The study also highlights the constructs of current linguistic theory, arguing for distinctive features and the notion 'onset' and against some of the claims of Optimality Theory and Usage-based accounts.
The importance of Henk Zeevat's new monograph cannot be overstated. [...] I recommend it to anyone who combines interests in language, logic, and computation [...]. David Beaver, University of Texas at Austin
The book under review represents a modified version of Brian O'Herin's 1995 University of California (Santa Cruz) Ph.D. dissertation. It concerns Abaza (Abaza Bëzsh*a), a Northwest Caucasian language, spoken by some 40.000 (other sources 31.000) people in the Karachai-Circassian Republic (Russian Federation), located at the northwestern slopes of the Great Caucasus mountain range (see Schulze 2002a for a recent presentation of the Abaza linguistic area). The main Abaza settlements are situated along the upper course of the Little and Great Zelenchuk rivers, as well as along the Laba and the Urup rivers. Here, Abaza speakers are to be found in thirteen villages, e.g. Abazakt, Tapanta, El'burgan, and Psysh. In addition, there are two Abaza villages near Kislovodsk and scattered settlements in the Kabardino-Balkarian Republic. Outside Russia, there are Abaza communities in Northern Turkey (near Amasya), as well as in Syria, Jordan, Egypt, and in the Balkans.
In spite of the fact that Abaza knows a written standard (see below), the language has to be described as endangered. According to the estimation of local speakers, Abaza is hardly ever used in school classes or among youngsters. Its use is mainly confined to the communication within the middle and older generation. Still, it must be added that recent sociolinguistic surveys draw a less pessimistic picture (see Schulze 2000a for some details).
Historically, the Abaza language had been spoken along the coast line of the Black Sea between what today is Tuapse in the North and the river Bzyb in the South. From this we can infer, that Abaza must have been in contact with the now extinct language Ubykh, historically spoken north of that area. Its earlier history is directly connected to that of its 'sister language' Abkhaz, see Schulze 2002b for a brief account. Accordingly, Abkhaz and Abaza form the southern branch of (North)West Caucasian, which again perhaps is related to a group of ancient northern Anatolian languages such as Hatti and Kashki (an early version of 'Circassian'?). Untenable is O'Herin's claim that "the potentially related languages include South Caucasian (...)" (p.6). Colarusso's suggestion to relate Northwest Caucasian to Indo-European is likewise difficult to follow (cf. Colarusso 1992).
Abaza speakers left their original homeland in the 13th-14th century and occupied their present locations devastated by Mongolian and Turkic raiders (1240). The two main dialects of Abaza (Tapanta and Ashkhar) seem to reflect the dialectal distribution given in the original homeland. Due to the supremacy of Kabardian (Eastern Circassian) groups (from the 17th century onwards), the Kabardian language started to influence especially the Tapanta variety. Due to the Tsarist efforts to russianize the region, more then four fifths of the original population left their homeland between 1862 and 1864. The remaining, by that time 9.000 Abaza were settled in their current locations.
In their historical homeland, the Abaza had been Christians. After their migration to the northern slopes of the Caucasus mountain range, they soon converted to Islam (Hanafiya).
Abaza is a written language (based on the Tapanta dialect). In 1923, the Abaza poet Talustan Talubov created an orthography based on Latin characters (discussed e.g. by Khashba 1931) that was replaced by a Cyrillic version in 1938. There is a considerable amount of literature available that is written in the Abaza language, both journalistic and poetic. Descriptions of Abaza do not start with Bouda (1940), as claimed by the author of the study under review p.1: Already in 1938, G. P. Serdjuchenko (the author of a small grammatical sketch of Abaza (Serdjuchenko 1956)) published an article on the dialects of Abaza (the earliest account seems to be Savinov 1850, a source, which reports (among others) on the Abaza language). As early as 1908, Abaza intellectuals in the Istanbul diaspora designed a written norm for Abaza, which, however, did not see success (nevertheless, the early history of Abaza grammar writing still remains an unstudied matter). In addition, we have to assume that from 1923-1929, several books on Abaza must have been prepared for school classes.
