1Minyanka is a Senufo language spoken in Southeastern Mali and recognized as one of the national languages of this state. The language is also known as Mamara. Despite its official status, the language lacked any detailed description so far, with the exception of several works dedicated to distinct grammatical aspects of the language, cf. Prost (1964) and Dombrowsky-Hahn (1999). Thus, the book under review is the first fundamental grammatical description of Minyanka and thus should be regarded as a very important contribution to Senufo and Gur linguistics.
2The grammar is divided into 14 chapters, each of which covers a separate subdomain of the grammar (phonology, class agreement, non-verbal predication, clause linkage, etc.). All the necessary components of a grammatical description can be found in the book. However, discourse organization (linkers, discourse markers) is not described. The grammar consistently employs IPA transcription and Leipzig Glossing Rules. The terminology used by the author meets all the standards of modern typology, and thus, the grammar is well adjusted to the needs of both typologists and specialists in West African languages. The author declares his commitment to functionalist approaches to language and thus, the categorization of phenomena subject to the study is based on functionalist theories and classifications.
3In what follows, I give a detailed overview of the grammar. Given the fact that it is written in French, this review aims to provide a wider access to the content of the grammar for non-francophone readers. The overview also highlights some typological features of Minyanka which are of particular interest for typology and West African language studies.
4Chapter 1 deals with the phonological system of Minyanka. 38 consonant sounds are distinguished, and the set of phonemes includes prenasalized, labialized, and palatalized vowels. The list of obstruent consonants is especially large and includes the sounds f, v, s, z, ʃ, ʒ, ʕ, h, and ɦ. For each sound, minimal pairs are given in order to establish their phonemic status. The vowel system of Minyanka is typical of the languages of the region and is represented by the sounds i, e, ɛ, a, u, o, and ɔ. The vowels have nasalized counterparts, with the exception of e and o. Long vowels are present, but are reported to be either combinations of short ones (and therefore are not phonemic) or be conditioned by morphosyntactic alternations.
5The tonal system of Minyanka comprises two phonological tones, unlike in other described Senufo varieties having three to four tones. This fact is accounted for by Bambara influence. Lexical tonal patterns include all possible combinations of high (H) and low (L) tones, and some complex combinations are attested as well – for instance, diminutives have LHLHL tonal melodies. The only general tonal rule is left-to-right tone propagation which is found in several different syntactic contexts; other tonal rules are specific to morphological and/or syntactic contexts. The syllable structure is typical of West African languages – closed syllables are prohibited, and consonant clusters are of limited use. Stem-initial consonant clusters Cy and Cw are only attested in a limited number of words but prenasalized consonants are quite common.
6Chapter 2 is dedicated to morphophonology. It begins with a discussion of the nasalizing prefix N-. This prefix can be represented as an abstract feature triggering consonant alternations on stem onsets and is attested in nominalizations: wò ‘pour out’ – ŋò ‘pour_out.nmlz’, lɛ̀ ‘be old’ – nɛ̀ ‘be_old.nmlz’. However, some cases where the output consonant is in fact not nasalized are also analyzed as instances of the nasalizing prefix: sè ‘be born’ – zè ‘be_born.nmlz’, fɛ̀ɛ̂ ‘run’ – vɛ̀ɛ ‘run.nmlz’. It does not seem clear why the underlying form in such cases is N-, although one can see that consonant voicing is a process conditioned by the same morphosyntactic context. A similar process is described for compounds where the second stem undergoes nasal assimilation: cá̰ŋì ‘day’ + pà ‘come’ → cá̰-mbà-ʕàyí <day-come-sg.def> ‘the day of arrival’. Other forms involving consonant alternations include participle forms with the prefix woN- (woN- + -fu ‘be hot’ → wòvúrì ‘hot’) and future forms, of which the latter are discussed at length separately. The future marker triggers consonant alternations of the same kind as described above but are restricted to intransitive constructions. In other words, the future auxiliary ní triggers nasalization on the following verb if the latter is intransitive (ní + dìî ‘eat’ → ní + nìî) but not if it is transitive. In the latter case, the future marker does not affect the right-adjacent direct object. The author convincingly argues that, contra Carlson (1994), this alternation regarded previously as the exponent of the future tense should be analyzed as a separate phonological process. The presence/absence of the alternation is justified by the fact that the syntactic relation between the auxiliary and the following direct object is not the same as the one between the auxiliary and the verb. However, the nature of the relation is not revealed. Similar alternations involving consonant nasalization are discussed for several other markers.
