Varying functions of the comitative preposition ábə̀ in Gavar
-
Melanie Viljoen
Abstract
The purposes of this study are to investigate and describe the varying functions of the comitative preposition ábə̀ with [1] in Gavar, compare these functions with cross-linguistic observations and propose an explanation for their diachronic development. This particle has a wide variety of functions. Firstly, it is used as a preposition marking phrases with a wide range of semantic roles. Secondly, its use has extended to Noun Phrase coordination, a common occurrence cross-linguistically. Its clause linking functions include its use as a temporal/sequential marker and elaboration. Of special interest to typologists is its predicative/complementizer use. This function is not observed in nearby languages and appears to be unusual cross-linguistically. Informed by previous research on the behavior of the comitative/instrumental preposition with cross-linguistically, general grammaticalization principles and an analysis of the semantic links between the different functions based on shared semantic features, a semantic map/network is provided showing the possible paths of development for the various functions of ábə̀ with in Gavar.
1 Introduction
Haspelmath (2003: 211) observes that “A recurrent problem in linguistic analysis is the existence of multiple senses or uses of a linguistic unit.” This is especially the case for function words which have a more general or abstract meaning. Adpositions in particular, being on the divide between grammar and the lexicon, have many interesting characteristics worth investigating (Hagège 2010: 332). Furthermore, adpositions are of interest to students of grammaticalization since they are often the sources of new diachronic processes (Hagège 2010: 97).
Comitatives, expressing accompaniment relations (marked by with in English), are particularly complex cross-linguistically, often exhibiting extensive polysemy with other semantic categories, as well as extending to cover a variety of grammatical functions. As a result, the multifunctionality of comitatives has been a topic of interest to linguistic researchers for several decades.
In examining both textual and elicited language data in Gavar, it was noted that the comitative preposition ábə̀ with has a number of different syntactic and semantic functions; as a preposition marking a wide range of semantic roles, as a noun phrase coordinating conjunction, as a predicative marker/complementizer and in other types of clause linking. It also appears that in Gavar the grammaticalization of this preposition has extended further than comitatives in nearby related languages.
Therefore a study was undertaken in order to gain further insight into the nature of the multiple functions of ábə̀ with, as well as the relations between them. The approach taken to this multifunctionality was a polysemic one, which assumes that there are different meanings or senses attached to each gram which are related in some way that needs to be specified (Haspelmath 2003: 212). The purposes of this study are therefore threefold:
Investigate and describe the varying functions of the comitative preposition ábə̀ with in Gavar from both a morphosyntactic and a semantic perspective.
Compare these findings with cross-linguistic observations of comitatives.
Propose an explanation as to how these various functions may have developed diachronically.
Such a study is expected to be of interest to Africanists as well as semanticists and typologists involved in research on comitatives and related categories.
Before reporting the findings of this study, some background information on the Gavar language is given in Sections 1.1 and 1.2. The nature of the database used will be outlined in Section 1.3 and an outline of the paper provided in Section 1.4.
1.1 Language background
Gavar is a Central Chadic language spoken by around 15,000 people. The Gavar homeland where around two thirds of the ethnic Gavar live, is located in the division of Mayo-Tsanaga, Mogode sub-division in the Far North Region of Cameroon. Sizable populations of ethnic Gavar can also be found in the Hina sub-division just to the south, Garoua (capital of the North Region), and across the border in Nigeria.
The classification of this language according to the Ethnologue (Eberhard et al 2019) is: Afro-Asiatic, Chadic, Biu-Mandara, A, A.7. Gravina (2014: 26) updated the classification of Gavar as follows: Chadic, Central, South, Daba, Buwal. Gavar and Buwal, which borders on the Gavar area to the east, make up the Buwal sub-group.[2]
1.2 Morphosyntactic profile of Gavar
There have been a limited number of in depth linguistic studies previously done on Gavar. Masters theses have been written on the phonology (Noukeu 2004) and nominal morphology (Tchikoua 2006). More recently papers have been published concerning the focus copula (Viljoen 2017) and verb morphology (Viljoen 2019). In view of this limited exposure, a short morphosyntactic profile of the language is given here.
The basic word order of Gavar is SVO/SV. Verbless clauses also exist in the language, having a subject-predicate structure. Some variation in word order does occur, although the pragmatic conditions for this variation are not yet well understood. There is very little noun morphology, apart from plural marking on a limited set of animate nouns and limited genitive marking. However, verb morphology is far more complex and includes: person/number marking, TAM marking and other verbal extensions. Two interesting features of Gavar and other languages of the area are sentence final negation and question words.
1.3 Database
The various functions of ábə̀ with in Gavar were determined based on a corpus of sixty natural texts of various lengths and genres, example sentences provided for lexical items plus other elicited data. This corpus has been entered in Fieldworks Language Explorer but is yet unpublished. Examples in this paper are transcribed phonemically. For the purposes of this paper, the texts and example sentences are considered to be natural language data. Elicited examples given in this paper can be recognized by ‘E’ in their code.
1.4 Outline of paper
This paper is divided into three main sections. Firstly, Section 2 gives a brief summary of the different theoretical models on which the analysis will be based. A general overview of Gavar prepositions is given in Section 3. Section 4 describes and delineates the functions found for ábə̀ with in the Gavar corpus. The findings are compared with what has been observed cross-linguistically and possible paths of diachronic development for the various functions are discussed. A brief conclusion is then given in Section 5.
2 Theoretical background
The label comitative refers to the grammaticalized expression of accompaniment relations (Stolz et al. 2006: 5). Two participants are involved, an accompanee and a companion with a relator which establishes the relation that holds between them (Stolz et al. 2006: 17). Comitatives are usually (but not always) encoded as adverbial or peripheral complements of a predication and are often marked with an adposition, but may also be marked with case markers, clitics or a combination of methods.
What follows is a brief summary of the theoretical frameworks used within this paper for the analysis of the multiple functions of the Gavar comitative preposition ábə̀ with. Section 2.1 gives an outline of the syntactic functions of adpositions and adpositional phrases in general. Section 2.2 covers the semantic variation of comitative adpositions. The further extension of the comitative to cover other grammatical functions is discussed in Section 2.3.
2.1 Syntactic functions of adpositions and adpositional phrases
In terms of the syntactic functions of comitatives, this paper follows the framework provided by Hagège (2010) for adpositions in general. He defines the function of an adposition as marking the grammatical and semantic links between its governed term, with which it forms an adpositional phrase, and another linguistic unit (Hagège 2010: 8). This linguistic unit is an external head which governs the entire adpositional phrase, and could be a verb, noun or adjective. The internal head of the adpositional phrase is the adposition itself.
Hagège (2010: 58) provides a list of eleven types of possible governed terms of adpositions found cross-linguistically. Most governed terms are noun-like, however adpositions also frequently govern clauses (Saint-Dizier 2006: 5). When adpositions mark clauses, many linguists refer to them as subordinating conjunctions. However Hagège (2010: 224–227) points out that some linguists do not distinguish between this subordinating use and the adpositional use since the adposition’s basic function is the same, it is simply the governed term which varies. This is the position which will be followed in this paper.
Adpositional phrases may also have a number of different syntactic functions: core or peripheral complements in a VP, adnominal complements, predicates and as heads with respect to certain dependent elements (Hagège 2010: 97).
2.2 Semantic variation of comitatives
From a semantic point of view, it has been observed that polysemy is especially developed in adpositions (Hagège 2010: 277). This is particularly the case for comitatives. Stolz et al. (2006: 8) note that the similarity of functions of the comitative cannot be explained simply by genealogy as the same pattern of syncretism can be found in unrelated languages and so may reflect a linguistic universal. Various semantic/conceptual models have been proposed by linguists to explain the observed multifunctionality of adpositions in general and comitatives in particular. The sub-sections below give a brief outline of the various models that have informed the analysis of ábə̀ with in Gavar in this paper. Section 2.2.1 introduces the idea of a semantic network/map. The nature of the semantic links between various functions is discussed in Section 2.2.2. Principles which can be used to determine the direction of their diachronic development are detailed in Section 2.2.3. Finally, Section 2.2.4 gives a brief description of Croft’s (1991) model based on the causal structure of verbal semantics.
2.2.1 Semantic network/map
The analysis of the semantic functions of Gavar comitative will be primarily based on the concept of a multi-dimensional semantic network where each sense is connected to others by a number of shared semantic features. Such a semantic network is drawn from the frameworks proposed by Heine et al. (1991), Haspelmath (2003), and Stolz et al. (2006).
The foundation of the semantic network is the idea that semantic roles or case functions are on a continuum which can be broken up into separate, overlapping senses, rather than involving a number of discrete categories (Heine et al. 1991: 163–164). For example, for comitative-instrumental syncretism, a continuum could be conceptualized with comitative at the one end and instrumental at the other with some meanings being closer to comitative and others closer to instrumental (Heine et al. 1991: 105)
Related to the continuum idea is the concept of prototypes where certain senses are thought to be prototypical and others less so. For example, according to Stolz et al. (2006: 26), a prototypical comitative involves a human companion and a nuclear predicate which is a verb of movement. On the other hand, a prototypical instrument is an inanimate object with an action verb as a nuclear predicate. However non-prototypical examples for each category are often observed such as non-human companions or animate instruments. Heine et al. (1991: 256) note that there are two major criteria for the ordering of the prototypicality of the function of a preposition; spatial versus non-spatial and the degree of association with human participants.
A one-dimensional continuum, in terms of which languages may distinguish a number of semantic roles, has been proposed within Role and Reference Grammar (Van Valin 2006: 685). However, Heine et al. (1991: 163–166) argue that just one dimension is inadequate to do justice to the state of affairs and that at least two dimensions (if not more) are necessary. For example possessive does not seem to be part of the comitative-instrument-manner continuum, but constitutes a separate branch. They propose a framework which involves an overall network showing the development of different senses of a linguistic unit undergoing grammaticalization. This network is based on a divergency model that can be represented in the form of a tree diagram with various senses connected with one another by means of lines of conceptual shift (Heine et al. 1991: 98).
An extension of this framework is the semantic map, which is a one or two dimensional geometrical representation of related functions in conceptual/semantic space (Haspelmath 2003: 213). The map constitutes a network with different functions linked by connecting lines (Haspelmath 2003: 216). The closer two functions are on the map, the closer they are conceptually. Which functions are included is based on systematic cross-linguistic comparison; if a function is marked in a distinct way in a particular language it will be added to the map (Haspelmath 2003: 217). For those interested in grammaticalization, semantic maps can make predictions about diachronic change by showing that some changes presuppose others and directionality can also be shown on semantic maps using arrows (Haspelmath 2003: 233–234). A number of researchers have attempted to make semantic maps covering comitatives and instrumentals both based on cross-linguistic studies (Narrog 2009: 599) and for individual languages (Haspelmath 2003: 229; Kilian-Hatz 1995: 122).
2.2.2 Semantic links between functions
While semantic maps do not provide information on the nature of the semantic links between various functions, attempting to characterize the nature of these links can help inform which senses are more closely connected and provide the motivation for their development over time. In this vein, Stolz et al. (2006: 160) take a componential approach to meaning which focuses on individual semantic features rather than holistic categories. The more features are shared between categories, the more likely syncretism is to occur. This leads to a kind of semantic network which displays a family resemblance in which immediate neighbors share at least one feature, whereas more distant members of the chain may display less similarity and are interconnected via another member with which they share more features than with each other. Table 1 below provides a list of semantic features/parameters, in no particular order, noted by Stolz et al. (2006) as being relevant to the semantic analysis of comitives and related categories. Some of these will be used later in the analysis of the Gavar comitative.
Semantic features relevant to comitatives and related categories (Stolz et al. 2006: 26, 61, 67, 120, 127, 130, 140, 164, 305).
| human | permanent |
| animate | concrete |
| dynamic | autonomous |
| co-presence | control |
| egalitarian | force |
| volitional | locative |
| same as | durative |
| group |
2.2.3 Direction and mechanisms of diachronic development
Also useful in the construction of a semantic network/map is a consideration of how meaning change occurs over time. Two processes have been identified: grammaticalization and lexicalization, the first being the most pertinent to the analysis of the multifuctionality of the Gavar comitative.