Just as it is true for its sister languages, Abaza is a strongly prefixing, agglutinating language, characterized by strong preferences for head marking strategies. Typologically speaking, Northwest Caucasian has much in common e.g. with the layout of Athapaskan grammatical systems. Contrary to Abkhaz, Abaza has extended the grammaticalization of 'pragmatic markers' as verbal suffixes. The unmarked word order is verb- final, preceded by a focal 'slot' as well as by referential segments marked for grammatical relations. As for these relations, Abaza follows a 'split ergativity' strategy (neutral/ergative, see below). Finally, Abaza, just as it is true for its sister languages, operated through a remarkable paradigm of phonemic variation that gives us for Abaza a system of roughly sixty consonants and two vowels (note that with respect to the number of phonemes, Abaza is rather moderate compared to e.g. Ubykh).
Morphological and syntactic complexity paired with a considerable amount of pragmatically relevant coding strategies render Abaza a typological treasure vault that still awaits a more comprehensive coverage. In this respect, O'Herin's book fills a significant gap: Hitherto, Abaza data have hardly been analyzed from the point of view of General Linguistics, in any framework whatsoever. O'Herin, who has undoubtedly managed to get deep into this language often (and falsely) denounced as an extremely difficult language, concentrates on the domains of Case and Agreement which constitute a major part of the Abaza 'relational' grammar. In this sense, the book promises to contribute not only to a better understanding of what is going on in a 'typical' Northwest Caucasian grammar, but also to the validation of these analytic domains themselves. On p. 1 of his book, O'Herin lists a number of references that are said to illustrate the small amount of linguistic work relevant for Abaza. His bibliography, listing some 110 titles, includes seventeen references that concern Abaza. Still, it should be noted that the author neglects a number of important sources, such as Genko 1954, Lomtatidze 1967, Lomtatidze & Klychev 1989; and Chirikba 1997. He likewise ignores the recent studies on Abaza, prepared by Iosif I. Gagiev (Gagiev 2000a, 2000b). Another relevant source neglected by O'Herin is given by a set of pedagogical grammars produced by Nur'ya T. Tabulova (e.g. Tabulova 1953, 1969, 1971). Also, he does not consider the vast literature available for Abkhaz, the sister language of Abaza, which shares many of the features discussed by O'Herin with Abaza. For instance, O'Herin remains in nearly complete silence as for the impressive work by George Hewitt, the grand-seigneur of Abkhaz studies.
Nevertheless, it goes without saying that O'Herin's analysis of Abaza that touches upon a very important theme crucial to both syntax theory and language typology is well-grounded with respect to the linguistic data exploited by the author. O'Herin has conducted a number of field trips to the Abaza communities assembling a vast collection of data. Unfortunately, the author does not tell us more explicitly of where and how he collected his data. The names referred to page xi-xii (when thanking his Abaza friends) suggest that the places of field work had been located both in Turkey and in Russia.
O'Herin's presentation of Abaza covers 286 (+ xvii) pages. It is divided into eight chapters, preceded by 'Acknowledgements' and a list of abbreviations, and followed by an appendix (summarizing constructional patterns of 'dynamic and stative predicates') and a list of references (see above). From the very beginning, the reader should be aware of the fact that O'Herin aims at an analysis of the Abaza data that is based on a formal framework, namely Principles and Parameters. This orientation explains what else would remain obscure at least for those readers acquainted with West Caucasian languages: It is standard knowledge that Abaza lacks case forms marked on referential (or nominal in its widest sense) forms. So, how can a book be entitled 'Case and Agreement in Abaza', if the language lacks what usually is associated with the term 'case'? This 'puzzle' is solved once the reader has adopted the formal Case Theory. Accordingly, "all Case assignment occurs in the specifier-head relationship within one of two types of agreement phrases, absolutive agreement phrases (...) and ergative agreement phrases (...)" (p.39). To illustrate this point, let me simply quote an example randomly taken from O'Herin's book (p. 59; here, I have retained O'Herin's glossing; 'sphas' (recte: s-ph/as) is not segmented by O'Herin):
sara wac*a s-ph/as (')al/ën lë-s-t-wëf-d I tomorrow 1s-woman ring 3sf-1s-give-FUT-DYN 'I will give my wife a ring tomorrow'
Neither sara 'I', nor s-ph/as 'my wife' or 'al/ën 'ring' are marked for case. But each case role (if we include the positionally defined zero-echo for 'al/ën) is reflected in the verb via agreement.
