Date: Tue, 19 Apr 2005 11:23:07 +0200 From: Olesya Khanina Subject: A Grammar of Moseten
AUTHOR: Sakel, Jeannette TITLE: A Grammar of Mosetén SERIES: Mouton Grammar Library 33 PUBLISHER: Mouton de Gruyter YEAR: 2004
Olesya Khanina, Moscow State University, Philological Faculty & Max Planck Institute for Evolutionary Anthropology, Department of Linguistics
This book is the first comprehensive description of a language of the Mosetenan family. The family consists of 3 languages forming a dialect continuum where adjacent languages are mutually comprehensible: Moseten of Covendo - Moseten of Santa Ana - Chimane. They are spoken in Bolivia (South America) by 600, 150-200 and 4000 people, respectively. Genetic affiliation of Mosetenan to other language families remains unclear, even though some hypotheses have been proposed in the literature (grouping with Chon-Ona and Tehuelche by Swadesh (1963); with Pano-Tacanan and further with Chona and Yuracare by Suarez (1969); including in Ge-Pano-Carib and then to Amerindian macrofamily by Greenberg (1987)). This state of affairs isn't actually surprising, as before the current grammar hardly any language data on the whole Mosetenan family was accessible to the linguistic community. That's why the grammar under review is not only a high-quality language description that surely will be actively used by typologists of all sorts, but also an invaluable point of reference for comparative studies of South American languages.
The Grammar of Moseten was written by Jeanette Sakel as her PhD thesis at University of Nijmegen and Max Planck Institute for Evolutionary Anthropology (Leipzig) and was published in the Mouton Grammar Library series without any substantial change. It describes mainly the Covendo dialect of Moseten spoken in the foothills of the Andes, in La Paz Department. The language data comes from the author's own fieldwork, but it includes also materials resulted from her close collaboration with a number of native speakers who recorded the language varieties which were beyond the reach of an outsider linguist.
The book contains 13 chapters, a list of abbreviations, some maps and Appendix with 3 Moseten texts, statistics of word order patterns' occurrence in written texts and a list of grammatical markers with meanings and relevant pages in the grammar. In addition to general references, there are 9 pages of bibliography concerned with all the aspects of Mosetenan culture. I will report one by one about all the chapters, highlighting the aspects of Moseten that may be of interest to typologists and theoreticians. I'll try to avoid, where possible, commenting on the author's strongs and weeks: it will follow in the critical evaluation.
Chapter 1, "Introduction" presents the language, its genetic affiliation, sociolinguistic situation, previous research and history and method of current research. It concludes with an informative 2-page synopsis of main structural characteristics of Moseten.
Chapter 2, "Phonology" introduces the reader to segmental and suprasegmental phonology and to the orthography of the language. Moseten of Covendo has 10 vowel phonemes (i, e, mid central unrounded 'shwa', o, a + their nasal counterparts), the length is phonemic only for the first three of them - for nasalized as well as for oral. There are 24 consonant phonemes, for some of them aspiration and palatalization is phonemic. The syllable structure is (C)V(C) with only the vowel being obligatory; no consonant clusters are possible, except 'm/n/r + glottal stop' at the end of the syllable. However, loans can violate this syllable structure. All affixes have nasal and oral variants and thus are subject to vowel nasal harmony governed by the root vowels. In addition, a restricted number of verbal roots (about 15) manifest vowel assimilation induced by certain cross-reference markers. Its nature being, unfortunately, not specified by the author, the place of the stress is the first syllable of the root with few exceptions. At last, a number of morphophonological changes occur at morpheme boundaries and all of them have nothing unexpected from typological point of view.
In 9-pages Chapter 3, "Morphological processes", Sakel gives the reader a general overview of morphosyntactic inventory of Moseten. The language makes an extensive use of suffixes, clitics and reduplication, while prefixes and infixes are very few. Verbs differ significantly from other parts of speech: "most of the verbal roots are bound morphemes that have to be followed by a verbal stem marker to be turned into an element to which inflectional markers can be added" (p.53).
