases where WM <g> reliably goes back to *g [as reconstructed from modern languages], we never observe [MMo] -h-”. Consequently, he tentatively reconstructs Pre-Proto-Mongolic *depere (314) (MMo <deere>, <dehere>) irrespective of WM <deger-e>.
Summing up, MPQGL provides a reliable and long-awaited overview of Mongolic phonology that takes the South Mongolic data into account. It can serve as a reference and basis for further research both for Mongolists and general linguists. For the latter, occasional lines of argument that do require very specific background knowlegde may cause slight problems. The more than thousand well-supported reconstructions will come in handy for etymological purposes. It is not, however, easily usable as an introduction to QG phonology.
REFERENCES Harnud, Huhe. 2003. A Basic Study of Mongolian Prosody. Helsinki: University of Helsinki.
Indjieva, Elena. 2009. Oirat Tobi: Intonational structure of the Oirat language. Ph.D. dissertation, University of Hawaii.
ABOUT THE REVIEWER:
Benjamin Brosig is a doctoral candidate in general linguistics at Stockholm
University with a master’s degree in Mongolian studies and general
linguistics. His research interests range from aspectuality, evidentiality
and other morpho-syntactic categories over historical linguistics, dialect
grammar and field linguistics to politeness, focusing on Mongolic,
especially the Central Mongolic branch. Next to a dissertation on aspect
and evidentiality in Middle Mongol, Khalkha Mongolian and Khorchin
Mongolian, current research activities concern negation in Mongolic, the
semantic field of temperature in Khalkha, and a documentation of Durvud
Oirat (led by Yu. Tsendee).