Grammaticalization is a “process leading toward greater grammaticality of linguistic terms” (Himmelmann 2004: 25). This process involves a number of steps involving both form and function. Heine (2009: 459) explains how these apply to the development of case markers.
Extension: lexical items are extended in their use to a larger range of complement nouns and their meanings become more general
Desemanticization (semantic bleaching): lexical meaning is lost and a schematic case function assumed
Decategorialization: loss in morphosyntactic properties
Erosion: loss in phonetic substance
The above parameters focus on the grammaticalizing item, but Himmelmann (2004: 31–32) points out that it is actually its syntagmatic context which is grammaticalized. He asserts therefore that the domain of grammaticalization for adpositions/case markers is probably the clause rather than the adpositional phrase. Therefore when considering the grammaticalization of the comitative preposition in Gavar we need to take into account the constructions in which it occurs as well as its semantic/grammatical functions. Grammaticalization in Himmelmann’s view involves context expansion on three levels: host-class (the class of items the element is in construction with), syntactic and semantic-pragmatic.
Heine et al. (1991: 156–157) proposed a number of guidelines on how to decide whether category X is more grammaticalized than category Y as applied to case marking. Based on these guidelines, they arranged case functions on a scale of increasing grammaticalization (see Figure 1 below). This scale can then be generalized to spatial relations > human relations > inanimate relations.
Chain of increasing grammaticalization of case functions (Heine et al. 1991: 159).
Croft (1991) identifies two mechanisms for the extension of meaning. The first of these, spread, “involves extension of a form from one element (use) in a semantic domain to a semantically contiguous or ‘nearby’ element in the same domain” (Croft 1991: 184). The second process is domain shift or metaphor in which “the form used for an element in one semantic domain is extended to apply to a ‘parallel’ or ‘similar’ element of another semantic domain” (Croft 1991: 193). Heine et al. (1991: 99–104) contend that the link between two senses can be metaphorical to a greater or lesser extent with the link between adjacent senses tending to be only “weakly metaphorical” if at all.
Metaphor, being one of the main processes underlying grammaticalization, has been discussed extensively in the literature. It can be divided in two types: categorial, which is general and broad, and conceptual, which is more limited in scope (Heine et al. 1991: 52). Categorial metaphorical categories can be arranged on a scale from more concrete to more abstract (see Figure 2) which represents the direction of metaphorical shift (Heine et al. 1991: 45). This scale can therefore assist in the determination of the probable direction of grammaticalization.
Scale of metaphorical categories (Heine et al. 1991: 48).
Meaning change can also occur through lexicalization. According to Himmelmann (2004: 28), this process occurs when “an originally productive, transparent, compositional formation loses its productivity, transparency and/or compositionality.” Idiomization and fossilization are examples of this. Lexicalization and grammaticalization have much in common. The difference between them is that lexicalization involves the narrowing, rather than the extension of the host class (Himmelmann 2004: 36).
2.2.4 Causal structure of verbal semantics
A different model from the semantic network described above, but which provides interesting insights into the motivation for the syncretism of various semantic roles, is that proposed by Croft (1991). He has developed an analysis of verbal semantics in terms of causal structure with a view to providing a semantic framework in which to define thematic (semantic) roles (Croft 1991: 178). In this model, major oblique thematic roles can be described in terms of the ordering of participants in the causal chain, relative to the subject and object. Oblique case markings are divided into two types: those that represent participants that precede the object in the causal chain (antecedent) and those that represent participants that follow it (subsequent) (Croft 1991: 184–186). How Croft categorises certain thematic roles is as follows (Croft 1991: 186):
Subsequent: benefactive, recipient, result
Antecedent: instrumental, manner, means, comitative, passive agent, ergative, cause
Croft asserts that there is a strong typological tendency that surface case markers do not subsume both subsequent and antecedent thematic roles (Croft 1991: 187) and a survey of oblique case marking largely confirms this (Malchukov and Narrog 2009: 527). However, the problem for this model is the observed cross-linguistic syncretism of comitatives with locatives (Stolz et al. 2006: 147). While locatives are not found in the list of thematic roles given above, Croft explains that spatial concepts can be integrated into a causal grammatical structure through a process of coercion (Croft 1991: 198). For a location relation the figure (the object which is located somewhere or moving), and the ground (the location), is coerced into an order which is always figure first, then ground (Croft 1991: 199). The implication is that the ground (location) would then be marked as a subsequent role and so should not syncretise with antecedent roles such as comitative and instrumental.
Like spatial concepts, possession is also considered to be neutral with regards to the antecedent, subsequent causal categories proposed. However, Croft (1991: 210) links the concept of possession, with the spatial metaphor, the possessor being the ground, and the possessee the figure. This is the source of what he terms possessee-first coercion (corresponding with figure-first coercion) which governs a possessee – possessor order in the causal structure and results in antecedent marking on the possessee. This conforms better with the cross-linguistic data, as it is not uncommon for a possessee to be marked in the same way as the comitative, also an antecedent role.
2.3 Further grammaticalization of comitatives
It has been observed that NP conjoining is a common grammaticalization of the comitative cross-linguistically and the direction appears to be well established (Hagège 2010: 309; Haspelmath 2013; Heine and Kuteva 2002: 80–82; Stolz et al. 2006: 21). In a cross-linguistic study, Stassen (2000: 21–23) divides the languages of the world into with languages and and languages. with languages use only the comitative for coordinative encoding. The result is two NPs of unequal structural rank, not being part of the same constituent, one being an oblique marked with the comititaive. There is also singular agreement marking on the verb (Stassen 2000: 18). and languages, on the other hand, differentiate between comitative and coordinative having distinct markers for these two functions. The coordinated NPs are of equal structural rank, can often (though not always) be seen to form one constituent and are associated with dual or plural verbal agreement marking (Stassen 2000: 7). He points out that these two strategies are two extremes on a continuum (Stassen 2000: 21) and that pure instances of with languages are relatively rare. There is usually some grammaticalization of the comitative taking place resulting in a hybrid between the two strategies (Stassen 2000: 26).
While the NP-coordinative function of the comitative is quite common cross-linguistically, clause coordination which develops from it is less so (Heine and Kuteva 2002: 83). However, it has been noted in some languages that comitatives and instrumentals can develop into a kind of temporal clause marker (Heine 2009: 468).
3 Overview of Gavar prepositions
Before looking in detail at the various functions of the preposition ábə̀ with in Gavar, it will be helpful to give a summary of prepositions in Gavar and the semantic roles they mark. In this way we can more easily see where the functions of ábə̀ differ from other prepositions and where they overlap.
Table 2 below gives a summary of other prepositions in Gavar apart from ábə̀ with (excluding those borrowed from Fulfulde), including glosses and the possible semantic roles of prepositional phrases marked by them. Some prepositions have extended their use to marking adverbial clauses. These are noted in italics.
Summary of other prepositions in Gavar.
| Preposition | Gloss | Semantic roles |
|---|---|---|
| á | at, to, from, for, in, into, (with?) | location, goal, source, time, recipient, beneficiary, emphatic reflexive, (manner ?) |
| á dìw | behind | location |
| á dìwʒá | behind, after | location, time |
| á támá | in front, later | location, time |
| á brà | next to | location |
| á xà | over, towards, next to | location, goal |
| á má | on the edge of | location |
| á rá | to, at (a person), because | location, goal, passive agent, reason, reason |
| á ská | under, for, so that | location, beneficiary, purpose |
| átá | on, about | location, topic |
| áná | in, into (a defined space) | location, goal |
| ándzá | like, as, if, when | similative, similative, condition, time |
| ɡʷàrà | towards, by | goal, location, passive agent |
| mavəj | because | reason, reason |
| pá | along, at the level of | location |
| tá | for | purpose |
From this summary it can be observed that Gavar has a general locative preposition á which marks a wide variety of semantic roles. Some of these semantic roles overlap with ábə̀ with (underlined). However when they do overlap there appears to be slight differences in the meaning expressed. This will be explored further in Section 4.1.1.2.1.
Of note is that many of these prepositions are clearly derived from a combination of the general locative preposition á and a body part noun. According to Heine (2011: 698), body-part terms are the most common conceptual source for adpositions in African languages. Other Gavar prepositions begin with á but their derivation is more opaque.
It is possible that the comitative preposition ábə̀ with originally developed from the spatial preposition á brà ‘next to’ which in turn is derived from the body-part term brà ‘belly’. This would be in line with a common general grammaticalization process involving the development of non-locative case markers from locative ones (Heine 2009: 461). This type of process has also been observed specifically for comitatives (Stolz et al. 2006: 360). The concept behind this is that the accompanee and companion occupy contiguous sub-regions in space (Stolz et al. 2006: 161). In terms of form of the Gavar comitative, phonetic erosion frequently occurs as part of the grammaticalization process (Heine 2011: 697) which would explain the loss of the final ra. The addition of the schwa could be explained due to the fact that word final voiced plosives are disallowed in Gavar. However, there are no historical sources of Gavar or related languages available which could either confirm or disprove this hypothesis.
4 Functions of ábə̀ with and their development
Before discussing each function of ábə̀ with in detail it will be helpful to give a brief overview of the functions discovered. Table 3 below summarises the syntactic functions of the preposition ábə̀ with in Gavar found in the corpus. The number of examples found in natural language data are included so as to give an idea of their frequency, although some of the examples provided below are taken from elicited data.
Summary of the syntactic functions of particle ábə̀ with in Gavar.
| Function | Sub-types | Number of examples |
|---|---|---|
| Preposition | NP complement | 94 |
| Clausal complement | 18 | |
| Coordination | Noun phrase | 20 |
| Numeral | 6 | |
| Infinitive | 1 | |
| Clause linking | Temporal marking | 13 |
| Elaboration | 3 |
The particle ábə̀ with, when functioning as a preposition, can mark a wide range of semantic categories (see Section 4.1.1.2.1 for more detail). A two-dimensional semantic network/map (see Section 2.2.1), while having its limitations, appears at present to be the most useful tool available for representing the semantic relations between the comitative and related semantic categories, as well as the more grammatical functions which have developed from it. It also provides a way of showing the possible grammaticalization chains linking these functions. Figure 3 above proposes a semantic network/map for the functions of the Gavar preposition ábə̀ with. Arrows show the directionality of the grammaticalization paths. Since there are no historical sources for this language, the information contained in this map is necessarily speculative. It is based on what has been discovered from cross-linguistic studies, general principles of grammaticalization and a consideration of the possible shared semantic features of the different functions (see Section 2 for theoretical background). The discussion which follows below attempts to provide a justification for the paths/links shown.
Semantic network/map of functions of ábə̀ with in Gavar.
The semantic network in Figure 3 is centered around the comitative. This is considered to be the basic sense from which all others have developed which is in line with what has been observed for languages where historical evidence exists (Heine and Kuteva 2002: 80–89, 2006: 198–200). Furthermore, according to various criteria it can be argued that comitatives are less grammaticalised than instrumentals (Stolz et al. 2006: 193). For example, if a category implies a human participant, it is less grammaticalized than if it does not (Heine et al. 1991: 156).
The semantic map indicates that there is a major division between comitative and instrument (bolded), certain categories having developed from the comitative and others from the instrumental. This reflects the findings of Stolz et al. (2006: 118) that distinct comitatives and instrumentals prefer to syncretize with different categories (see Section 4.1.1.2 for further discussion).
The network has a number of branches representing continua where meaning has spread through extension from one function to another. The more grammaticalised functions beyond the marking of prepositional phrases with NP complements are indicated with italics.
The sub-sections below provide a description of each of the functions of ábə̀ with, a comparison with cross-linguistic findings and a discussion of how each function may have developed over time. Its use as a preposition with an NP complement and with a clausal complement are covered in Section 4.1. Its extension to NP coordination is examined in Section 4.2. Other clause linking functions are discussed in Section 4.3.
4.1 Prepositional use
The preposition ábə̀ with may take either an NP (Section 4.1.1) or a clausal complement (Section 4.1.2). These will be treated separately as they exhibit quite different syntactic and semantic properties.
4.1.1 Prepositional phrases with NP complements
This section begins with a summary of the different syntactic functions which prepositional phrases with an NP complement can fulfill (Section 4.1.1.1). The semantic roles which such phrases can express and how they may have developed from the comitative is discussed in Section 4.1.1.2.