The fact that O'Herin adopts a formal framework to illustrate the basic morphosyntactic strategies of Abaza renders the book somewhat hermetic. Readers not interested in or not used to formal approaches to language structures have to single out passages relevant for their proper research interests. In the introductory section, O'Herin makes clear that one of his goals is to convince "those not familiar with formal theories that there is much to be gained from such theories in terms of understanding language". For those not used to the framework of Principles and Parameters, the author offers a (admittedly very) brief overview in section 1.2 of his book (pp.33-41). Maybe that this section is helpful for those who want to learn of how this formal framework accounts for the Abaza data. Also, specialists in this framework will find numerous arguments that help to refine some assertions of the framework. Still, the 'ordinary' user interested in typological variation together with its historical and pragmatic instantiations in Abaza, probably misses allusions to other explanatory paradigms which have turned out to be at least as powerful as formal theories to account for typological variation (e.g. Cognitive Typology, Cognitive Semantics, historical comparative linguistics etc.) Note that O'Herin's contribution is written from a nearly completely synchronic perspective, although it comes clear that quite a number of findings do not have synchronic motivation but stem from the habitualization of older communicative patters (see for instance Lomtatidze 1977 for a preliminary presentation of the linguistic history of Abaza).
Not being a specialist in formal approaches to language, I will refrain from presenting the individual analyses prepared by O'Herin. I leave it to such specialists to judge upon the appropriateness and correctness of O'Herin's analyses with respect to the underlying framework. Rather, I will simply summarize the main categorial, constructional, and morphological patterns elaborated by the author.
The introductory section starts with a rather condensed description of the grammar of Abaza. He briefly considers the phonological system, which, nevertheless is crucial to some aspects of Abaza morphosyntax. In fact, morphophonological features often help to decide which kind of functional properties we have to deal with. For instance, in Abkhaz, the sister language of Abaza, the first person singular prefix (s-) is assimilated to a voiced onset of the verbal stem (when immediately following the prefix) (> z-). This process, however, is confined to the 'agentive' role (horribile dictu: transitive subject). In case s- reflects a first person singular in objective function ('object'), this process does not apply. According to my consultants, the same holds for Abaza.
O'Herin then briefly considers the 'morphology and syntax' of Abaza, concentrating on postpositional phrases, nominal phrases, and verbal phrases. Each of these domains is further elaborated in the subsequent chapters. Here, it is sufficient to note that Abaza postpositions (echoing the feature 'person/class' of the 'object' of the postposition with the help of the set of possessive prefixes) in fact are not postpositions at all, but grammaticalized possessive structures, e.g. (I have changed the glosses to render them more explicit):
a-s'ëys a-/*ara (p.50) DEF-bird 3sg:nhum:A-nest 'the bird's nest'
The hypothesis that both patterns share the same underlying constructional pattern is standard knowledge in West Caucasian linguistics. Nevertheless, O'Herin "posit(s) the possessor in a specifier position within the nominal extended projection and not in a complement position" (p.51). In other words: The issue again touches upon the question whether one gives preference to syntactic 'principles' etc., or whether one addresses 'natural' constraints (i.e., constraints resulting from cognitive and communicative parameters).
Perhaps the most crucial aspect of Abaza morphosyntax is represented by its system of 'agreement' present with verbal structures. O'Herin nicely summarizes the basic facts, that is what he calls 'Ergative Case assignment' (p.49) and 'Absolutive agreement' (p.63). [Unfortunately, he does not use a parallel terminology for both strategies (2.2. vs. 2.3)]. In addition, he informs on the paradigmatically salient opposition of dynamic vs. stative, as well as on some relevant phenomena related to the cluster Tense-Aspect-Mood. After a brief presentation of the underlying framework (see above), O'Herin -- in chapter 2 -- turns to Basic Case Assignment (pp. 43-90). This "core chapter" (p.4) describes and analyses the above mentioned agreement patters of Abaza, as illustrated in the following examples (p.47, 55, 58, 63, 64, 237, glosses modified; DIR = Directive Preverb):
sh*ë-l-ba-wash-d 2pl:O-3sg:f:A-see-FUT-DYN 'She will see you (pl.).'
y-/a-sh*ë-l-t-wash-t' 3sg:nhum:O-DIR-2pl:IO-3sg:f:A-give-FUT-IND 'She will give it to you (pl).'
sh*-/*ë-y-d 2pl:S-run-PRES-DYN 'You (pl.) run.'
d-yë-c-lë-z-ca-t' 3sg:hum:S-3sg:m(:A)-COM-3sg:f(:A)-BEN-go-DYN 'S/he went with him for her.'