7Another important alternation discussed in the chapter is fusion between the subject pronoun and the following copula. The latter contains a semi-vowel w which is frequently elided when preceded by a pronoun and fuses with it: yè wà ‘you are’ → yà:.
8Chapter 3 discusses the properties of nominal suffixes marking number. Traditionally, these suffixes are treated as class markers in Gur linguistics, and the book under review follows the tradition. The system of number marking includes several classes whose labels reflect the phonological form of the corresponding pronoun, that is, the pronoun replacing nouns of a given class. Singular markers are wú, kí, and lí, and plural markers are pí, yí, and tí. Singular nouns of the same class can have different corresponding plural forms – that is, nouns with the singular suffix -wú may have plural counterparts ending in pi or yi, etc. Each pair of class suffixes is labeled as “gender” (genre), and each gender has segmentally different suffixes for definite and indefinite forms. If a noun triggers agreement on its modifiers, the agreement suffixes partially resemble the corresponding pronouns but do not copy them. Thus, the pronoun-agreement marker correspondences are as follows: wú – Ø, ki – kV, li – lV, pí – bVlV, yí – yV, tí – kVlV. It is noteworthy that Minyanka has frozen suffixes for count nouns that do not correspond to any regular pronoun. The frozen suffixes are -rV and -mV, and their pronominal counterparts are tí and kí/yí, respectively. At the same time, definite markers are identical to pronouns, the -rV and mV suffixes going back to -ri and mi, respectively. The chapter gives lists of words belonging to each class and allomorphs of each suffix.
9The agreement system in Minyanka is quite rich, and the elements having class features and/or agreement markers include determiners, relativizers, demonstratives, some pronouns (‘other’, reflexives, anaphoric elements) and adjectives, the modifier ‘(the) rest of’, as well as identificational markers which are discussed in several sections. Class pronouns have special emphatic forms (wû ‘class W pronoun’ – wéré ̀ ‘emphatic’). The author also reports cases of semantic agreement, where nouns belonging to certain noun classes are replaced by pronouns of different classes, usually by those of the class K.
10Nominal derivational morphology is represented by several suffixes including augmentatives and pejoratives, the associative plural marker (-yé), the caritive suffix (-bàlì ‘without X’), the nominalizer -má, the prefix deriving instruments (yá-), and the prefix ká- with the meaning ‘motive of X’ (ká- + pà ‘come’ → ká-pání ‘reason for coming’). As is typical for the languages of the area, the augmentative and pejorative meanings are associated with particular class markers. For example, the noun há-rì ‘meat’ belonging to the T-class has an augmentative counterpart há-yì ‘a huge amount of meat’. Interestingly, one of the class suffixes -li, when used as a diminutive, it is stacked over other class suffixes (nà̰ɦɔ̀ ‘man’ – nà̰ɦɔ̀-rɔ́ ‘man-sg.indf’– nà̰ɦɔ̀-rɔ́-lì ‘a little man’), suggesting that its morphological distribution is different from that of regular class markers. Nominal prefixes include the elements yá- (going back to the noun yárá ‘thing’) and ká- of unknown origin. In the section on compounding, we learn about an unusual feature of the compound system, namely that some lexemes contain reduplicated verbal stems: lòfà-nì-bṵ̀-bṵ́-mbè-lì <donkey-sg.indf-hit-hit-stick-sg.def> ‘a stick to beat donkeys’.