4.1.1.1 Syntactic functions
As outlined in Section 2.1 above, while adpositional phrases often function as peripheral complements within a verb phrase, they may also have other functions. Below is a description of syntactic functions found for prepositional phrases marked with ábə̀ with in Gavar.
Peripheral/Adverbial Complement
This is by far the most common function, accounting for 75 out of the 94 examples in natural language data.
| də̀rlàŋ | àːɡí | à:-lawaɗ | ábə̀ | dàmbə̀j. | |
| youth | pl | 3pl.sbj-play | with | stick | |
| ‘The youth play with a stick.’ | [LL5-E:2.5] | ||||
Predicate
Three examples of this function were found in the natural language data, two of these encoding possessive relations. These occur in verbless clauses (2a), as well as clauses containing the copula verb ɗa in the future tense (2b). The copula verb is not included in the predicate since it is semantically empty. Its function is to convert what follows into a predicate and to carry any tense/aspect marking necessary.
| a. | [sà]S | [ábə̀ | dàlà.] PRED | |
| 1sg | with | money | ||
| ‘I have money.’ (lit. ‘I (am) with money.’) | [LL14-E:12.1] | |||
| b. | [ní-]S á:-ɗa | [ábə̀ | skə̀n. ]PRED | |
| 1excl.sbj-fut-be | with | thing | ||
| ‘We will have thing(s).’ (lit. ‘We will be with thing(s).’) | [LL19-E:8.6] | |||
Adnominal complement
Two examples were also found in the natural data of prepositional phrases marked with ábə̀ with functioning as a noun modifier in a noun phrase (3).
| tá-bəw | tə̀ | á | àlà | ábə̀ | skə̀n. | |
| pfv-change | 3sg.poss | loc | someone | with | thing | |
| ‘He became a rich person.’ (lit. ‘He changed into someone with thing(s).’) | ||||||
| [LL10-SE:19.5] | ||||||
Complement of property terms
An interesting construction in Gavar involves a prepositional phrase marked with ábə̀ with functioning as a complement of an adjective[3] within an adjectival phrase. Nine examples of this were found in the natural data. The adjective expresses a property of the entity encoded by the noun marked with the preposition (4a and 4b). The adjectival phrase can either function as a predicate (4a) or as a depictive secondary predicate (4b).
| a. | [rə̀kʷə̀ts | ndə̀ | sà-tsa-ə̀ | wàtsə́]S | |||
| piece.of.clothing | rel.nsbj | 1sg.sbj-put-3sg.dobj | dem.prox | ||||
| [ tʃáp | ábə̀ | rà. ]PRED | |||||
| cover.completely | with | arm | |||||
| ‘This shirt I am wearing has long sleeves (lit. (is) covered completely with arm).’ [2303] | |||||||
| b. | ándzá | xʷà-tá-ɗa | ɡwár | xʷà-áː-ɡra | |||
| when | 2sg.sbj-pfv-be | arrived | 2sg.sbj-fut-see | ||||
| dʒàk | tə̀ | sə́ksə́k | ábə̀ | tə̀f. | |||
| hut | 3sg.poss | opposite | with | road | |||
| ‘When you have arrived, you will find his house opposite (with) the road.’ | [2749] | ||||||
Two examples were also found of adverbs (5a), including an ideophone (5b) being modified by the prepositional phrase.
| a. | ndə̀ | sà | ɗa | ɡʷàr | káɗ | ábə̀ | də̀mə̀s. | |||
| rel.nsbj | 1sg.sbj- | be | arrived | directly | with | dance | ||||
| ‘When I arrived, it was dancing straight away (lit. directly with dance).’ | [2472] | |||||||||
| b. | ábə̀ | vìjà | mpì | àːɡí | ɡbə́v | |||||
| with | wet.season | plant | PL | abundance.of.leaves | ||||||
| ábə̀ | vàdʒíɗ. | |||||||||
| with | leaves | |||||||||
| ‘During the wet season, plants are very leafy (lit. plants (are) abundant with leaves.)’ | [2360] | |||||||||
4.1.1.2 Semantic functions
4.1.1.2.1 Overview of semantic functions
Prepositional phrases marked by ábə̀ with with NP complements can have a wide variety of semantic roles in Gavar. Table 4 below gives a summary of the fifteen semantic roles found in the corpus, with the number of examples for each type found in natural language data. Semantic roles which were only found in elicited data are still included in the list. The semantic roles in the table are listed from the most to the least frequent.
Summary of semantic roles of PPs marked by ábə̀ with with NP complements.
| Semantic role | Number of examples |
|---|---|
| Comitative (human/animate companion) | 20 |
| Instrument | 15 |
| Mental/physical states | 9 |
| Confective (inanimate companion) | 8 |
| Material | 8 |
| Means | 7 |
| Time | 6 |
| Activity | 6 |
| Direction | 3 |
| Emphatic Reflexive | 2 |
| Possession | 2 |
| Manner | 2 |
| Ornative (temporary bodily property) | 1 |
| Equative | 0 |
| Part-whole (permanent bodily property) | 0 |
As discussed in Section 2.2.2 all these related senses could be thought to be part of a continuum. Being all marked in the same way in Gavar, it is therefore somewhat arbitrary where to put the divisions between closely related meanings. Certain categories can be broken down into further sub-senses (Different sub-senses for comitatives and instruments in Gavar will be detailed below). Other functions could be grouped together. For example confective, ornative and mental/physical states could be classified as non-prototypical comitatives (Stolz et al. 2009: 605).
We could ask how this list of semantic roles compares with those found for comitatives cross-linguistically and for related languages to Gavar. Table 5 below gives the forms of the comitative preposition(s) for nearby languages of Gavar and summarizes the semantic roles marked by them which are mentioned in the literature. The information available for some of these languages is at times sketchy.[4]
Comitative prepositions and their semantic roles for nearby languages to Gavar.
| Language | Preposition | Semantic roles |
|---|---|---|
| Buwal | á | comitative, instrument, material, manner, spatial, time, recipient, benefactive |
| ɮàɓá | comitative, confective | |
| Mina | ábə̀ (plural form íbə̀) | comitative, instrument, time |
| Mbudum | ábə̀ | comitative, instrument, manner, possession, time |
| Daba | tété | comitative, instrument |
| Mofu-Gudur | tà | comitative, instrument, material, manner, location, time |
| Kapsiki | le | comitative, instrument |
The paucity of semantic roles for comitatives listed for the languages in Table 5 is likely to be a result of inadequate research. Be that as it may, it can be observed that in the majority of cases the semantic roles line up with what is observed in Gavar. Interestingly, this is not the case in the closely related language of Buwal where the general locative preposition á has extended its usage to cover the comitative and related functions (Viljoen 2013: 399–401). A specialized comitative preposition ɮàɓá ‘with/along with’ is used when there is a desire to emphasize the accompanying entity (Viljoen 2013: 402–403) and can be used with both animate and inanimate companions. It should be noted also that Buwal does not have the form ábə̀, whereas Mina and Mbudum do.
The semantic roles marked by ábə̀ with for Gavar (see Table 4) line up well with the kinds of categories comitatives tend to syncretize with cross-linguistically (Hagège 2010: 307–308; Heine et al. 1991: 257; Narrog 2009: 598; Stolz et al. 2006: 38–42, 2009: 605). From Table 4 it can be seen that comitative and instrument senses are the most frequent senses in the Gavar corpus. It was previously thought that comitative-instrumental syncretism was universal, however cross-linguistic studies on languages from around the world, such as those conducted by Stolz et al. (2006: 98–102) and Narrog (2009: 598) have demonstrated that this is not the case. The world’s languages can be divided into three types:
A-type: Distinct markers for both categories
B-type: Same marker used for both
C-type: Mixed (Either a special marker for comitative and another which combines both categories or a special marker for instrumental and another which combines both categories.)
We see here that Gavar is a B-type language, using the same marker for both comitative and instrument. This is not unusual in Africa with 30% of African languages in Stolz et al.’s (2006: 105) sample, and 55% in Narrog’s (2009: 598) sample exhibiting instrumental-comitative polysemy. Furthermore, all the nearby related languages to Gavar listed in Table 5 are B-type, apart from Buwal which is C-type.
Languages can vary greatly in the number and type of semantic categories which syncretise with comitatives. For example in Seychelles Creole the comitative also covers instrument, conjunctive, passive agent, source, cause and recipient (Haspelmath 2003: 227), while Kilian-Hatz (1995: 122) found twenty-four different functions for the comitative preposition in Baka including such grammatical categories as reflexive, benefactive, dative and habitual and progressive aspect.
That the senses encoded by ábə̀ with in Gavar are relatively numerous is not surprising considering that Stolz et al. (2006: 115) found that relators which combine comitative and instrumental meanings are more prone to further syncretism than distinct relators. This is because distinct comitatives and instrumentals prefer to syncretize with different categories. Instrumentals tend towards highly agentive categories such as causative, agentive and ergative while comitatives do not (Narrog 2009: 598; Stolz et al. 2006: 115). Furthermore, Stolz (2001: 322) found that instrument prefers to syncretize with locative, while the comitative prefers the possessive.
It can be observed from Figure 3 and Table 4 that in Gavar the extension of instrument to other functions has not occurred to the same degree as in some other languages. For example, it does not cover the agentive categories listed above. It has also not extended to locative (the limited ‘direction’ sense will be discussed further below). This is in constrast to Narrog’s (2009: 598) findings that Afro-Asiatic and Nilo-Saharan languages show a strong affinity between instrumental and locative. As a result, categories which would likely extend from the locative such as benefactive/dative and tense/aspect marking are also not found in Gavar.
4.1.1.2.2 The development of semantic functions
Examples for each of the semantic roles shown in Table 5 are given below with further explanatory detail when necessary. Where a preposition other than ábə̀ with can be used to mark the role in question, examples are given and the difference in meaning discussed. Concurrently, an exploration of the semantic links between the different semantic categories and a justification for the possible direction of their development one from the other (see Figure 3) is given. Each branch or chain on the semantic network is examined in turn, beginning with the central category of comitative.
Comitative (human/animate companion)
Comitative is the semantic category from which all others have developed. However, comitative can be divided into different sub-categories based on animacy, involvement and control of the different participants (Stolz et al. 2006: 56–60). In Gavar, comitative with a human companion can be divided into a number of sub-types as identified by Stolz et al. (2006: 42–43) including co-operative (6a), reciprocal (6b) and active comitative (human companion) (6c).
| a. | ní-ndá | lam-xà | klàŋ | ábə̀ | wàlà | nàkà. | ||
| 1excl.sbj-go | repair-vnt | threshing.circle | with | wife | sg.poss | |||
| ‘I (lit. we) went to repair the threshing circle with my wife.’ | ||||||||
| [NH7-SN:1.1] | ||||||||
| b. | xʷə̀nì-tá-ɗal-xà | mbə́lí | ábə̀ | àlà. | ||||
| 2pl.sbj- pfv-do-vnt | greeting | with | someone | |||||
| ‘You have passed a greeting with someone …’ | [DE7-SN:6.3] | |||||||
| c. | sà-tá-ɗal-xà | ɮə̀n | ábə̀ | xə̀dʒì | mà | xʷələm | ||
| 1sg.sbj-pfv-do-vnt | work | with | person | rel.sbj | paralyse | |||
| tà | à:ɡí. | |||||||
| 3sg.poss | pl | |||||||
| ‘I have worked with handicapped people.’ | [LL10-E:8.4] | |||||||
Note that when the situation involves two human participants, both having the same level of involvement and control, Gavar uses an inclusory construction with plural subject agreement marking on the verb as examples (6a and 6b) above show. This kind of construction is also found in Buwal (Viljoen 2013: 356–357). When the accompanee has more control than the companion as in example (6c), singular subject agreement marking is used.