Chapter 4, "The nominal system", describes nominal grammatical categories and derivation of nouns from nouns/verbs and explores noun phrase structure. Nouns are reported to have one of two genders that are inflected on other constituents of the NP, being as well often, but not always, represented in the cross-reference ending of the verb. While the masculine is less marked formally, the feminine surprisingly appears to be unmarked functionally: a group of people of different gender are referred to as 'feminine plural', the feminine form of benefactive is used to refer to any group of people (even exclusively men!), when their gender is not focused, and verbs taking object complement clauses receive feminine object marking. There are two numbers, the singular being formally unmarked and the plural marked by a clitic or by global/partial reduplication of the root. Moseten nouns don't inflect for case, but there are a number of 'case-resembling' clitics: local (adessive, inessive, "downriver", superessive), instrumental, comitative, associative, benefactive, 'only', 'former'. They are attached first to determiner, if no, to modifiers, if no, to head NP itself; in addition, they can appear as words on their own. At last, there are two possible noun-to-noun derivations: the augmentative, marked by a prefix, and the diminutive, marked by a lexically determined change of the quality/nasality of the root (stem?) vowels (the latter can be applied as well to other parts of speech, except verbs). The main feature of Moseten NP structure is that modifiers (adjectives, relative clauses, possessors) carry the so-called 'linker morpheme', i.e. a marker of nominal syntactic dependency, which appears after 'case' clitics. The elements of NP can be linearly split, e.g. by a verb, for focus purposes.
Chapter 5, "Pronouns and reference", is concerned with all types of reference maintenance. Personal pronouns can be used on their own or can be cliticized to a verb or to a possessed noun. Here, they show a high flexibility: cliticized to a verb, they can refer both to the subject and to the object, and cliticized to a possessed noun, they can agree in person, number and gender both with the possessor and the possessum. The other way to encode possession is to mark the personal pronoun with the 'linker morpheme' obtaining, thus, a possessive pronoun. There are two demonstrative pronouns - masculine and feminine - that can refer both to animates and inanimates (the same holds true for personal pronouns). Moseten has an extensive class of interrogative pronouns that, being themselves constructed on the base of a single root (with one or two exceptions), function at the same time as the first part of indefinite pronouns (the second is indefiniteness markers) and as negative quantifiers in the context of markers of negation. A couple of interrogative pronouns can also be used as relative and adverbial clause markers. Finally, apart from a number of reference tracking pronouns, Moseten has a nice 'proform' used as a filler in discourse when a speaker isn't sure about the content of the word, but knows it syntactic status: various derivational and inflectional markers can be then added to it in order to express the desired syntactic function.
Chapter 6, "Adjectives and adverbs", consists of a number of general remarks about these parts of speech: their different syntactic functions, their derivational patterns and comparison techniques. However, the existence of true separate classes of adjectives and adverbs and their differentiation from stative verbs and nouns isn't properly presented and lacks if not argumentation, at least clarity. There seem actually to be a true separate class of adjectives, but in the light of overwhelming typological discussion of adjectives that have been occurring the last decade (cf., among others, Bhat 1994, Dixon & Aikhenvald 2004, Whetzer 1992), a general linguist would expect the 2004 grammar to be more pronouncing on the matter.
Chapter 7, "Quantification", gives an account of the numeral system and quantifiers. The former forms "a decimal system, which may have arisen from a quinary system" (p.167).
Chapter 8, "The verbal system", and Chapter 9, "Voice", treat the Moseten verb. Remembering quite agglutinative nature of the language, it's possible to describe a typical morpheme structure of the verb, even though it appears to be rather complex. It consists of a (bound) root followed by a 'verbal stem marker', then eventually by associated motion and/or voice markers, then eventually by aspect markers and obligatory by cross- reference endings. In addition, the root can be sometimes preceded by causative and applicative prefixes. Unfortunately, the relative order of the suffixes of different categories is not discussed in the book, so for some cases - like associated motion and voice - it remained unclear (the reader might be able to find the answer in numerous examples all over the grammar, although the superficial check by the reviewer didn't succeed).