11Chapter 4 discusses nominal modifiers that do not exhibit class agreement, namely numerals and quantifiers. Minyanka has a quinary system for numerals below 10, a decimal one for numerals below 20, and a vigesimal one for higher numerals. Numerals above 30 are frequently replaced by Bambara loanwords. The numeral ‘5’ is etymologically related to the word ‘hand’, and the numeral ‘15’ is reported to resemble the verb ‘fold’. Numerals, like nouns, have definite and indefinite forms and agree with the head noun in definiteness. Ordinal numerals, as was mentioned before, are formed by adding the possessor relator: pí-kɛ́ tá̰á̰rí mò-kî <house-sg.indef three poss-cl> ‘third house’.
12Another type of non-agreeing dependents is represented by quantifiers, one of which (bɛ́ɛ ‘also’) is a Bambara loanword. Distributivity is expressed by reduplication with the element wó intervening between the reduplicants: ʃɛ̰́ wó ʃɛ̰́ ‘each person’. Other non-agreeing modifiers include ɲjɛ̀ɛ̀rì ‘many’, bɛ̀rɛ̂ ‘sufficient’, dàmádàmà ‘some’, yɔ̀rɔ̂ ‘even’, wónɔ̂ ‘the same’, bɛ́ ‘also’, and yɛ́ ‘only’, many of which are Bambara loanwords. A separate class of non-autonomous adjectives is distinguished, which are inserted between the noun and the class suffix: ŋgò-fyɛ̰̀-ɔ̀ <chicken-white-sg.def> ‘the white chicken’. These are opposed to so-called “empathic” adjectives formed with the prefix woN-: nù wò-ŋô <cow adj-be.black.sg.def> ‘the cow that is black’. Other classes of qualifying items include verbs incorporated into nouns in the same way as non-autonomous adjectives: ʃùkárí-mò-kî <sugar-poss-cl> ‘sweet’.
13Genitive constructions are subdivided into those involving possessive relators and those involving direct juxtaposition of the possessor and the possessee. The former construction is used in cases where the possessor is focalized.
14Chapter 5 deals with adverbial constituents, namely, postpositional phrases, adverbs and ideophones. Some temporal expressions that can be used as NPs and thus have nominal distribution are regrouped under the term “quasi-nominals”. True adverbs, most of which belong to the lexical class of manner adverbials, can appear both to the left and to the right of the clause, even if the latter position is preferred by most items. The large class of adpositions is subdivided into prepositions and postpositions. Interestingly, most prepositions are Bambara loanwords (fɔ́ ‘till’, kàbî ‘since’, yánì ... cɛ́nì ‘before’), which is a common feature of many languages of the region generally having postpositions but employing Bambara prepositions. One of the most intriguing features of the Minyanka system is the presence of the circumfixed adposition ní ... ní, expressing the comitative/ instrumental meaning. Postpositions are predictably items going back to nouns; it is noteworthy that postpositions are morphologically indefinite forms of their nominal counterparts.