It is also possible to have a non-human animate companion (7).
| ndə́v | á-ndəv | tə̀ | ábə̀ | dàkʷ | tə̀. | |
| fall | 3sg.sbj-fall | 3sg.poss | with | horse | 3sg.poss | |
| ‘… he fell with his horse.’ | [NH4-N:2.11] | |||||
Comitative-Emphatic Reflexive
The emphatic reflexive in Gavar is formed by replacing the companion in a comitative phrase by the noun xà ‘head’ modified by a possessive pronoun which agrees with the subject (8). This function, where the entity marked by ábə̀ with is co-referential with the subject, can be thought of as being similar to reciprocal or reflexive which is a common syncretism cross-linguistically.[5]
An emphatic reflexive may or may not involve a human or animate entity. As seen in example (8) below, it may be an inanimate force of nature. This differentiates it semantically from the comitative. However it has other semantic features in common with comitatives which enables this syncretism such as egalitarian, same as and control.
| kʷáxʷə́ | wàtsà | mà | kəɗaŋ | ábə̀ | xà | tə̀ | |
| fire | dem.dist | rel.sbj | finish | with | head | 3sg.poss | |
| á | ba | bə̀rlà | wàtsà. | ||||
| inf | burn | mountain | dem.dist | ||||
| ‘It is the fire which finished burning the mountain by itself .’ | |||||||
| (lit. ‘This fire is the one which finished with its head to burn this mountain.’) | [GE1-SE:7.1] | ||||||
Note that it is also possible for emphatic reflexives to be formed with the general locative preposition á (9). According to one language informant, the use of the comitative preposition gives greater emphasis to the fact that the subject really did the action alone.
| xʷə́ní-təŋɡʷəl | nkə̀ɗàŋ | á | xà | kʷə̀nì. | |
| 2pl.sbj-roll | rock | loc | head | 2pl.poss | |
| ‘You roll the rock yourselves.’ (lit. ‘You roll the rock to/for your head.’) | |||||
| [GE1-SE:6.4] | |||||
Comitative-Equative
The preposition ábə̀ with is also used to express the equative when used in conjunction with the verb saɓ ‘resemble’ (10a and 10b). This type of use is similar to a comparison of equality, which is found in Estonian (Stolz et al. 2006: 30). In this construction the companion is the entity with which the subject is compared. Since the equative use of this preposition is limited to this verb, it could be considered to be part of its argument structure, and therefore an example of lexicalisation rather than grammaticalisation (see Section 2.2.3).
The examples below show that for the equative, the subject and the comparee are either both human (10a) or both non-human (10b). When two human participants are involved, the verb is marked with a plural subject marker (10a) as described above for inclusory constructions (6a and 6b). When the participants are not human, a singular subject marker is used (10b).
| a. | xʷə́ní-ká-saɓ | ábə̀ | màtsáxʷ. | ||||
| 2pl.sbj-ipfv-resemble | with | mother.2 poss | |||||
| ‘You resemble (with) your mother.’ | [LL5-E:3.4] | ||||||
| b. | ɮə̀n | wàlà | à-ká-saɓ | kálkál | ábə̀ | ||
| work | woman | 3sg.sbj-pfv-resemble | equal | with | |||
| ɮə̀n | xə́ládə̀mà | =w | |||||
| work | girl | =q | |||||
| ‘Is a woman’s work the same as a girl’s?’ (lit. ‘Does a woman’s work resemble equally with a girl’s?) | [LL1-E:27] | ||||||
As example (10b) shows, unlike the comitative, the equative may involve non-human, even non-concrete entities. Furthermore they do not share the semantic feature control as the equative function involves equivalence in property, not in action. In this sense, we see an extension in meaning in terms of broad metaphorical categories from person to quality. The semantic features comitative and equative have in common include egalitarian and same as.
Comitative-Confective-Possession-Part-whole/Confective-Ornative-Mental/physical states
Categories such as confective, ornative, and mental/physical states (abstract properties) are referred to by Stolz et al. (2009: 605) as non-prototypical comitatives. They assert that on a semantic map, these functions are located between the comitative prototype and possession since “they relate to temporary or permanent relations in which the companion is physically or ideally very close to the accompanee, thus almost in his/her possession.” One could therefore say that the semantic feature all of these categories have in common is co-presence (This equally applies to the category part-whole). Therefore, as these functions are so closely related, they are being discussed together here.
Examining the syntactic constructions such prepositional phrases occur in, plus looking at the other semantic features which they share (see Table 6 above), a different proposal than that of Stolz et al. (2009) can be made concerning their order of development. Rather than these functions constituting a single continuum, the network branches in two directions from the confective node (see Figure 3). More detailed discussion on the justification for this is given below. In terms of the scale of metaphorical categories (see Figure 2), there is a progression from person to object to quality overall for this branch.
Semantic features for prototypical and non-prototypical comitatives and possession.
| Comitative | Confective | Possession | Part-whole | Ornative | Mental/physical states | |
|---|---|---|---|---|---|---|
| permanent | − | − | − | + | − | − |
| animate | + | − | − | − | − | − |
| control | + | + | + | − | − | − |
| concrete | + | + | +/− | + | + | − |
| co-presence | + | + | + | + | + | + |
Confective (inanimate companion)
Confective involves an animate accompanee carrying or transporting a concrete object (Stolz et al. 2006: 42) as in example (11) below.
| àvándzá | áká | nxìl | míndʒá | à-ndá | nxil-xà | ||
| former.times | exist | thief | ind.det.sg | 3sg.sbj-go | steal-vnt | ||
| á | dámə̀ | ábə̀ | ɡʷə̀làŋ | tə̀ | |||
| loc | bush | with | quiver | 3sg.poss | |||
| ‘There was a time, a certain thief went to steal in the bush with his quiver.’ | [NF3-WN:2.1] | ||||||
In terms of semantic features, confective can be distinguished from comitative in that there is a shift from [+ animate] to [− animate]. These two functions are also strongly linked in terms of form, with both comitative and confective phrases mostly functioning as peripheral complements within the Verb Phrase.
Possession
The preposition ábə̀ with can also be used to mark the possessee (Y) in an expression of the form: X is with Y. This is what is termed the Companion Schema by Heine (1997: 53), a type of metaphor in which the possessee (Y) is conceptualized as a companion to the possessor (X). This schema is widely found in Bantu and Niger-Congo languages but also occurs all over the world (Heine 1997: 55). Stolz et al. (2006: 148) also note that the association of comitatives with possession is very in strong in Africa.
This construction in Gavar is used to express alienable possession of inanimate objects (12a, also 2a and 2b above) as well as abstract possession (12b).
| a. | xàjàk | ábə̀ | tàmtàkʷ | áská | rà. | |
| Hayak | with | club | under | hand | ||
| ‘Hayak has a club in his hand.’ (lit. ‘Hayak (is) with a club under (his) hand.’) | [NH11-SN:5.1] | |||||
| b. | sà | ábə̀ | dèŋ. | |||
| 1sg | with | thought | ||||
| ‘I have an idea.’ (lit. ‘I (am) with thought.’) | [LL14-E:12.3] | |||||
Note that possession can also be expressed in other ways in Gavar, using Heine’s (1997: 51) location schema, Y is at X’s place and also the genitive schema, X’s Y exists (Heine 1997: 58). However the companion and genitive schemas appear to be the most frequent.
Concerning the development of companion possession from confective, a major shift in syntactic form occurs between confective and possession. The possessee is encoded as a predicate (2a and 2b; 12a and 12b). There is also a metaphorical shift in meaning in that it is possible for the possessee to be abstract. It can also be seen that comitative to possessive conforms to the order of the scale of increasing grammaticalization of case functions (see Figure 1).
Part-whole (permanent property)
Part-whole involves “two entities … permanently associated with each other” (Stolz et al. 2006: 43). One could view this as a type of inalienable possession such as that which occurs with body-parts as in example (13).
| màwə̀l | ábə̀ | dàvàzà. | |
| man | with | body.hair | |
| ‘The man has body hair.’ (lit. ‘Man is with body hair.’) | [LL14-E:19.19] | ||
There is a close relationship between possession to part-whole relations at the level of both syntactic and semantic function. The one example of part-whole available involves the prepositional phrase functioning as a predicate (13), which as was seen above is the most common construction for possession. On the semantic level, we see a shift from alienable (or temporary) to inalienable (or permanent) possession. There is also a shift from [+control] to [−control].
Ornative (temporary bodily property)
Stolz et al. (2006: 43) define ornative as “an ephemeral bodily property of an animate participant.” This is always associated with adnominal attribution (see Section 4.1.1.1) with the attribute helping to identify the person in question.
Example (14) below illustrates this role.
| àː-tá-gəra | kʷá | àlà | ábə̀ | skə̀n | áká | skʷə́. | |
| 3pl.sbj-pfv-see | even | someone | with | thing | exist | neg | |
| ‘They didn’t see anyone with something.’ | [NH12-SN:5.8] | ||||||
On the semantic level, there is a close connection between possession and ornative. They share the semantic feature of [+temporary]. However, there is a shift from [+control] to [−control]. Syntactically there is shift from the predicative function to the adnominal attributive function.
Mental/physical states
The preposition ábə̀ with can also mark temporary mental (15a) and physical states (15b).
| a. | sà | ŋmkpéf | ábə̀ | ɣʷə̀ɗèk. | ||||
| 1sg.stat | full | with | anger | |||||
| ‘I (am) filled with anger.’ | [LL14-E:22.1] | |||||||
| b. | là-ɣʷə́ɗék | tə̀ | ànà | mà | wan | ká | ||
| nom.act-get.angry | 3sg.poss | cop | rel | sleep | purp | |||
| ábə̀ | míntíʃ. | |||||||
| with | hunger | |||||||
| ‘It was his anger that had him sleep hungry (lit. with hunger).’ | [GE3-SN:18] | |||||||
The category mental/physical states is connected to confective in terms of syntactic form since they are often encoded as peripheral complements within a VP (15b). Here we see also mental/physical state phrases functioning as complements of property terms such as adjectives (15a, see also 4a and 4b). In terms of semantics there is a shift from [+concrete] for confective, to [−concrete].
Comitative-time
The preposition ábə̀ with can be used with time expressions involving a period during the day (16a), a particular day itself (16b) or a season (16c).
| a. | á | béɮí | áː-nda | tà | ábə̀ | |||
| loc | ceremonial.place | 3pl.sbj-go | 3pl.poss | with | ||||
| ntsə̀ɓàxʷ | áʒá. | |||||||
| evening | so | |||||||
| ‘So to Beldi (ceremonial place), they go during the evening.’ | ||||||||
| [DP7-SN:3.3] | ||||||||
| b. | kàtà | àːɡí-áː-nda-xà | ábə̀ | lə̀wmà | psàkàj. | |||
| perhaps | 3pl.sbj-fut-go-vnt | with | market | Psiki | ||||
| ‘Perhaps they will come on Monday.’ | [LL26-E:1.3] | |||||||
| c. | ábə̀ | vìjà | mpì | àːɡí | ||||
| with | wet.season | plant | pl | |||||
| ɡbə́v | ábə̀ | vàdʒíɗ. | ||||||
| abundance.of.leaves | with | leaves | ||||||
| ‘During the wet season, plants have an abundance of leaves.’ | [2360] | |||||||
The majority of temporal expressions in Gavar are marked with the general locative preposition á. It frequently occurs cross-linguistically that temporal relations are conceptualized in terms of spatial relations (Heine et al. 1991: 188). In Gavar however, some temporal expressions can be marked with either á loc (17) or ábə̀ with (cf. 16a).
| sà-áː-nda-xà | ápát | á | ntsə̀ɓàxʷ | á | zàɗə̀. | |
| 1sg.sbj-fut-go-vnt | tomorrow | lo c | evening | loc | night | |
| ‘I will come tomorrow in the evening at night.’ | [LL19-E 13.9] | |||||
The difference in the prepositions used relates to a different conception of the time period in question. When the locative preposition is being used, a particular point in time is being referred to. When ábə̀ with is used, the whole of the time period is in view. In English we might use the preposition during to express this second perspective of time. It is likely in Gavar that the use of ábə̀ with to express time developed directly from the comitative rather than via a spatial use which is virtually non-existent in this language. Adding weight to this hypothesis is the situation in Maltese which has separate markers for comitative and instrumental, both of which can be used with temporal expressions (Stolz et al. 2006: 300). When the comitative is used, the reading is one of duration or of a time span, while the instrumental marks a more punctual event. Stolz et al. (2006) argue that with the comitative, the time period could be conceived of as being a companion, sharing the feature co-presence with the comitative.
The extension from comitative to time conforms to Heine et al.’s (1991: 159) chain of increasing grammaticalization for case functions (see Figure 1) as well as their scale of broad metaphorical categories (Heine et al. 1991: 48) (see Figure 2), there being a shift from person to time.