The assumption according to which we have to deal with two different sets of agreement morphemes (S(ubjective)= O(bjective) vs. A(gentive)) turns out to be highly problematic: In fact, the distinction between S=O (= absolutive) and A (=ergative) becomes evident mainly in the third person, but not in those morphemes that encode speech act participants, compare the following paradigm of 'personal' prefixes: S=O A(=IO)
It comes clear that the formal opposition ABS vs. ERG is present with the third person, but not with a second and a first person. Here, the positional arrangement becomes crucial, compare (field notes):
wë-s-k'ë-y-t' 2sg:m:O-1sg:a-love-PRES-IND 'I love you (sg, masc.)'
së-w-k'ë-y-t' 1sg:O-2sg:m:A-love-PRES-IND 'You (sg, masc.) love me.'
In other words: The alleged ergativity of Abaza is confined to mainly the third person, or the 'non-personne', to use a term coined by É. Benveniste. Thus Abaza conforms (at least synchronically) to the well- known person/agentivity hierarchy (the so-called Silverstein Hierarchy). From this it follows that positional parameters are more salient than the degree of formal distinction. In this, Abaza comes amazingly close to e.g. Athapaskan languages, compare Chiricahua (Pinnow 1988:37):
nishbéézh < *ni-sh-l/-béézh 2sg:O-1sg:A-CL-cook:IMPERF 'I cook you (sg.)'
In addition, it should be noted that in Abaza, ergative strategies are strongly coupled with anaphoric constructions. Speech Act Participants do not occur as overt pronouns except for emphasis, whereas any third person prefix cross-references a deictic or nominal segment. Crucially, a third person non-human referent is not cross-referenced on the verb in case it immediately precedes the verb (p.20), compare (p.20, glosses modified):
sara a-msh* s-ba-y-t' I DEF-bear 1sg:A-see-PRES-DYN 'I see the bear.'
sara a-msh* shashta yë-s-ba-y-t' I DEF-bear early 3sg:nhum:O-1sg:A-see-PRES-DYN 'I see the bear early.'
In sum, it comes clear that O'Herin's description of Abaza as an "ergative- absolutive language" (p.75) is difficult to support. Rather, we should speak of 'split ergativity' with respect to the domain 'Person'. In case the superordinated strategy of word order is taken as a decisive parameter, we should define 'ergativity' in terms of serialization parameters (e.g. Schulze 2000), for instance:
ABS ERG #S-V #O-A-V
Here, the diagnostic feature has to be defined as the left 'word border' to arrive at an ergative strategy (S=O vs. A). There are several additional arguments that support the claim according to which Abaza is 'configured' in terms of an ergative strategy. But none of them is directly connected to the category of Case.
Chapter 3 turns to 'Stative Predicates' (pp.91-124). Stative verbs are differently tense/mood-framed than dynamic verbs, an opposition that is well-documented for all West Caucasian languages. Most importantly, stative verbs may be intransitive and transitive. Note that the decision whether a verb is stative or dynamic does not necessarily depend from the actual semantics of the verb. On p.93, O'Herin points out that for instance the verb -c*ëmagh- 'hate' is stative, whereas its 'positive correlate', -bzëyba- 'love' is dynamic. He correctly refers to the original (i.e. historical) reading of the two verbs: 'love' is dynamic, because it originally meant 'to see well', and 'hate' is stative, because it originally meant 'be one's enemy'. This example sufficiently illustrates that actual syntactic frames are not necessarily motivated (and processed) on a synchronic level. In fact, all synchronically transitive stative verbs seem to result from the reanalysis or metaphorization of former intransitive constructions, which may involve not only verbs, but also nouns, adjectives, and so-called postpositions. An example for the non-verbal use of stative constructions is (p. 94, glosses are modified):
a-sara d-rë-c-p' DEF-sheep 3sg:hum:S-3pl(:A)-COM-STATIVE:PRES 'S/He is with the sheep.'
In fact, we have to deal with the grammaticalized version of an older copula construction (copula *wp' (non-Past), *-n (Past)). Again, if we start from this diachronic scenario, much of what O'Herin discusses for stative verbs in terms of formal grammar becomes immediately transparent. Nevertheless, O'Herin's analysis helps to better understand the dimension of stative constructions in Abaza, just because it offers important data hitherto less observed.