15Chapter 6 discusses non-verbal predication. Identificational, locative, existential, equative, simulative, and presentative constructions do not form a homogenous set, since some of them are analyzed as clauses involving pronouns as predicates and others as clauses headed by verbs. The discussion of the second type is included into Chapter 7 which deals with verbal clauses. The description of the TAM system is preceded by a description of verb tonal classes. A separate class of predicates is represented by verbal idioms (locutions verbales) that should be regarded as nominal-based complex predicates. Those consist of nouns integrated into a verbal root, cf. (1), forming non-compositional expressions. Interestingly, these sequences can be interrupted by grammatical markers, cf. (2)
|
(1)
|
wú
|
bá
|
bɔ́ɔrɔ̂
|
ɲù-cô
|
|
3sg.w
|
cond
|
work.sg.def.w
|
mouth-take
|
|
‘If he begins working ...’ (p. 159)
|
|
(2)
|
bɔ́ɔrɔ̂
|
ɲù
|
bá
|
ɲjò
|
|
work.sg.def.w
|
mouth
|
cond
|
take
|
|
‘If the work begins ...’ (p. 160)
|
16The TAM system of Minyanka is characterized by complex interactions between perfective/imperfective forms, which are basic forms, and post-subject predicative markers following the subject (future, subjunctive, habitual, perfect, etc.) accompanied in some cases by ‘be’-auxiliaries. The full clausal template consists of the sequence subject – (‘be’) – other post-subject markers – V, where V is the basic verb form. The most complex verb form is found in the future tense, where the ‘be’-auxiliary (wá) can be followed by a sequence of function morphemes. Thus, the sequence expressing imminent future consists of four elements: subject – wá sá níN ̀ ndà V. The morphologically most complex form is the imperfective, which has numerous exponents (-ni, -li, -ri, -ki, -ŋi), e.g. fwɔ́ – fwólí ‘fry’, yéré – yéení ‘stop’, as well as irregular alternations, e.g. jò – jú ‘say, speak’, dúkú – dúrú ‘ascend’, dìî – dí ‘eat’. Imperfective suffixes are reported to resemble class markers, and the author puts forward a hypothesis whereby imperfective forms go back to nominalized verbs. Given the fact that this diachronic scenario is widespread both typologically and areally, the hypothesis seems quite plausible.
17Negation in Minyanka is double-marked. Its exponents are the clause-final marker mɛ̰́, accompanied by morphophonological or/and tonal alternations on post-subject markers. For instance, the past marker bé (positive) alternates with bêe (negative). In some cases, post-subject markers are replaced in negative clauses: the positive subjunctive ʃíN ́ has a negative counterpart kàN ́. The negative elements go back to the proto-Gur negative auxiliary, whereas tonal alternations seem to be reflexes of negation markers that have fused with post-subject markers.
18The TAM system is represented by a large set of post-subject markers including the perfect expressed by the copula wá (going back to the verb ‘be’), conditional báN ̀, habitual màN ́, perfective yá, remote past mà, and subjunctive ʃíN ́. The progressive form is complex and composed of the imperfective copula wà followed by the progressive ná; however, verbs can be inserted between the two elements, as shown in (3).
|
(3)
|
wú
|
wá
|
ʃìnɛ̰́
|
ná
|
ŋɔ̰́ɔ̰́ní
|
|
3sg.w
|
prf.aff
|
lie.down
|
prog
|
sleep.ipfv
|
|
‘He lay down and is sleeping.’ (p. 194)
|
19The semantic properties of the markers that are described successively in separate chapters are in general typologically predictable. Minyanka habituals exhibit an unusual property – they are incompatible with negation. Another interesting property of the verbal inflectional system is the presence of two imperatives, one of which is labeled as “progressive imperative”. Imperative forms of this type consist of the progressive marker tà and the imperfective verbal stem. Semantically, progressive imperatives are forms used to cause the addressee to carry on the action denoted by a lexical verb or to begin an action. Negative imperatives (= prohibitives) constitute a separate subparadigm within the negative conjugation – whereas most negative forms involve copulas/ post-subject markers accompanied by the clause-final negative marker, prohibitives use post-subject elements kàN ́ or its allomorph kàmbáN ́.
20Subjunctives expressed by the element (ʃíN ́) are defined as subordinate to other verbal forms. Like in many languages of the West African zone, subjunctives are used not only as markers introducing sentential arguments but also as elements marking sequences of events; in the latter case, they are labeled “consecutives”. A special construction expressing wishes involves subjunctive forms heading clauses where the subject is expressed by the distributive form of the noun ‘God’ (kílɛ́ wó kílɛ́). Subjunctives pattern with imperative forms in using the same markers of negation.
21A separate section is dedicated to evidentiality which is expressed by the quotative marker ná. This grammatical marker appears clause-initially, preceding the subject.