Comitative-instrument
The category instrument can be divided into various sub-types (Narrog 2009: 593–594; Stolz et al. 2006: 63). In Gavar we find such instruments as tools (18a), body-parts (18b), animate (18c) and means of transportation (18d).
| a. | wàlà-í | ndíɓí | àːɡí | àː-ká-jaɓ | |||
| woman-pl | ind.det.pl | pl | 3pl.sbj-ipfv-wash | ||||
| tékíɗ | ábə̀ | ɡìʃkìl. | |||||
| calabash | with | folere.leaves | |||||
| ‘Certain women were washing calabashes with folere leaves.’ | |||||||
| [TN12-SN:3.1] | |||||||
| b. | sà-ndá | mbəla-zə̀ | ábə̀ | rà. | |||
| 1sg.sbj-go | catch-trans | with | arm | ||||
| ‘I will go and catch (him) with (my) arms.’ | [NH11-SN:5.4] | ||||||
| c. | dìw | tə̀ | káx | sá-kax-ə̀ | |||
| beginning | 3sg.poss | plough | 1sg.sbj-plough-3sg.dobj | ||||
| ábə̀ | màdàmbə̀rʒàxʷ. | ||||||
| with | donkey | ||||||
| ‘… in the beginning I plough it (field) with a donkey.’ | [DP13-SN:1.1] | ||||||
| d. | kàtà | à:ɡi-á:-nda-xà | ábə̀ | mə̀wtá | áká | skʷə́. | |
| perhap | 3pl.sbj-fut-go-vnt | with | car | exist | neg | ||
| ‘Perhaps they will not come by (lit. with) car.’ | [LL26-E:1.4] | ||||||
As mentioned above, the direction of syncretism from comitative to instrument has been well-established from historical studies as well as on the basis of grammaticalization principles. In terms of semantic features Stolz et al. (2006: 164) suggest that the features these two functions have in common are control and force. One could also add the feature co-presence to this list.
In terms of categorial metaphorical categories, there is a shift from person to object. The conceptual metaphor of this type often associated with comitative-instrumental syncretism is an instrument is a companion which was proposed by Lakoff and Johnson (1980: 135) to explain what they considered at the time to be the universal syncretism of comitative and instrument in the one marker. Although further studies have demonstrated that many languages outside of Europe have separate markers for these two senses (Hagège 2010: 313; Heine and Kuteva 2006: 183; Stolz et al. 2006: 109), this metaphor does provide an explanation for the significant number of languages which do exhibit such polysemy.
Instrument-material-means-activity/means-manner
Looking at syntactic structure does not greatly help in determining the direction of development for the instrument-activity/means-manner continuum(s). In the vast majority of cases phrases encoding these semantic categories function as peripheral verbal complements and so it could be argued that their development is almost purely due to semantic extension. From Table 7 below we see that the semantic feature that all these functions have in common is that the participant marked with ábə̀ with in some way is under the control of the participant encoded by the subject. Furthermore, instrument, material and means all share the feature force, meaning that they have some kind of effect on a patient/undergoer which may be encoded either by the subject or direct object of the verb. In terms of metaphorical categories, there is a progression from object to activity, and object to quality (see Figure 2).
Semantic features relevant to the instrument-activity/manner continuum.
| Instrument | Material | Means | Activity | Manner | |
|---|---|---|---|---|---|
| control | + | + | + | + | + |
| force | + | + | + | − | − |
| concrete | + | + | − | − | − |
| object | + | − | − | − | − |
| dynamic | − | − | +/− | + | − |
Material
A material is a “substance which a user employs to achieve certain goals” (Stolz et al. 2006: 43). It differs semantically from instrument in that it is a substance rather than an inanimate object and so is [−object].
| à-ban | ábə̀ | yə̀m | kʷáxʷə́ | kʷá | ntrà | pál. | |
| 3sg.sbj-wash | with | water | fire | even | month | single | |
| ‘She washes with hot water even for one month.’ | [DE4-WN:1.3)] | ||||||
Means
Means, for the purpose of this paper, is considered to be a type of abstract instrument. Therefore it is differentiated semantically from instrument and material in that it is [–concrete]. Means in Gavar can be divided into sub-types. For example it can be an internal force or capacity (20a) or an action (20b), so that it may or may not have the semantic feature dynamic.
| a. | nìkʷə́kʷə́ɗ | tá-ŋɡa-ànə̀ | ŋɡàs | tə̀ | ká | |
| Nikukud | pfv-break-3sg.iobj | foot | 3sg.poss | purp | ||
| ábə̀ | mbàɗ | tə̀ | ||||
| with | magic | 3sg.poss | ||||
| ‘Nikukud had broken its foot with his magic.’ | [TN8-WN:2.6] | |||||
| b. | xì-baɗ | wàlà | ɡàvàr | ábə̀ | nxìl | |
| 1incl.sbj-court | woman | Gavar | with | theft | ||
| ‘We court Gavar women by stealing (them).’ (lit. ‘We court Gavar women with theft.’) | [DE1-SN:1.1] | |||||
As example (21a) below shows, the Gavar view language as a means by which to speak. This idea extends as well to the style or genre of language (21b). This is similar to what is found in Maltese which also uses the instrumental marker bi for words, language and and other abstract instruments (Stolz et al. 2006: 287).
| a. | ɮáp | ní-ɮap-àtànà | á | tar-xà | ábə̀ | |||
| speak | 1incl.sbj-speak-3pl.iobj | inf | pray-vnt | with | ||||
| mà | ɡàvàr. | |||||||
| mouth [6] | Gavar | |||||||
| ‘We told them to pray in (with) the Gavar language.’ | [NH9-SN:1.7] | |||||||
| b. | ndə̀ | sà-ɮap-ànə̀ | ábə̀ | mà | ɡʷə̀và | |||
| rel.nsbj | 1sg.sbj-speak-3sg.iobj | with | word | side | ||||
| mɮà | tá-ɮima=s. | |||||||
| blacksmith | pfv-hear=neg | |||||||
| ‘When I speak to him with paraboles, he doesn’t understand.’ | [481] | |||||||
Activity
The preposition ábə̀ with also not infrequently marks an activity which may be expressed with a noun (22a) or a nominalized verb[7] (22b). This type of use of the comitative is not often mentioned in the literature. It is not uncommon however, for verbal nouns to be governed by a preposition (Hagège 2010: 58). An example of the comitative preposition tɛ marking a nominalized verb can be found in Baka (Kilian-Hatz 1995: 160).
| a. | à-lam-zə̀ | bàbà | tə̀ | ábə̀ | ɮə̀n. | |||
| 3sg.sbj-help-trans | father | 3sg.poss | with | work | ||||
| ‘She helps her father with work.’ | [DE12-SN:4.2] | |||||||
| b. | àlà | wə̀lə̀f | à-sa | jàká | ábə̀ | ndà. | ||
| someone | blind.person | 3sg.sbj-drink | difficulty | with | walk | |||
| ‘A blind person has difficulty with walking.’ | [LL15:5.1] | |||||||
The activity function shares the semantic feature dynamic with means.
Manner
Narrog (2009: 598) notes that polysemy between instrumental and manner in Africa is common and this syncretism is also found in Gavar (23).
| ní-bəl | ábə̀ | ɓàs. | |
| 1excl.sbj-dig | with | joy | |
| ‘We dig (it) joyfully.’ (lit. We dig (it) with joy.’) | [LL6-E:23.5] | ||
In Gavar, certain prepositional phrases expressing manner can be marked with the general locative preposition á (24). All examples of this type found in the corpus involve the idea of non-physical ‘strength’. Therefore this could be interpreted as ‘in strength’ or ‘strongly’ rather than ‘with strength or force.’
| à-dza | wə̀là | á | ɓə́r-ɓə́r | tə̀. | |
| 3sg.sbj-hit | voice | loc | strong | 3sg.poss | |
| ‘He cries out loudly.’ (lit. ‘He hits voice in(?) its strength.’) | [LL4-E:32.1] | ||||
Two examples were found in the corpus of ‘strength’ being used with ábə̀ with. In these cases it is physical strength being referred to (25). More examples in context will be needed to more fully determine the difference in meaning between the two prepositions.
| à:-nda-xà | ábə̀ | ɓə́r-ɓə́r | tə̀ | |
| 3pl.sbj-walk-vnt | with | strong | 3sg.poss | |
| ‘… they walk back quickly …’ (lit. ‘… they walk to here with its strength …’) | [DE6-SN:6.2] | |||
Manner does not share the semantic feature dynamic with activity and therefore it is directly connected to means as a separate branch on the semantic map.
Manner-direction
The spatial use of prepositional phrases marked with ábə̀ with appears to be limited to either ‘forward’ and ‘backward’ (26).
| àː-ʃi | ábə̀ | má | á | támá, | àː-tak-àtà. | |||
| 3pl.sbj-run | with | rel.sbj | loc | front | 3sg.sbj-obstruct-3pl.dobj | |||
| àː-ʒin-xà | ábə̀ | má | á | dìwʒá | ||||
| 3pl.sbj-return-vnt | with | rel.sbj | loc | behind | ||||
| àː-tak-àtà | ||||||||
| 3pl.sbj-obstruct-3pl.dobj | ||||||||
| ‘They ran forward (lit. with that which is in front), they blocked them. They came back (lit. with that which is behind), they blocked them.’ | [NH4-SN:4.3-4] | |||||||
The grammaticalization chain given by Heine et al. (1991: 159) for case functions (see Figure 1) shows all spatial concepts (locative, allative, ablative and path) at the beginning of the chain, while manner is at the end. Therefore the development of direction from manner in Gavar is the reverse of what would be expected according to this framework. It also contravenes the scale of metaphorical categories (Heine et al. 1991: 45) which shows space before quality (see Figure 2).
However, Croft’s (1991) causal chain model helps to explain how this kind of development could come about. This is a metaphorical shift from causal relations to spatial relations. It was seen that the spatial use of ábə̀ with in Gavar is limited to ‘forwards’ and ‘backwards’. Direction in this sense is not a typical locative and so does not relate to the ground, which would be marked with subsequent marking. It relates instead to the path. In the causal chain path precedes the ground and so naturally takes an antecedent type marking, in this case ábə̀ with. In a sense, direction or path could be viewed as a type of ‘manner’ of movement.
4.1.2 Prepositional phrases with clausal complements
Eighteen examples of prepositional phrases marked by ábə̀ with with clausal complements were found in the natural Gavar data plus a significant number in elicited data. Their syntactic functions do not mirror exactly those for prepositional phrases with NP complements. For example, they do not function as peripheral complements within a VP, nor as complements of nouns, adjectives or adverbs. Rather in the majority of examples (16 out of 18), they function as predicate of a verbless clause. In the remaining two examples, they are used directly as complement clauses. In terms of semantics, ábə̀ with in these constructions appears to have developed into a general non-core participant marker whose interpretation depends on the context. More details of these functions are given in Section 4.1.2.1 below. A discussion of how they may have developed diachronically is found in Section 4.1.2.2.