Verbal stem markers derive verbal stems from bound verbal roots and other parts of speech: the choice of one of six possible markers entails the level of transitivity and of subject participant control over the event; for a number of verbs, the choice procedure is claimed to be lexicalized. In the footnote (f.167, p.477), the author draws a parallel between these markers and 'light verbs' and 'verbal classifiers' of other languages, but doesn't explain what is the difference between them and Moseten stem markers, if any. Unbound roots are very few, about 15-20. The final of the stem marker further determines which allomorph of derivational/inflectional suffix will be used; this information is also lexicalized to a great extent (e.g. the stems ending in -ki can be either "consonantal" or "vowel" (p.202)).
There are six affixes of 'associated motion': four of them refer to the movement to/from a deictic center ('go away to do an action', 'come to do an action', 'to perform on the way there', 'to perform on the way here'), one is a distributive ('movement in several directions to perform an action' - cf. reduplication distributive 'action performed in several locations') and the last one is used only in combination with some of the deictic suffixes above ('interrupted movement').
Moseten voices are reported to be causative, applicative, middle, antipassive, passive and reflexive/reciprocal (expressed by a single suffix and differentiated only by meaning). Valency increasing causatives and applicatives stay apart from the others as they are encoded exclusively by prefixes and both by prefixes and suffixes, respectively. Even though they have several variants, none of them overlaps with valency decreasing affixes. The valency decreasing affixes, on the contrary, are expressed just by two suffixes that spread over all valency decreasing operations: -ti- is one of antipassives and reflexive/reciprocal, -ki- is another antipassive and the only middle, the passive is encoded by combination of any of two causative prefixes and -ti- suffix. Surprisingly enough, this coincidence of form doesn't bring the author to any generalizations: she doesn't seem to see any problem in such a mess of 'homonymic' affixes (e.g. -ti- 'antipassive' and -ti- 'reflexive'), on the one hand, and the expression of the same concept by totally different affixes (antipassive -ki- and -ti-), on the other hand. Thus, to sum it up for the review reader, there are actually only two valency decreasing suffixes (-ki-, -ti-) and the latter can co-occur with causative prefixes to encode also a type of valency decreasing operation.
Aspectual derivations are progressive (differentiating between transitive and intransitive variants), inceptive (expressed by two different suffixes), iterative (expressed by total reduplication, suffix or infix, all responsible for different meaning: 'fast repetition', 'repetition over a long period of time' and 'neuter repetition', respectively) and durative (expressed by partial reduplication). In addition to them, there are a number of analytic constructions, particles and clitics that are used to express aspectual meanings. They are treated partly in this Chapter (habitual analytic structures with auxiliary verbs 'be, sit' and 'know' and habitual clitic), partly in Chapter 11 (see below).
Speaking about cross-reference system, there are intransitive and transitive paradigms. In the former, there is agreement only in gender with the subject (i.e. A/S argument, as Moseten is an accusative language), 1st person plural inclusive subject being an exception: it has a special suffix, which doesn't show gender distinctions though. In the latter, there is agreement in person, gender and number of both subject and primary object, but all verbal forms 'only refer to a subset of these features' (p.185), i.e. there is no verbal form displaying the agreement in all three categories for both participants. This transitive cross- reference system is far from being straightforward, and no general logic was found by the author. Sakel simply lists possible suffixes resulting from all combinations of subject's and object's gender, person and number, leaving thus the discovery of overall strategies operating on them for further research.
There are special affixes for 2nd person imperative; commands to other persons are expressed by general cross-reference forms. Negative imperative doesn't have special form and is expressed by standard verbal negation of positive imperative. It's worth noting that reflexive verbs take slightly different imperative affixes which might probably turn out to be trivially derived by their etymology, though. At last, few verbs have lexical hortative forms whose etymology is opaque.
Chapter 10, "Negation", discusses negation strategies and related problems. Moseten has general negation marker that can be applied to the entire close, as well as to any of the constituents. It is also used as an answer to negative questions. There are also a negative existential marker and a negative possessive marker, even though in positive existential and possessive clauses no copulas are used at all (see below).