Chapter 4 addresses the question of Causatives (pp.125-165). Abaza is marked for highly productive strategies to morphologically mark causative constructions (prefix r-). The position of the causative morpheme (immediately before the root) suggests that we have to deal with a derivational strategy rather with an inflectional pattern. In fact, the serialization of 'agreement' prefixes corresponds to that of transitive structures, compare (p.127, glosses are modified):
d-a-r-q*ëc-i-t' 3sg:hum:O-3sg:nhum:A-think-PRES-DYN 'It makes him/her think'
With transitive verbs, the pattern is O - A' - A - CAUS - V (I use A' to indicate the fact that the embedded agent or causee is encoded with the agentive/ergative morphemes, compare (p.133, glosses again modified):
yë-l-së-r-sa-t' 3sg:O-3sg:f:A'-1sg:A-CAUS-cut:PAST-DYN 'I had her cut it.'
O'Herin aims at elaborating the verbal nature of the -r-Causative, which is said to account for the agreement patterns just described. From the point of view of Caucasian linguistics, such an assumption unnecessarily complicates the matter. It is a well-known pattern in some other languages to use instrumental/causal features to encode let-causation. In this sense, the sentence above would read: 'With/because-of me, she cuts it.' In other words: the segment -së-r- represents nothing but a heavily grammaticalized, postpositional structure that later became incorporated into the verbal frame. There is one phenomenon, which might go against this analysis: O'Herin (p.138-9) shows that with a 3pl embedded agent (usually r(ë)-), dissimilation occurs, e.g. (p.138):
y-d-yë-r-ba-t' 3sg:nhum:O-3pl:A'-3sg:m:A-CAUS-see:PAST-DYN 'He caused them to see it.' (= He showed it to them')
This type of dissimilation does not occur e.g. with incorporated postposition, compare (p.138):
y-r-a-r-h/*-t' 3sg:nhum:O-3pl(:A)-DIR-say-:PAST-DYN 'They said it to them.'
Nevertheless, it should be noted that the two patterns are not alike. The causative pattern mentioned above posits the causee in front of the causer, but not before the 'subject' of the postpositional phrase. Although assimilation and dissimilation may occasionally be motivated by a syntactic arrangement, it is rather unlikely that we have to deal with a synchronically 'transparent' type of dissimilation. This comes also true from the fact that the same 'process' can be observed in Abkhaz. Most likely, the dissimilation had already been fossilized in the time of the Abkhaz-Abaza unit that is roughly some 1500 years ago (if ever it had been a dissimilation at all).
After having discussed reflexive strategies, O'Herin turns to 'Derived Inversion' in chapter 5 (pp. 167-191). Here, the author considers derivational patterns that are marked for the 'inversion' of agreement patterns. This includes for instance the Potential, marked by a prefix -z (ë)-, compare (p.168):
y-së-z-lë-ta-t' 3sg:nhum:O-1sg:A-POT-3sg:IO-give:PAST-DYN 'I was able to give it to her.'
Here, the agentive morpheme has drifted further to the left, opening a 'slot' that may for instance be exploited by an IO prefix. Contrary to O'Herin's view, I cannot really see that inversion would be at work. The main point is that the IO-domain drifts to the right. The same holds for the Potential of Causatives, e.g. (p.189):
yë-s-zë-l-rë-f-t' 3sg:nhum:O-1sg:A-POT-3sg:f:A'-CAUS-eat:PAST-DYN 'I was able to make her eat it.'
The fact that the embedded agent undergoes the same shift as it occurs with the IO of a ditransitive verb (see above), gives us another clue for determining the nature of the embedded agent in causative constructions. Accordingly, we would have to deal with an IO (Indirect Object(ive)) rather than with an 'ergative' agreement marker (just as it is true for instance for the distantly related language Kabardian). As for the Potential, O'Herin suggests (in simplified, non-formal terms) that the marker -z/ë)- functions as some kind of 'capability auxiliary' followed by the lexical complex. Literally, the above mentioned example isëzlëtat' (y- së-z-lë-ta-t') would read: 'it I could her give' (instead of a non- inversed reading *'it her I would give' (*y-lë-s-zë-ta-t'), compare y-lë-s- ta-t' 'I gave it to her'). This auxiliary hypothesis, which in fact is said to hold for the causative, too, is rather attractive - however, up to now, it lacks clear historical evidence.
Chapter 6 deals with what O'Herin calls 'Lexically Inverted Verbs' (pp.193- 211). In this brief, nevertheless extremely interesting section, the author turns to a class of superficially transitive verbs that are marked for the inversion of the position of A and O functions. Such verbs (among them 'bite', 'touch', strike, hit', attack, forgive', 'help' and 'shoot at') are said to have Inherent Case just as it is assumed for e.g. German 'ich helfe dir (dative)' 'I help you'. AN example for the framing type in Abaza is (p.196, glosses again modified):
h/ë-y-g*ëk*s-t' 1pl:S-3sg:m:IO-attack-DYN 'We attack him (the enemy).'