22Chapter 8 is dedicated to argument structure alternations. These are represented by 1) alternations demoting arguments to obliques and 2) verbal lability. Obliques can be marked by one of several postpositions (ní ‘in’, ná ‘on’, má ‘benefactive’), and with certain verbal items, the alternation of postpositions yields different meanings. For instance, the verb ʃìkè is translated as ‘wait’ when combined with a zero-marked direct object but as ‘deprive’ when used with an oblique marked by the postposition ná. Lability is the defining feature of the verbal system, and most verbs can be used both in transitive and intransitive configurations. If a verb is used intransitively, it can have passive and anticausative interpretations, which is typical of West African languages. Interestingly, the overt expression of an agent in passive constructions is possible in Minyanka (in which case it is marked by the benefactive má), but the agent is not interpreted as a volitional entity. Reflexive and reciprocal meanings are not expressed by labile verbs; instead, they are encoded by the suffix -yɛ (attaching to pronouns) that goes back to the determiner ‘only’. The system of causatives comprises two suffixes, -kV and -ŋV̰, and most of the causativized verbs are statives.
23Chapter 9 describes serial constructions, which are classified on the basis of semantic criteria and subdivided into asymmetric and symmetric constructions. The former include manner and causative constructions. Symmetric constructions are further subdivided into benefactive and consecutive ones, and a special type of so-called cause-effect constructions is identified as well. These are defined as constructions where V2 expresses the result of the action denoted by V1 (‘beat X die’ = ‘beat X to death’). The description of semantic types is followed by a list of grammaticalized verbs, which are mostly V1 elements: pà ‘come’ is able to combine with several copulas yielding different TAM values, tàʕà ‘put’ contributes an instrumental meaning, wɛ́rí ‘be hot’ functions as an equivalent to the adverbial ‘early’, and hɔ̌ɔrì ‘return’ expresses repetition (‘again’). However, it follows from the data that these constructions represent several different structural schemas, since some of the serial verbs (= V1) project their own arguments. Thus, the intransitive verb pà ‘come’ follows the post-subject marker and is followed by an OV sequence, yielding the structural schema in (4). In contrast, the transitive verb tàʕà ‘put’ is preceded by its argument, and the whole serial construction consists of two OV sequences following the TAM marker, as in (5).
|
(4)
|
S
|
+
|
(TAM)
|
+
|
pà ‘come’
|
+
|
DO
|
+
|
V2
|
|
|
|
(5)
|
S
|
+
|
(TAM)
|
+
|
DO
|
+
|
tàʕà ‘put’
|
+
|
DO
|
+
|
V2
|
24Overall, I had an impression that formal criteria would be better suited to describe the diversity of serial constructions and to capture significant differences between them. Thus, “manner constructions” are subdivided into those where the manner verb precedes and those where it follows the verb, cf. (6) and (7).
|
(6)
|
py-ɔ̀
|
yà
|
fɛ̀ɛ
|
pà
|
|
child-sg.def.k
|
pfv
|
run
|
come
|
|
‘The child came running.’ (p. 238)
|
|
(7)
|
pí
|
wá
|
bɔ̀ɔrɔ̂
|
pè
|
ɲɔ̰̀
|
|
3pl.p
|
prf.aff
|
work.sg.def
|
do
|
be.nice
|
|
‘They worked well.’ (p. 238)
|
25Even more diversity is found in constructions of other types. For instance, the verb tò ‘fall’ used in the sense ‘begin’ is always followed by a progressive marker according to the schema in (8). Another syntactic pattern is represented by the infinitive construction involving the element mà preceding V2, as shown in (9).
|
(8)
|
S
|
+
|
(TAM)
|
+
|
tò ‘fall’
|
+
|
PROG
|
+
|
V2
|
|
(9)
|
S
|
+
|
(TAM)
|
+
|
V1
|
+
|
mà
|
+
|
V2
|
26Taking into consideration all these facts, one can see that serial constructions do not constitute a homogenous set. The functionalist approach adopted by the author determines the division of serial constructions according to semantic criteria (such as the (a)symmetric nature of relations between V1 and V2). However, I do not find this approach completely satisfactory – one class can contain constructions belonging to several different structural patterns. Furthermore, the absence of a syntactic analysis in addition to identifying structural schemas can be regarded as a shortcoming of the book. Cross-linguistically, serial constructions can have different syntactic properties – for instance, some OV constructions in West African languages are headed by nominalized verbs, and criteria confirming or disproving this could be provided for each case. Infinitival constructions do not seem to fit well in the definition of “serial constructions” – as far as I understand, they do not contain non-linked verbs but do contain a linker (whose syntactic properties are still to be investigated). In general, considering the great diversity in the domain of serial constructions, I would expect more information on their syntax.