4.1.2.1 Functions of clauses marked with ábə̀ with
Predicate
A clause marked by ábə̀ with can function as a predicate of a verbless clause. A common example of this type is the reversed question construction. Ordinarily question words occur clause finally in Gavar as in example (27a). However it is possible to move the question word to the front of the clause, with the main part of the clause preceded by ábə̀ with (27b). This yields a verbless clause construction where the question word becomes the subject and the prepositional phrase the predicate. In such examples which follow, the subject and the predicate are marked with square brackets.
| a. | sà-áː-lam-zə̀-àxʷə̀ | áváŋɡá? | ||
| 1sg.sbj-fut-help-trans-2sg.dobj | how | |||
| ‘How can I help you?’ | [LL1-E:23.2] | |||
| b. | [áváŋɡá] S | [ábə̀ | sà-áː-lam-zə̀-àxʷə̀?]PRED | |
| how | with | 1sg.sbj-fut-help-trans-2sg.dobj | ||
| ‘How can I help you?’ (lit. ‘How (is) that I will help you?’) | [LL1-E:23.1] | |||
Both of these interrogative clause constructions can function as complement clauses. The examples below show such clauses functioning as direct object of the verb ɣal ‘tell, show’ (28a and 28b).
| a. | [sà-ɣal-axʷə | [ sà-ɗal | vèpíj.] CC]MC | ||
| 1sg.sbj-tell-2sg.iobj | 1sg.sbj-do | when | |||
| ‘I tell you when I did (it).’ (lit. ‘I tell you I did (it) when.’) | [LL17:11.2] | ||||
| b. | [sà-ɣal-axʷə | [[vèpíj] S | [ábə̀ | sà-ɗal-ə̀.] PRED]CC]MC | |
| 1sg.sbj-tell-2sg.iobj | when | with | 1sg.sbj-do-3sg.dobj | ||
| ‘I tell you when I did it.’ (lit. ‘I tell you when is that I did it.’) | |||||
| [LL17:11.1] | |||||
Clauses marked by ábə̀ with can also function as predicates in non-interrogative clauses as in the following examples (29a and 29b).
| a. | [aw də zhavel]S | [ábə̀ | àː-ká-ja-ə̀ | |
| ‘eau de javel’ | with | 3pl.sbj-ipfv-call-3sg.dobj | ||
| á | mà | násárá.]PRED | ||
| loc | mouth | white.man | ||
| ‘ “Eau de javel” (is) that they call it in French.’ | [DE7-SN:4.1] | |||
| b. | [ákátʃíxʷá]S | [ábə̀ | xì-ntir-ə̀.]PRED | |
| today | with | 1incl.sbj-write-3sg.dobj | ||
| ‘It’s today that we wrote it.’ (lit. ‘Today (is) that we wrote it.’) | [1370] | |||
It is probable that the constructions described above are a kind of pseudo-cleft construction, although the pragmatic conditions for their use need more study. They are used to express focus on non-core participants. Examples in both natural and elicited data show that this construction can express comitative, instrument, manner, passive agent, quantity, reason, spatial and temporal roles as well as indirect object. The interpretation of the meaning of ábə̀ with in this construction depends on the context, such as the presence of an interrogative word/phrase (27b and 28b) or a temporal adverb (29b).
The majority of examples in natural data express either manner (29a) or time (29b). Manner clauses marked with ábə̀ with are also not infrequently found in the concluding statement in procedural texts describing how something is done (30a and 30b). The preposition may even occur at the beginning of the sentence as in example (30b). In this case the subject is omitted.
| a. | [íjkʷà]S | [ábə̀ | xə̀dʒì-ká-dʒi | wàlà | á | ɡàvàr.]PRED | ||||
| like.this | with | 1incl.sbj-ipfv-take | wife | loc | Gavar | |||||
| ‘Like this is how (that) we take a wife in Gavar.’ | [DE1-SN:5.1] | |||||||||
| b. | [ábə̀ | xì-ká-ɗa | mánáván. ]PRED | |||||||
| with | 1incl.sbj-ipfv-draw | rain | ||||||||
| ‘(That’s) how (that) we make it rain.’ | [DP6-SN:5.1] | |||||||||
Another example of this type of construction expressing time, is given below (31). In this case the subject in the subject-predicate structure is a headless relative clause marked by the relative pronoun ndə̀. This relative pronoun is used whenever non-subjects are relativized and has also developed into a marker of temporal adverbial clauses.
Six out of the nine temporal examples in the natural data are of this type, the construction expressing the simultaneity of two actions.
| [ndə̀ | sà-nda | àká | á | lə̀wmà,]S | |
| rel.nsbj | 1sg.sbj-go | imm | loc | market | |
| [ábə̀ | sà-daː-zə̀ | àːrá.]PRED | |||
| with | 1sg.sbj-bring-trans | sim | |||
| ‘When I came back from the market, (was) that I brought it back.’ | [2727] | ||||
A possible example of a reason clause from a natural spontaneous text is found in (32) below which comes from a description of a cholera epidemic. In this case the sentence begins with the preposition ábə̀ with. It ends a short section describing the efforts of the government agents who were sent to combat the spread of the disease and elaborates on why they should be thanked.
| xìjə̀-ká-ɗal-àtànà | wásíj | dàɡàlà | á | xì | |||
| 1incl.sbj-ipfv-do-3pl.iobj | thanks | a.lot | loc | person | |||
| ɡʷàmnà | àːɡí. | ábə̀ | mtʃì | ndə̀ | àː-kax-ə̀ | ||
| government | pl | with | corpse | rel.nsbj | 3pl.sbj-bury-3sg.dobj | ||
| mtʃì | à-ká-ŋɡəɗ-ə̀ | skʷə́. | |||||
| corpse | 3sg.sbj-ipfv-count-3sg.dobj | neg | |||||
| ‘We give many thanks to the government people. (It’s) because (that) the corpses that they buried cannot be counted.’ | [DE7-SN:5.3-4] | ||||||
The focus construction with ábə̀ with cannot be used for core grammatical roles such as subject and direct object which uses a different construction involving the focus copula ànà (Viljoen 2017: 157–159). Interestingly however, both construction types can be used for indirect object (33), which is introduced by a preposition, and also cross-referenced on the verb. This could be an indication that indirect object is on the borderline between core and non-core roles.
| á | vàjà | ábə̀ | xʷà-vəla-ànə̀ | wə̀dà? | |||
| loc | who | with | 2sg.sbj-give-3sg.iobj | food | |||
| ‘To whom did you give food?’ (lit. ‘To whom is that you gave food?) | [LL1-WE:42.2.] | ||||||
Complement clauses
As was seen in examples (28a a and 28b) above, interrogative sentences with relative clauses marked with ábə̀ with in the predicate role can function themselves as complement clauses. It is also possible for the interrogative word to be omitted resulting in the prepositional phrase functioning directly as a complement clause (34a and 34b). Again the interpretation of ábə̀ with depends on the context.
| a. | [sà-ɣal-àxʷə̀ | [ábə̀ | wàlà-í | |
| 1sg.sbj-tell-2sg.sbj | with | woman-pl | ||
| àː-ká-da | kʷə́nə́.]CC]MC | |||
| 3pl.sbj-ipfv-prepare | millet.drink | |||
| ‘I (will) tell you how (that) women make millet drink.’ | [DP9-SN:1.1] | |||
| b. | [sà-ɣal-àxʷə̀ | [ábə̀ | sà-ɗal-ə.]CC]MC | |
| 1sg.sbj-tell-2sg.sbj | with | 1sg.sbj-do-3sg.dobj | ||
| ‘I tell you the moment (that) I do it.’ | [LL17-E:12] | |||
4.1.2.2 Development of comitative-predicative-complementizer
A comitative prepositional phrase with a clausal complement which functions as a predicate or a verbal complement clause appears to be a function completely unique to Gavar when compared with neighboring languages. Furthermore, cross-linguistically this seems to be also unusual, with adpositions more commonly developing into subordinating conjunctions marking adverbial clauses. How the predicative/complementizer function may have developed in Gavar is discussed in this section.
The semantic features egalitarian or same as can help to explain the extension of ábə̀ with to marking clauses as predicates of verbless clauses. Such clauses express equivalence between the subject and the predicate. This is illustrated by the use of this construction to express simultaneous events (31)
The direction of grammaticalization from the marking of NPs to clausal marking is in line with Heine et al.’s (1991: 157) hypothesis: “If a given morpheme governs both noun phrases and clauses, then the latter use is more grammaticalized than the former.” The preposition ábə̀ with with NP complement is limited to marking peripheral participant roles. Therefore, in its extension to governing clauses, it is not surprising that this limitation should still apply. However its use has extended even further to become a general non-core participant marker, covering categories not seen for the preposition with NP complement such as: passive agent, quantity, reason and space. This kind of generalization is a further indication of grammaticalization. One could envisage the grammaticalization process following these steps:
Peripheral participant marked with ábə̀ with shifted to the focal position at the beginning of the clause.
The preposition ábə̀ with copied to mark the clausal predicate
Meaning extension to cover all non-core participant roles
Steps 1 and 2 above can be illustrated by the comitative (35a) and instrument (35b) uses of this construction.
| a. | [ábə̀ | vàyè]S | [ábə̀ | à-ɗal | ɮə̀n | wàtsə́?] PRED | ||
| with | who | with | 3sg.sbj-do | work | dem.prox | |||
| ‘With whom did he do this work?’ (lit. ‘With who (is) that (with) he did this work?’) | [LL1-E:32.1] | |||||||
| b. | [ábə̀ | vémé]S | [ábə̀ | xʷə́ní-da | wìrì | wàtsə́ | ákatʃíxʷá?]PRED | |
| with | what | with | 2pl.sbj-prepare | sauce | dem.prox today | |||
| ‘With what did you prepare this sauce today?’ (lit. ‘With what (is) that (with) you prepared this sauce today?’) | [LL1-E:33.2] | |||||||
The complementizer use appears to be in the process of developing from the predicative use via the reversed question structure as described in Section 4.1.2.1 (27). As mentioned there, the reversed question can function as a complement clause within the VP (28). The interrogative word can then be omitted, resulting in the clause marked by ábə̀ with functioning directly as a complement of the verb (34). The fact that there are only a few examples of this type in the corpus and that the full reversed question structure can still function as a verbal complement is an indication that the grammaticalization process is not yet complete.
4.2 Coordination
The function of the preposition ábə̀ with in Gavar has extended its use to a coordinating conjunction. The following sections outline the types of coordination it can be used for (Section 4.2.1). Section 4.2.2 briefly discusses how this development may have come about.
4.2.1 Coordination sub-types
4.2.1.1 Noun phrase coordination
The preposition ábə̀ with can be used in the coordination of noun phrases as the following examples illustrate. The entities being coordinated can be animate (36a), inanimate (36b) or human (36c).
| a. | tíw | ʃìvà | ábə̀ | drə̀kʷ | àː | tsá | ||
| ass.pl | billy.goat | with | ram | 3pl | be.there | |||
| ká-nda | tà | á | dámə́ | |||||
| ipfv-go | 3sg.poss | loc | bush | |||||
| ‘The billy-goat and the ram were leaving for the bush.’ (TN3-WN:1.1) | ||||||||
| b. | kʷár | á-kʷar-xà | ɡʷə̀làŋ | tə̀ | ábə̀ | |||
| pick.up | 3sg.sbj-pick.up-vnt | quiver | 3 sg.poss | with | ||||
| kʷápá | tə̀. | |||||||
| spear | 3sg.poss | |||||||
| ‘He picked up his quiver and his spear.’ [NH4-N:2.4] | ||||||||
| c. | tsáː | ká-jaɓ-ànə̀ | rə̀kə̀ts | bàbà | ||||
| be.undertaking | ipfv-wash-3sg.iobj | clothes | father | |||||
| tə̀ | àːɡí | ábə̀ | àtáː | màmə̀n | ||||
| 3sg.poss | pl | with | gen | mother.3poss | ||||
| tə̀ | àːɡí | ábə̀ | ɮìŋɡìn | àːɡí. | ||||
| 3sg.poss | pl | with | sibling.3poss | pl | ||||
| ‘She undertakes washing her father’s clothes and those of her mother and her siblings.’ [DE12-SN:2.4] | ||||||||
Note that pronouns can also be coordinated in this way (37). In this example three elements are being coordinated with ábə̀ with occurring between each.
| mə̀ | ábə̀ | xʷà | ábə̀ | mbə̀ | kʷà | ndə̀ | |
| 1dual | with | 2sg | with | child | 2sg.poss | rel.nsbj | |
| ɓrə̀s | á | màzrə̀f | wàtsə́ … | ||||
| tail | loc | backside | dem.prox | ||||
| ‘Me and you and your child who has a tail on his backside …’ | |||||||
| [TN9-SN:1.8] | |||||||
It is also possible for ábə̀ with to precede only the final item in an enumerated list (38).
| … skə̀n | zə̀m | dàgàlà | mávə́ | wə̀dà kàŋkə̀s | ||||
| thing | eat | a.lot | beer | bean.doughnut | ||||
| ábə̀ | skə̀n | tá | zə̀m | m à | párpár | párpár. | ||
| with | thing | for | eat | rel.sbj | different | different | ||
| ‘… a lot of food: beer, bean doughnuts and many different things to eat.’ | [DE8-SN:2.2] | |||||||
4.2.1.2 Numeral coordination
The preposition ábə̀ with is also used to coordinate numerals in the formation of numerals greater than one hundred or a thousand as in the following examples (39). Lower additive numerals are formed using the preposition á xà ‘over’ as seen for ‘seventy-nine’ in example (39b) below.