It comes clear that, here, Abaza uses the set of 'absolutive' prefixes instead of the standard 'ergative' series to encode A. O'Herin correctly suggests that "inverted verbs in Abaza are parallel to the dative verbs of Russian and German" (p.196). From a functional point of view, the verb frame mentioned above represents nothing but fossilized antipassives, compare:
Ergative: O-A-V AP: S-IO-V (AP: A>S, O>IO)
Antipassives are well-known in related Circassian and Kabardian, compare for Kabardian (e.g. Colarusso 1992b:177). The IO-character of the 'former O' becomes immediately evident, if we have a look at the Potential. Here, again, the so-called Object prefix (i.e., the IO prefix) shifts to the right of the Potential prefix:
s-z-y-ësë-y-d 1sg:S-POT-3sg:m:IO-hit-PRES-DYN 'I can hit him.'
In chapter 7, O'Herin discusses 'Postposition Incorporation' (pp.213-248). The author carefully analyses the relevant data that are marked for the incorporation of a postpositional complex into the prefix chain (PP-S-V or O-PP-A-V). Semantically speaking, this strategy concerns Benefactives, Adversatives, Comitatives, Locatives, and Instrumentals. A simple example is (p.214):
y-l-zë-s-dz*-d 3sg:nhum-3sg:f(:A)-BEN-1sg:A-drink:PAST-DYN 'I drank it for her.'
This process, which is well-known e.g. from Athapaskan languages, seems to be linked to (among others) the parameter of definiteness: In case the 'subject' of a postposition is marked for (strong) indefiniteness, incorporation applies. This tendency goes together with the preference for (pro)nominals carrying strong reference not to be incorporated (p.224). In fact, the incorporation conditions nicely meet the basic typology set up by Mithun 1984.
Finally, O'Herin turns to Wh-Agreement (chapter 8, pp.249-276). The author observes: "When an argument is [+wh], the agreement with that argument is realized as wh-agreement. This places wh-agreement squarely within the normal agreement paradigm" (p.250). Crucially, wh-agreement is also present e.g. with relative clauses. This fact sets Abaza apart for instance from the East Caucasian language Udi which knows wh-agreement only for questions (see e.g. Harris 2002). The general Abaza wh-marker is z (ë)- (A=IO) and y(ë)- (S=O). Examples are (252, 252, 252, glosses modified; Q = 'wh-agreement marker'):
a-c*wal yac'*ëya yë-ta-wa DEF-sack what Q:S-be=in-PRES:STAT 'What is in the sack?'
dëzda s-axcja zë-ghëcj who 1sg(:A)-money Q:A-steal(:PAST) 'Who stole my money?'
ismir dzac*wëya yë-r-ba-k*a-z Izmir who Q:O-3pl:A-see-PL-PAST 'Whom did they see in Izmir?'
Just as it to be expected from the linguistics of the given area, Abaza prefers to place wh-words in the preverbal focus field. O'Herin nicely analyses this positional preference without, however, alluding to the fact that we have to deal with an areal phenomenon, common to many languages spoken in and around the Northern Caucasus. Unfortunately, the author does not ouch upon the question of how and why the special wh-agreement pattern has emerged. It is perhaps more than just a guess that both z(ë)- and y(ë)- represent residues of older wh-words (pace Nikolaev & Starostin 1994:492). Note that the wh-agreement prefixes do not distinguish degrees of animacy, whereas the overt wh-pronouns do (dëzda ~ dzac'*ëya 'who' vs. yac'*ëya 'what'). Relative clauses are clearly derived from wh-strategies. Relative clauses operate in the same way as participle-based relativization happens e.g. in Turkic languages, compare (p.260):
y-awë-y-shtë-z a-h/aq*-dëw Q:O-PV-3sg:A-throw-PAST:REL DEF-stone-big 'The big stone that he threw....'
Still, note that the relative segment occurs to the left of its head, whereas a (usually incorporated) attribute follows it.
Finally, O'Herin draws the reader's attention to a very interesting fact, namely there is an alternative reading of yac'*ëya 'what' > 'why'. The use of 'what' when asking for a reason is also know e.g. from German, e.g. 'was guckst du?' ('why do you look (at me)'). In Abaza, the use of the pronoun as a 'why'-marker is coupled with a special wh-agreement morpheme, compare (p.265):
yac'*ëya (...) sh*-zë-në-m-xa-wa what>why (...) 2pl:S-Q-PV-NEG-work-PRES:NEG 'Why don't you work [even harder]?'