27Chapter 10 is dedicated to topicalization and focalization. Minyanka allows left-dislocation of topics; however, movement of locative constructions involving postpositions is reported to be rare. There are asymmetries between the topicalization of subjects and of non-subjects; the former are marked by a special particle ʃí. Another element used to topicalize subjects is the Bambara loan kɔ̀nì, which is also used as a marker of certainty when placed clause-finally. The element dè functions as a focus marker; if it is placed after a non-emphatic pronoun, the whole proposition is focused (10). Focus marking in situ yields narrow focus readings (11).
|
(10)
|
pí
|
dè
|
yà
|
yɔ̀-ʕɔ̀yì
|
gà
|
|
3pl
|
foc
|
pfv.aff
|
water-sg.def.k
|
drink
|
|
‘What they did was drink water.’
|
|
(11)
|
pí
|
yá
|
yɔ̀-ʕɔ̀yì
|
dè
|
gà
|
|
3pl
|
pfv.aff
|
water-sg.def.k
|
foc
|
drink
|
|
‘It is the water that they drank.’
|
28The discussion of constituent focus is followed by a description of clause-final focus particles (bànì, kɛ̀, kóyí, wálá, dɛ́, dóró, wó, fyéwú, sá, dó), many of which seem to be particles that widespread in the languages of the area and possibly borrowed from Bambara.
29Chapter 11 describes the syntax of interrogation. General (yes/no) questions are marked by one of three particles lá, yɛ́, or wá. Conditions on their use are described; interestingly, some of the particles are associated with certain emotions communicated by the speaker – in particular, lá is reported to be associated with contexts where the speaker is angry. Alternative questions are formed using the element lâa or lǎa presumably related to lá ‘or’. The suggestion seems plausible, given the fact that such polysemy is typical of other languages of the area. The most prominent strategy of partial interrogation is in situ questions, but interrogative adjuncts can be placed both on the left and on the right periphery of the clause.
30Chapter 12 describes strategies of complementation. Quoted clauses are introduced by the preclausal conjunction ná which is analyzed as a quotative marker. Clauses marked by ná are selected by utterance predicates (‘say’, ‘answer’) as well as by cognition verbs (‘know’, ‘forget’). In contrast, predicates of other semantic types require different morphosyntactic marking. Commentative predicates (‘regret’, ‘be satisfied’) select for nominalized verbs, whereas manipulative verbs combine with infinitives or subjunctives. Modal predicates are compatible with subordinate subjunctives, futures or infinitives depending on the types of modality expressed. Verbs expressing epistemic modality (‘believe’, ‘think’, ‘doubt’) require complement clauses headed by indicative verbs. Another strategy that employs preclausal elements are indirect questions – subordinate interrogative clauses are introduced by the marker ní.
31Chapter 13 discusses relative clauses. These employ the in situ strategy where the relativized constituent is not dislocated leftwards within the relative clause. Relativizers are placed to the left of the relativized constituent, and they come in two forms – mî and dem-mì, the latter involving a demonstrative pronoun. Only in free relatives the variant dem-mî cannot be replaced by a simple relativizer mî. Relativized nominal forms usually occur as indefinites. An intriguing feature of Minyanka is that relative clauses which are usually attached to the left of the relativized constituent can be extraposed to the right. In this case, they should obligatorily use relativizers with demonstratives. Like in many West African languages, relativization is accompanied by resumption in the main clause, the resumptive pronouns being from the emphatic series. Minyanka is reported to have preserved traces of other relativization strategies, namely of that involving the subordinator ké/gé found in other varieties of the language. Its disappearance is attributed to the contact influence of Bambara. Relative markers can co-occur with conditional post-subject markers, the strategy frequently used in free choice relatives (‘any X who does V’).