| a. | á | và | ndìm-ndìm | ɡbàk | ábə̀ | wàm. | |||
| loc | year | thousand | two | with | ten | ||||
| ‘In the year two thousand and ten.’ | [DE7-SN:1.1] | ||||||||
| b. | má-báɮ-á-bàɮ | àːɡí | xì | mà | |||||
| nom-exterminate-nom.pat-exterminate | pl | person | rel.sbj | ||||||
| də̀ɡnàk | blàkʷ | nfáɗ | ábə̀ | rà-n ʃ ílíɗ | á-xà | ||||
| black | thousand | four | with | hand-seven | loc -head | ||||
| dzánfáɗ. | |||||||||
| nine | |||||||||
| ‘Four thousand and seventy-nine people were killed.’ (lit. ‘The ones exterminated were four thousand and seventy nine human beings.’) | [NH4-N:4.19] | ||||||||
4.2.1.3 Infinitive coordination
It is possible for ábə̀ with to coordinate clauses consisting of infinitival forms in Gavar (40). Infinitives are marked with the general location preposition á. As infinitives are in a sense a nominal form of the verb which are functioning in this case as a complement to the verb ji ‘accept’, this could be regarded as a similar construction to noun phrase coordination.
| à-ká-dʒi | á | ntir | skə̀n | ábə̀ | á | ja | ||
| 3sg.sbj-ipfv-accept | inf | write | thing | with | inf | read | ||
| skə̀n | skʷə́. | |||||||
| thing | neg | |||||||
| ‘She doesn’t want to read and write.’ | [DE14-SN:1.4] | |||||||
4.2.2 Development of comitative-NP coordination
As has been seen above, Gavar uses the same particle for encoding both comitative and NP coordination. Gavar is not unusual in this regard. Stassen (2000: 41) states that most of the languages of Africa in and below the Sahara are of this type. Furthermore, this strategy is also common in Chadic languages, applying to thirteen out of twenty-four languages listed by Frajzyngier (1996: 27). This is also the case for the majority of nearby languages, the one exception being Buwal, which though closely related to Gavar, uses a distinct conjunction éj ‘and’ for NP coordination (Viljoen 2013: 354–355), possibly borrowed from Fulfulde which uses e to connect nominals (Stennes 1961: 48).
As previously mentioned, comitative to NP-coordination is a very common grammaticalization cross-linguistically with the direction being well established. We see that this process is in line with general grammaticalization parameters (Heine 2009: 459), with the preposition ábə̀ with having undergone a process of desemanticisation. Lexical meaning has been lost, and the particle has taken on a generalised, logical function.
According to Stassen’s (2000: 26) categories, Gavar is a mixed-with language as there has been some grammaticalization of the comitative (see Section 2.3) but Gavar has not yet developed fully into an and-language with distinct markers for comitative and coordinative. In terms of the path of grammaticalisation, for SVO languages he proposes the following steps for diachronic shift from the with to the and strategy (Stassen 2000: 27):
the comitative NP moves to the front of the verb
NP with-NP is reanalyzed as a single constituent
subject agreement becomes dual or plural.
We see for Gavar, that steps (a) and (c) above have occurred. It is difficult, however to find evidence for (b), particularly as the inclusory constructions highlighted above (6a and 6b, 10a) involve both accompanee and companion being encoded in a plural subject agreement marking when they are human. Such constructions could be seen as an intermediate step between comitative phrases and NP coordination. Stassen (2000: 29) cites a similar situation in Tera, a Western Chadic language spoken in Nigeria, where the NP cannot be moved but plural agreement is possible. In Gavar the grammaticalization has gone further in that the NP can be moved to the front of the verb. Potential evidence for (b) would require examples of inanimate coordinate NPs functioning as a subject, co-occurring with plural agreement marking on the verb. Unfortunately all subject coordinate NPs in the corpus involve human (or at least animate) nouns.
From the point of view of semantics, Stolz et al. (2006: 131–139) propose the following semantic features which are shared by prototypical comitatives and and-coordination and which facilitate their syncretism: co-presence, same as, egalitarian and group. Both categories involve more than one participant who are of equal status, involvement and control in the event. This particularly applies to the inclusory constructions discussed above where both participants are also both human.
4.3 Clause linking
The NP-coordinating function of ábə̀ with in Gavar has not extended to general clause coordination. Nor has this type of extension occurred in any of the neighboring languages. This is not surprising as cross-linguistically the use of comitatives for clause coordination is less common than NP coordination. For African languages, Haspelmath (2013) states that differentiation between nominal and verbal (encompassing clausal) coordination is widespread, with juxtaposition often being used for verbal conjunction. Chadic languages tend to fit this pattern. Frajzyngier (1996: 26–28) notes that most Chadic languages do not have a sentential coordinating conjunction and where they do, it appears to be a product of a relatively recent grammaticalization.
In Gavar, the most frequent method of clause conjoining in natural texts is juxtaposition as in example (41) below.
| à- | nda | pá | á | tə̀f, | ||
| 3sg.sbj- | go | at.level.of | loc | road | ||
| ntsə̀ɓə̀n | dzá | á-dza | ábə̀ | nfá, | ||
| knocking.foot | hit | 3sg.sbj-hit | with | flour | ||
| tə̀ | wàtsà | nfá | tə̀ | |||
| 3sg.poss | dem.dist | flour | 3sg.poss | |||
| wàtsà | záx | á-zax | tə̀. | |||
| dem.dist | spill | 3sg.sbj-spill | 3sg.poss | |||
| ‘He went a short distance along the road, he knocked his foot (and) the flour spilled.’ | [TN12-SN:2.1] | |||||
However, the preposition ábə̀ with in Gavar has developed certain clause linking functions which are discussed below. These are temporal marking (Section 4.3.1) and elaboration (Section 4.3.2), both of which are likely to have developed from the NP-coordination function.
4.3.1 Temporal marking
In some languages comitatives and instrumentals grammaticalize into temporal clause markers (see Section 2.3). An example of this type is sequential marking. Frajzyngier (1996: 24) identifies sequential marking as a particular formal device commonly used in clause combining in Chadic languages. This marker indicates a very specific temporal connection between two events. Firstly, that the event described by the second clause occurs after the event described by the first clause. Secondly, that the second event results from the first event (Frajzyngier 1996: 40). This second clause marked by the sequential marker could also be interpreted as a result clause, which Dixon (2009: 17) defines as describing a natural consequence of what is described by its supporting clause.
The Gavar particle ábə̀ with can be used as such a sequential marker, meaning ‘then’. This construction is relatively frequent in the language, twelve examples being found in the natural language data as well as a number of elicted examples. For some examples, the second event is clearly a consequence of the first (42a and 42b).
| a. | sà-tá-xaʃaʃaŋ | mà | wàtsà | ká | ábə̀ | |||||
| 1sg.sbj-pfv-forget | word | dem.dist | purp | with | ||||||
| dàwə̀r | à-mbəla-àkà. | |||||||||
| disease | 3sg.sbj-catch-1sg.dobj | |||||||||
| ‘I had forgotten those precautions, so then I caught the disease.’ | ||||||||||
| [GE11-SE:4.1] | ||||||||||
| b. | tíw | átʃá | tá-drakʷ | skə̀n | á | da | áʒá | ɣʷàɗèɗè | ||
| finally | if/when | pfv-learn | thing | inf | cook | so | clean | |||
| ábə̀ | à-ká-áː-da | skə̀n, | à-ɗal | |||||||
| with | 3sg.sbj-ipfv-fut-cook | thing | 3sg.sbj-do | |||||||
| á | wàtà | màwə̀l | tə̀. | |||||||
| loc | home | husband | 3sg.poss | |||||||
| ‘Finally if/when she has learned to cook things well, then she will cook things, she does (it) at her husband’s home.’ | [DE12-SN:3.2] | |||||||||
For other examples, the result sense of the event expressed in the second clause is not particularly strong and one could argue that it is simply marking temporal succession (43a and 43b). Therefore it seems the result sense is not inherent within ábə̀ with but arises from contextual interpretation.
| a. | xì-ja | wə̀ʒí | bə̀zà, | ||||||
| 1incl.sbj-call | children | compl | |||||||
| ábə̀ | xì-áː-diw | wə̀dà | á | zəm. | |||||
| with | 1incl.sbj-fut-begin | food | inf | eat | |||||
| ‘We call the children first, and then we will begin eating food.’ | [GE8-SE:4.1] | ||||||||
| b. | sà-tá-pik | mánámádàbà | bə̀zà | ábə̀ | sà-nda | nàkà | |||
| 1sg.sbj-pfv-fill | water.pot | compl | with | 1sg.sbj-go | 1sg.poss | ||||
| ‘I filled the water pot, then I left.’ | [2722] | ||||||||
In terms of nearby languages, Mbudum also appears to use ábə̀ with to mark temporal succession. Mofu-Gudur uses a particle which has a similar form to the comitative (Hollingsworth and Hollingsworth 2008: 304), but it is not clear whether they are diachronically related. In Buwal (Viljoen 2013: 587), Mina (Frajzyngier and Johnston 2005: 373–377) and Kapsiki (Frajzyngier 1996: 50) sequential markers are completely unrelated to the comitative preposition.
However, in Gavar the function of ábə̀ with in such structures is not limited to this specific type of sequential marking. It is also possible for the same structure to express simultaneous events, rather than temporal succession as in example (44) below. Only one example of this type was found in the corpus, however. The majority of the time when simultaneity is being expressed, the predicative structure described in Section 4.1.2.1 is used.
| sà-ká-tadəkʷ-xà | bə̀rlà | má | dʒíŋ | wàtsà, | |||
| 1sg.sbj-ipfv-descent-vnt | mountain | rel.sbj | high | dem.dist | |||
| ábə̀ | xʷà-ká-ja-àkà. | ||||||
| with | 2sg.sbj-ipfv-call-1sg.dobj | ||||||
| ‘I was descending that tall mountain while you were calling me.’ | [2233] | ||||||
It appears therefore that ábə̀ with is functioning as general marker of a temporal relationship between two clauses, the interpretation of which depends on the tense/aspect marking within each clause. For example in (42a and 42b) and (43b) verbs are marked with the perfective prefix tá- in the first clause, indicating completion of the first event before the second begins. Likewise, the completive marker bə̀zà occurs in the first clause of examples (43a and 43b). While in (42b) and (43a), verbs in the second clause are marked with the future prefix á:-, again indicating temporal succession of some kind. In example (44) however, the verbs of both clauses are marked with the imperfective prefix ká-.
The essential semantic component of the temporal relationship between the two events therefore seems to be proximity in time, either simultaneous or one directly following the other. A similar situation was found by Stolz et al. (2006: 301) in Maltese. As in Gavar, in this language the comitative marks temporal expressions with a durative type meaning. However it may also mark a more punctual event if it occurs close in time to another one. Stolz et al. (2006) explain this with the idea that the two events in a sense “accompany each other”, they are in a sense “co-present”.
While this may explain the semantic motivation for the extension of the use of the comitative to temporal/sequential marking in Gavar, it does not explain the grammaticalization path taken in terms of syntactic form. Kilian-Hatz (1995: 163–165) proposed that the extention of the comitative preposition tɛ in Baka to that of a temporal conjunction expressing simultaneity occurred via the temporal marking of prepositional phrases. The grammaticalization process involved verbo-nominal constructions marked with tɛ being reinterpreted as subordinate clauses. This also appears to be the process followed by Maltese, with all the examples provided involving the events marked with the comitative being encoded by nominals (Stolz et al. 2006: 301). This does not appear to be the path followed by Gavar however, as the clause marked by ábə̀ with, is the main clause and contains a fully inflected verb. Therefore it is likely that the temporal clause linking use in Gavar has developed via NP coordination. This is one of the grammaticalisation paths proposed by Frajzyngier (1996: 84) for the sequential marker in Chadic languages. The fact that thirteen out of twenty-four languages listed by him use the same form for NP coordination and sequential marking (Frajzyngier 1996: 24) adds weight to this suggestion.