Chapter 11, "Modality and discourse markers", treats particles and clitics all having in common that they are a part of sentential semantics and they are not obligatory. Clitics are differentiated from particles on linear order criteria: the latter can appear everywhere in the clause, even in the very beginning, while the former are always attached to the right of their host (first element of the clause for sentential clitics). Moseten has two productive evidential particles ('hearsay' and 'sensory experience'), a number of modal particles and clitics (only two of them are treated as 'grammatical': irrealis and necessity, nine others encoding speaker's certainty), five emphasis markers, six 'referential discourse markers' (cf., 'only, just', 'again', 'also', etc.) and eight temporal/aspectual reference particles. The latter being facultative, Moseten temporal reference can thus always remain unspecified and the same clause can refer either to past, present or future.
Chapter 12, "Clause types", is concerned with different types of independent clauses of Moseten (verbal, non-verbal and interrogative) with special attention to their constituent order. The language is 'pro-drop' and full NPs appear only for introduction of new referents or for emphatic purposes. The least pragmatically marked word order is SV(O) in these cases. Non-verbal clauses usually don't contain any copula and the combined elements are just juxtaposed (the only exception are negative non- verbal clauses, see above). Interrogative clauses have a question particle that follows the interrogated element, the latter being always fronted, i.e. in a clause-initial position. There are also a number of focus constructions involving special focus particles.
Chapter 13, "Clause combinations", explores clause coordination and subordination in Moseten. Basically, clauses are coordinated in the same way as NPs are: either by juxtaposition, or by a special particle. The contrastive coordination can be marked by one of two particles (clitics?): emphasizing subject non-coreference 'but', and frustrative, used to mark a contrast within a clause as well. Relative (restrictive, non-restrictive and headless), adverbial and complement clauses have the same structure as independent clauses. Their subordinate function is either marked only by a suffix ('linker morpheme', see Chapter 4), particle or clitics in case of relative, complement and adverbial clauses, respectively, or isn't marked at all in the case of same-subject complement clauses. There are two minor non-finite clause types: nominalizations that are only reported to encode complement clauses of ditransitive matrix verbs like 'forbid' and 'beg' and participles used almost exclusively for description of successive actions of the same participant. The word order in subordinate clauses is more or less the same as in independent clauses.
The book provides a detailed and comprehensive description of Moseten. It is the result of a deep and careful investigation of the language that was completely unknown before this study. The grammar is really reader- friendly: the language is extremely clear (the only small problem is the confusing usage of the word 'marker' instead of 'suffix', 'particle' or 'clitic') and all the numerous examples are carefully glossed and translated, for many of them the pragmatic context is provided as well. It's evident that the author has a thorough knowledge of Moseten that goes far beyond general linguist's needs. Moreover, this study represents an exemplary collaboration between a linguist and speech community: e.g. the choice of orthography for this purely scientific study was ruled not by methodological considerations, but resulted from the discussions in the community induced by the researcher.
However, the main minus of the grammar is somehow connected with its pluses: the author seems to prefer stating linguistic facts as they are, instead of trying to find a better analysis and to provide plausible arguments for it. Still, I think it will be quite easy for typologists and other general linguists to find examples relevant if not to their theoretical discussions, but to their topic for sure. All main linguistic facts are carefully documented and if one is interested in presence or absence of a feature in Moseten, (s)he can find the necessary information in the book. Unfortunately, it's not always possible to go into further details, though sometimes it is.
Summing up, even though the analysis presented here is not probably as detailed and insightful as that found in some other descriptive grammars of the last decade, one can't simply disregard the fact that the basics of the whole language family were unknown before this book. It seems to be a matter of metaphysical reflections whether it is better to publish that precious information one has for the moment or to pursue never-ending analysis leaving the whole subject a mystery to the others.