32Subordination not involving verb serialization is described in Chapter 14. Minyanka has both postposed and preposed conjunctions as well as some suffixes such as the gerundive ndɔ̀ɔ̂. Interestingly, many preposed conjunctions are Bambara loanwords, e.g. kàbí ‘since’, fɔ́ ‘till’, yánì ‘before’, hàlí ‘even’, wálásá ‘so that’, jɔ̰̀ŋgɔ̀ ‘in order to’, and yánì ‘instead of’. The preposed conjunction mà tà ‘while, before’ seems to be the only one that was grammaticalized from a language-internal source (← mà kì tà <inf 3sg.k find>). A special case of subordination is presented by a distributive construction involving reduplicated nominalized forms (S + V.nmlz+ wô + V.nmlz ‘each time when V’). It is also noted that some TAM markers can head subordinate clauses – in particular, clauses headed by future forms without conjunctions are largely used to express purpose.
33Coordination is discussed in the last chapter of the book. It is expressed by a range of conjunctions, most of which are loans from Bambara. Conjunction (‘and’) is reserved to the element ní that can link nominal but not verbal constituents. Another element associated with nominal conjunction is wò which is used in enumeration. A typologically interesting feature is the co-existence of two strategies for coordination of propositions. Minyanka disposes of two strategies employing preclausal elements, one of which (à) requires that the subjects of two clauses be different and the other requires that the subjects be coreferential. The former strategy involves subjunctive markers. The latter strategy does not employ any post-subject markers but the tonal melodies of the verb involved resemble those found in sequences of subjunctives as if a verb were preceded by the subjunctive marker ʃíN.
34The Minyanka grammar is well-structured and can be regarded as an excellent contribution to Senufo studies and to Gur linguistics in general. Detailed grammars of Senufo languages are almost entirely missing, with the exception of Carlson (1994), and Coulibaly’s work is undoubtedly comparable with the latter. It is typologically informed, presents a plethora of data and applies a consistent terminology for the phenomena under study, thus making Minyanka data easily accessible to specialists in Senufo and Gur languages. From a descriptivist point of view, the grammar is almost perfect; some issues that may raise questions were discussed above. The book is replete with accurate observations, many of which were reproduced in this review. The main advantage of the grammar is its clarity and the high quality of the data. Examples originate both from elicitation and from texts, and various uses of the same construction are discussed in detail throughout the book. Overall, I find the analyses presented by the author convincing, with some possible exceptions mentioned above.
35Of course, it is inevitable that some aspects of the Minyanka grammar are not covered. Given the fact that Senufo languages are understudied, comparisons to other languages of the same family are only sporadic and are mostly limited to Supyire, as described in Carlson (1994). However, I do not consider it to be a major drawback of the book under review, since descriptive grammars are not supposed to provide extensive historical linguistics-based overviews. However, I am sure that the Minyanka data will allow many interesting insights to study the relationship of Senufo to central Gur, especially taking into account lexical affinities between Minyanka and other non-related Gur languages.
36I find it very difficult to formulate any critique. The shortcomings discussed in this review are minor and do not question the exceptional scientific quality of the book. The level of detail of the data presented in the grammar makes it an outstanding descriptive work. Perhaps, the only component of the description that could be developed further are constraints on the use of grammatical structures. For instance, I would have appreciated if serial constructions had been described with more detail highlighting potential structural asymmetries between different types. However, such finer-grained theoretically-oriented studies are not commonly included in descriptive grammars and their absence cannot be regarded as a serious drawback. The grammar will undoubtedly become a model for Senufo linguistics and a base from which more theoretically oriented linguists can start in order to reveal more about the intriguing aspects of Minyanka.