The morpheme is z(ë)- and thus equals the standard Q:A. Unfortunately, O'Herin does not give us an explicit transitive construction (e.g. 'why do you kill the horse?'). Still, the examples given by the author suggest that the Q-marker in why-constructions actually plays the role of the agentive, 'demoting' the standard personal agreement prefix to the Objective. Hence, the example above would read: 'What makes you not to work [harder].' This analysis goes together with the fact that the 'why'- reading of the pronoun presupposes that it is placed clause-initially, that is in just the place that usually is occupied by an overt A-referent. In terms of cognitive linguistics, we have to deal with the metaphorization of 'what' as a 'reason-related agent'.
Unfortunately, O'Herin's book lacks a summary or a concluding chapter. Especially those readers who are unacquainted with the marvelous world of West Caucasian languages may have difficulties to arrive at a more general picture of the morphosyntax of Abaza. O'Herin has put much effort in giving a detailed account of what is actually going on in the language. The wealth of data (which often include new material) is coupled with a highly sophisticated analysis which sets the reader at risk to concentrate more on details than 'on the whole'. Still, it is my deepest conviction that without understanding the overall strategies and 'mechanisms' of a language (together with their communicative and historical settings), the analysis of particular phenomena may rest episodic.
The reader would perhaps have welcomed the illustration of Abaza with the help of a longer text, fully glossed and commented upon with the help of the analyses presented in the book. I am well aware of the fact that such a presentation would not be in the scope of the formal framework adopted by the author. Still, I assume without a closer look at the organization of textual data (in terms of 'context'), much of what O'Herin proposes in his highly sophisticated and undoubtedly learned analysis remains fragmentary. For instance, pragmatic strategies, the interaction of TAM- framing and clausal organization, variation in the degree of referentialization etc. only become apparent if textual embedding is considered. O'Herin surely has an impressive knowledge of Abaza, at least as far as the synchronic layer is concerned. His data are accurate, well- chosen and highly illustrative. Nevertheless, many questions remain open. In this sense, the book cannot serve as an introduction into the morphosyntax (and morphosemantics and morphopragmatics) of the language, nor does it replace what may be called the pragmasyntax of Abaza. The reader will certainly enjoy the scrutiny of the analyses, as well as the careful and balanced arguments put forward by the author in his analyses. However, as I have pointed out in the beginning of this review, the framework adopted by the author hinders him from approaching alternative explanatory perspectives. Here, it would perhaps have been wise if O'Herin had more frequently consulted grammatical and typological work on other (West) Caucasian languages, readily available on the market. This holds both for synchrony and diachrony. In fact, at least some of the phenomena explained by the author in terms of the Principles and Parameters framework, reflect older layers of the language, the functionality of which can today be only viewed in terms of 'habitualized routines' (or fossilized strategies). The decision to base his analysis on the 'formal paradigm' may help to bring further progress to this framework. But at the same time, the book becomes less useful for those who take a more functional perspective.
Nevertheless, it is a great pleasure to read the book (once one has adopted the formal framework). Even functionalists, 'business-as-usual' typologists, and cognitive linguistics will enjoy the impressive wealth of data that will undoubtedly contribute to the revision of some generalizations hitherto thought to be 'standard'. In addition, specialists in Caucasian linguistics are strongly motivated by O'Herin's data to take up the enterprise to unearth hitherto neglected categories and functional domains in other (West) Caucasian languages. In this sense, the book, which by itself is extremely well-done, must be welcomed. The only point the reader should be aware of is the fact that it does not (and probably cannot) tell the whole story. It is an important contribution to the morphosyntax of Abaza, but it is (hopefully) not designed to be a reference book of Abaza morphosyntax. At least the reader should not take it as such.
Bouda, Karl 1940. Das Abasinische, eine unbekannte abchasische Mundart. ZDMG 94,2.
Chirikba, V. A. 1997.: Abazinskij. In: Jazyki Rossijskoj Federacii i Sosednix Gosudarstv. Enciklopedija v trex tomax. I A-I,1-8. Moskva: Nauka
Colarusso, John 1992a. Phyletic links between proto-Indo-European and proto-Northwest Caucasian. In: H.I. Aronson (ed.). The Non-Slavic Languages of the USSR: Linguistic studies, 19-54. Chicago, CLS.