At last, I'd like to make some comments on the structure of the grammar. It possesses a very detailed table of contents, making it in principle easy to find necessary information. Such clarity is particularly important, as some facts appear not in the place you would expect it:
1. Nominalization in treated in Chapter 4 "The nominal system", not in Chapter 8 "The verbal system", while verbalization (that is confusingly referred to as 'incorporation markers'), on the contrary, is presented in the latter, not in the former chapter; 2. The fact that each modifier of an NP is marked by the 'linker morpheme', not only one of them, is mentioned in the Chapter 6 "Adjectives and adverbs", not in Section 4.7. "Noun phrase structures"; 3. "Voice" is treated in a separate Chapter 9 and not as a part of Chapter 8 "The verbal system", even though it's encoded by the same type of verbal affixes and Chapter 8 is said to treat "inflectional and derivational verbal structures" (p.181); 4. Reflexive/Reciplocal are described in Section 8.1. "Verbal inflection", not in Chapter 9 "Voice", even though they are expressed by the same type of affix, as other valence-changing derivations; 5. Section 8.5. "Aspect" of Chapter 8 "The verbal system" is announced to treat only derivations on verbs, postponing all other ways of expressing aspect to further chapters; but 'habitual' treated there is encoded either by an analytic structure, or a by a clitic of the type presented in Chapter 11 "Modality and discourse markers"; 6. Frustrative that actually is a sentential discourse marker used not only in coordinated structures, is treated in Chapter 13 "Clause combinations", and not in Chapter 11 "Modality and discourse markers".
Moreover, the structure of the entire Chapter 8 "The verbal system" is quite confusing and doesn't facilitate understanding: the exposition goes from inflection to stem types, then to stem markers and back to derivational affixes. It seems to be logical to go in the reverse order: first to describe stem formation by stem markers from bound roots, then the types of stems resulting from the previous procedure, then the derivational affixes operated on these stems and at last inflectional categories.
And if the structural oddities can be generally overcome by the mentioned table of contents, a very strange choice of abbreviations turns out to cause some problems in reading the book. While in descriptive and typological linguistics abbreviation of almost all grammatical meanings are conventionalized to a great extent (cf. such an outcome of this convention as Leipzig Glossing Rules at http://www.eva.mpg.de/lingua/files/morpheme.html) and thus can be often understood without reference to the List of abbreviations, it's not true for this grammar. Examples include ITI = 'iterative', RE = 'reflexive', NO = nominalization, CS = stative causative, HA = 'habitual', etc.
It's worth noting that the quality of the editing is really high: for 504 pages, I haven't encountered any real misprint (can 'focuSSing'(p. 91) be considered to be a misprint?) and only one mistake in glossing (example 8:284 must have DK, not AN for -ki- morpheme (p.274)). The only fairly serious error is in the footnotes: in the text, the footnotes from 52 to 59 appear as 19-26, in the body of the notes it's the correct 52-59.
A Grammar of Moseten is by sure a very important contribution to the field of South American languages and general typology. There is no doubt that comparative studies in the area may now go much further and that many linguists will find it as valuable source of data on the whole range of phenomena.
Dixon, R. M. W. and Aikhenvald, A. Y. 2004. Adjective classes: A cross- linguistic typology. Oxford: Oxford University Press.
Bhat, D. N. S. 1994. The adjectival category: criteria for the differentiation and identification. Amsterdam: Benjamins. (Studies in Language Companion Series, 24)
Greenberg, J. H. 1987. Language in the Americas. Stanford: Stanford University Press.
Suarez, J. A. 1969. Moseten and Pano-Tacanan. Anthropological Linguistics 11(9), 255-266.
Swadesh, M. 1963. On aboriginal languages of Latin America (Acerca de languages aborigines de America Latina). Current Anthropology 4, 317-318.
Wetzer, H. 1992. "Nouny" and "verby" adjectivals: A typology of predicate adjectival constructions // Michel Kefer, Johan van der Auwera (eds.). Meaning and grammar: cross-linguistic perspectives. Berlin: Mouton de Gruyter, 223-262.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Olesya Khanina is a PhD student of Moscow State University, Philological Faculty, Department of Theoretical and Applied linguistics and a visiting scientist at Max Planck Institute for Evolutionary Anthropology (Leipzig), Department of Linguistics. She is currently working on a cross- linguistical study of desideratives, with attention both to its semantics and morphosyntax. Beside the desideratives, her research interests includes typology of argument structure (interaction between parameters of argument structure and actionality). She has an extensive field-work experience in a number of languages of Russian Federation (Tatar, Chuvash, Balkar (Turkic), Nenets (Uralic)).