Colarusso, John 1992b. The Kabardian Language. Calgary: The University of Calgary Press.
Dixon, R. M. W. 1994. Ergativity. Cambridge: CUP.
Gagiev, Iosif I. 2000a. Prostoe predlozhenie v abazinskom jazyke. Mosvka: Nauka.
Gagiev, Iosif I. 2000b. Sopostavitel'nyj analiz prostogo predlozhenija v abazinskom i anglijskom jazykax. Moskva: Nauka.
Genko, A. N. 1954. Abazinskij jayzk. Grammaticheskij ocherk narechija Tapanta. Moskva: Izd. AN SSSR.
Harris, Alice 2002. Endoclitics and The Origins of Udi Morphosyntax. Oxford: OUP.
Ionova, S. Kh. & A. K. Shagirov: O sociolingvisticheskom sostojaniii, perspektivax i xode izuchenija abazinskogo jazyka. Rodnoy Jazyk 1 (Moskva).
Khashba, A. 1931. K voprosu o sozdanii pis'mennosti abazin Severnogo Kavkaza. In: Prosveshchenie nacional'nostej, 7-8.
Klychev, R. N. 1995. Slovar' sochetamosti preverbov s suffiksoidami i glagol'nymi korjami v abazinskom jazyke. Cherkessk.
Lomtatidze, K. 1977. Abxazuri da abazuri enebis ist'oriul-shedarebiti analizi. Tbilisi. Mecniereba.
Lomtatidze, K. & R. Klychev 1989. Abaza. In: G. Hewitt (ed.). The Indigenous Languages of the Caucasus, vol. 2.: The North West Caucasian Languages, 89-153. Delmar, NY: Caravan.
Mithun, Marianne 1984. The evolution of noun incorporation. Language 60:847-895.
Nikolaev, S. L. & S. A. Starostin 1994. A North Caucasian Etymological Dictionary. Moscow: Asterisk.
Pinnow, Jürgen 1988. Die Sprache der Chiricahua-Apachen mit Seitenblicken auf das Mescalero. Hamburg: Buske.
Savinov, V. 1850. Dostoverye ob Abazii. Panteon 2-12.
Serdjuchenko, G. P. 1938. On abazinskix dialektax. In: Pamjati akademika N. Ja. Marra (1864-1934). Moskva-Leningrad: Nauka.
Serdjuchenko, G. P. 1956. Kratkij grammatichestij ocherk abazinskogo jazyka. Kh. D. Zhirov & N. B. Ekba 1956. Russko-abazinskij slovar'. Moskva: GosIzd Inostrannyx i nacional'nyx slovarej.
Schulze, Wolfgang 2000. The Accusative Ergative Continuum. General Linguistics 37:71-155.
Schulze, Wolfgang 2002a. Abasinisch. In: Milos Okuka & Gerald Krenn (Hrsg.). Wieser Enzyklopädie des Europäischen Ostens, Band 10. Lexikon der Sprachen des europäischen Ostens, 837-841. Klagenfurt: Wieser Verlag.
Schulze, Wolfgang 2002b. Abchasisch. In: Milos Okuka & Gerald Krenn (Hrsg.). Wieser Enzyklopädie des Europäischen Ostens, Band 10. Lexikon der Sprachen des europäischen Ostens, 843-850. Klagenfurt: Wieser Verlag.
Tabulova, N. T. 1953. Grammatika abazinskogo jazyka. Chast' 2. Sinstaksis (dlja 6-7. kl.). Cherkessk.
Tabulova, N. T. 1969. Grammatika abazinskogo jazyka. Chast' 2. Sinstaksis (dlja 7-8. kl.). Cherkessk.
Tabulova, N. T. 1971. Abazinskij jazyk. Fonetika I morfologija (dlja 5-6 kl.). Cherkessk.
Tabulova, N. T. 1996. Dialektologicheskij slovar' abazinskogo jazyka (tapanta-shkaraua, shkaraua-tapanta). Cherkessk.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Wolfgang Schulze is the Head of the Institute for General Linguistics and Language Typology at the University of Munich. His main research topics include Language Typology, Cognitive Typology, Historical Linguistics, language contact, the languages of the Caucasus, of Inner Asia, and 'Oriental' languages. He currently works on a Functional Grammar of Udi, on the edition of the Caucasian Albanian Palimpsest from Mt. Sinai, and on a comprehensive presentation of the framework of a 'Grammar of Scenes and Scenarios' in terms of a 'Cognitive Typology'.