4.3.2 Elaboration
At times ábə̀ with appears to mark clauses which provide an elaboration of the situation which has been described in the previous clause (45). Only three examples of this type were found in the corpus. Dixon (2009: 28), states that for elaboration clause linking, “the second clause echoes the first, adding additional information about the event or state described.” For example in (45a), the first clause states how a sequence of songs are sung during a ceremony to celebrate the birth of twins, and the second clause details some of the words which are sung. Example (45b) concerns the actions of initiation candidates, how they go to a special place to continue the initiation ceremony. The first clause states that they go there, while the second clause, which is marked by ábə̀ with, clarifies that people usually first tie palm leaves to their feet before going.
| a. | xə̀dʒì | ŋɡə́ɗ | ŋɡə́ɗ | ŋɡə́ɗ | ŋɡə́ɗ | ŋɡə́ɗ | |||
| 1incl | recount | recount | recount | recount | recount | ||||
| xì-ŋɡəɗ | drə̀f | tə̀ | já, | nìnì | |||||
| 1incl.sbj-recount | song | 3sg.poss | top | (hesitation) | |||||
| ábə̀ | xì-ká-ja-ə̀ | já | wáw | wə̀ʒí | |||||
| with | 1incl.sbj-ipfv-say-3sg.dobj | top | Oh! | children | |||||
| àː | má | far | ánɮìŋ | wàlà-í | wə̀ʒí | àː | |||
| 3pl | rel.sbj | dig.up | peanuts | woman-pl | children | 3pl.stat | |||
| má | far | ánɮìŋ | wàlà-í | xə̀dʒíj. | |||||
| rel.sbj | dig.up | peanuts | woman-pl | 1incl.quot | |||||
| ‘We recount the songs, um, and we sing, “Oh, it’s the children who dig up the women’s peanuts, it’s the children who dig up the women’s peanuts.”’ [DP5-SN:7.1] | |||||||||
| b. | á | béɮí | áː-nda | tà | ábə̀ | ntsə̀ɓàxʷ | |||
| loc | ceremonial.place | 3pl.sbj-go | 3pl.poss | with | evening | ||||
| áʒá | ábə̀ | xə̀dʒì-tá-dza | ɬìbà | ábə̀ | ntsə̀ɓàxʷ | áʒá | |||
| so | with | 1incl.sbj-pfv-hit | palm | with | evening | so | |||
| ábə̀ | xì-nda | á | béɮí | ||||||
| with | 1incl.sbj-go | loc | ceremonial.place | ||||||
| ‘So to Beldi, they go in the evening, and so we tie on palm leaves in the evening, and then we go to Beldi.’ [DP7-SN:3.3] | |||||||||
Elaboration clauses provide more detail about a certain event. In that sense the two events could be considered as co-present, as was seen for temporal marking above. Elaboration and temporal clause linking are also closely connected in terms of syntactic structure. One could perhaps argue that in fact they are the same construction, with the interpretation of the exact nature of the relationship between the two clauses depending on context.
5 Conclusions
This study of the widely varying functions of the comitative preposition ábə̀ with in Gavar has unearthed some interesting features which should be of interest to Chadicists and Africanists in particular, and semanticists and linguistic typologists in general. Here are some of the most salient points.
Firstly, prepositional phrases consisting of the comitative with an NP complement mostly function as peripheral complements within a VP, but can also modify nouns, adjective and adverbs. The meaning of ábə̀ with has extended from the comitative to cover a wide range of other semantic roles such as: emphatic reflexive, equative, confective, possession, part-whole, ornative, mental/physical states, time, instrument, material, means, activity, manner and direction. This type of extension and the semantic categories covered are not unusual cross-linguistically. What is of note for Gavar however, is the very limited syncretism with spatial/locative categories which can be sometimes observed in other languages and as a result there is no syncretism with more agentive categories such as passive agent, causative and ergative. Furthermore, it means that the temporal use of the comitative cannot have developed from the locative as is usually the case. Rather, it seems to have arisen directly from the comitative and indicates duration.
The grammaticalization of the comitative preposition ábə̀ with to a noun phrase conjoining function is not surprising as this process is common cross-linguistically and also in Africa. Apart from Buwal, this function was also found in nearby languages. The so-called ‘inclusory’ constructions found in Gavar could be seen as an intermediate step in this process.
The extension of the NP coordinating function to general clause coordination is less frequently seen in the languages of the world, and also does not occur in Gavar. However the comitative preposition has developed other clause linking functions, in particular temporal and elaboration. The temporal function covers its use as the sequential marker often seen in Chadic languages. A similar function was noted for the related language, Mbudum. It was noted, however, for Gavar that depending on the tense/aspect marking on the verbs in each clause, the two events could be interpreted as simultaneous rather than sequential. It seems likely that this temporal use has developed via NP coordination. The elaboration clause linking function appears to be closely linked to the temporal and this meaning may simply arise based on the context.
What is of particular interest in Gavar is the predicative/complementizer use of ábə̀ with. It seems to have arisen from a kind of focus construction, though the pragmatic factors behind its use are not yet well understood and would bear further investigation. This development is not observed in neighboring languages and seems to be unusual cross-linguistically. The variable word order found in Gavar may have facilitated this process.
Underlying the diachronic development of the multiple functions of ábə̀ with in Gavar are the semantic links between the different categories based on shared semantic features, as well as general principles of grammaticalization gleaned from cross-linguistic observations. Based on these, a semantic map/network was constructed, showing the possible grammaticalization paths for the various functions of ábə̀ with. Since there are no historical sources available for this language, this proposal is necessarily speculative but may provide interesting insights.
Croft’s (1991) semantic framework based on causal structure was found to fit well with the Gavar patterns of syncretism. All the semantic roles covered by ábə̀ with fit into Croft’s category of antecedent roles. Furthermore, Croft’s idea of coercion from the causal to the spatial domain helps to explain how the direction use of ábə̀ with may have come about. However, all spatial relations are to the far left of the chain of increasing grammaticalization of case functions given by Heine et al. (1991: 159). One could conclude therefore that no theoretical model is perfect in its predictions of the linguistic phenomena observed in actual languages.
Abbreviations
- 1, 2, 3
-
1st, 2nd, 3rd person
- ass
-
associative
- cc
-
complement clause
- comp
-
complementizer
- compl
-
completive
- cop
-
copula
- dem
-
demonstrative
- det
-
determiner
- dist
-
distal
- dobj
-
direct object
- dual
-
dual
- gen
-
genitive
- excl
-
plural exclusive
- exist
-
existential
- imm
-
immediate
- incl
-
inclusive
- ind
-
indefinite
- inf
-
infinitive
- iobj
-
indirect object
- ipfv
-
imperfective
- fut
-
future
- loc
-
locative
- mc
-
main clause
- neg
-
negative
- nom
-
nominalizer
- nsbj
-
non-subject
- pat
-
patient
- pfv
-
perfective
- pl
-
plural
- poss
-
possessive
- pot
-
potential
- pred
-
predicate/predicative
- prox
-
proximal
- purp
-
purposive
- quot
-
quotative
- rel
-
relative marker
- s/sbj
-
subject
- sg
-
singular
- sim
-
simultaneity
- stat
-
stative
- temp
-
temporal marker
- top
-
topic marker
- trans
-
transitivity marker
- vnt
-
ventive
References
Barreteau, Daniel. 1983. Description du Mofu-Gudur. Paris: Université de la Sorbonne Nouvelle PARIS III.Search in Google Scholar
Croft, William. 1991. Syntactic categories and grammatical relations: The cognitive organization of information. Chicago: The University of Chicago Press.Search in Google Scholar
Dixon, Robert M. W. 2009. The semantics of clause linking in typological perspective. In The semantics of clause linking: A cross-linguistic typology, 1–55. Oxford: Oxford University Press.10.1093/oso/9780199567225.003.0001Search in Google Scholar
Eberhard, David M., Gary F. Simons & Charles D. Fennig. 2019. Gavar. In Ethnologue: Languages of the world. Dallas, Texas: SIL International. http://www.ethnologue.com (accessed 23 July 2019).Search in Google Scholar
Frajzyngier, Zygmunt. 1996. Grammaticalization of the complex sentence: A case study in Chadic. Amsterdam & Philadelphia: John Benjamins.10.1075/slcs.32Search in Google Scholar
Frajzyngier, Zygmunt & Eric Johnston. 2005. A grammar of Mina. Berlin & New York: De Gruyter.10.1515/9783110893908Search in Google Scholar
Gravina, Richard. 2014. The phonology of Proto-Central Chadic. Utrecht: LOT.Search in Google Scholar
Hagège, Claude. 2010. Adpositions: Function-marking in human languages. Oxford: Oxford University Press.10.1093/acprof:oso/9780199575008.001.0001Search in Google Scholar
Haspelmath, Martin. 2003. The geometry of grammatical meaning: Semantic maps and cross-linguistic comparison. In Michael Tomasello (ed.), The new psychology of language: Cognitive and functional approaches to language structure, vol. 2, 211–242. Mahwah, New Jersey: Lawrence Erlbaum Associates.Search in Google Scholar
Haspelmath, Martin. 2013. Nominal and verbal conjunction. In Matthew, S. & Martin Haspelmath (eds.), World atlas of linguistic structures online. Leipzig: Max Plank Institute for Evolutionary Archeology. https://wals.info/chapter/64 (accessed 5 October 2019).Search in Google Scholar
Heine, Bernd. 1997. Possession: Cognitive sources, forces and grammaticalization. Cambridge: Cambridge University Press.10.1017/CBO9780511581908Search in Google Scholar
Heine, Bernd. 2009. Grammaticalization of cases. In Andrej Malchukov & Andrew Spencer (eds.), The Oxford handbook of case, 458–469. Oxford: Oxford University Press.10.1093/oxfordhb/9780199206476.013.0030Search in Google Scholar
Heine, Bernd. 2011. Grammaticalization in African languages. In Heiko Narrog & Bernd Heine (eds.), Oxford handbook of grammaticalisation, 696–707. Oxford: Oxford University Press.10.1093/oxfordhb/9780199586783.013.0057Search in Google Scholar
Heine, Bernd & Tania Kuteva. 2002. World lexicon of grammaticalization. Cambridge: Cambridge University Press.10.1017/CBO9780511613463Search in Google Scholar
Heine, Bernd & Tania Kuteva. 2006. The changing languages of Europe. New York: Oxford University Press.10.1093/acprof:oso/9780199297337.001.0001Search in Google Scholar
Heine, Bernd, Ulrike Claudi & Friederike Hünnemeyer. 1991. Grammaticalization: A conceptual framework. Chicago: University of Chicago Press.Search in Google Scholar
Himmelmann, Nikolaus. 2004. Lexicalization and grammaticalization: Opposite or orthogonal? In Walter Bisang, Nikolaus Himmelmann & Björn Wiemer (eds.), What makes grammaticalization? A look from its fringes and its components, 21–42. Berlin & New York: De Gruyter.10.1515/9783110197440Search in Google Scholar
Hollingsworth, Judith H. & Kenneth Hollingsworth. 2008. Dictionnaire Mofu-Gudur-Français-Fulfude, vol. 2. Cologne: Rüdiger Köppe.Search in Google Scholar
Hollingsworth, Kenneth. 2004. A survey of clause types in Mofu-Gudur. Yaounde: SIL Cameroon, ms.Search in Google Scholar
Kilian-Hatz, Christa. 1995. Das Baka: Grundzüge einer Grammatik aus der Grammatikalisierungperspektive. Cologne: Institut für Afrikanistik, Universität zu Köln.Search in Google Scholar
Lakoff, George & Mark Johnson. 1980. Metaphors we live by. Chicago: University of Chicago Press.Search in Google Scholar
Lienhard, Ruth & Marti Giger. 1976. Daba grammar phrase level sorting work paper. Yaounde: SIL Cameroon, ms.Search in Google Scholar
Malchukov, Andrej & Heiko Narrog. 2009. Case polysemy. In Andrej Malchukov & Andrew Spencer (eds.), The Oxford handbook of case, 518–534. Oxford: Oxford University Press.10.1093/oxfordhb/9780199206476.013.0035Search in Google Scholar
Narrog, Heiko. 2009. Varieties of instrumental. In Andrej Malchukov & Andrew Spencer (eds.), The Oxford handbook of case, 593–600. Oxford: Oxford University Press.10.1093/oxfordhb/9780199206476.013.0041Search in Google Scholar
Newman, Paul. 1977. Chadic classifications and reconstructions. Afroasiatic Linguistics 5. 1–42.Search in Google Scholar
Noukeu, Serge. 2004. Etude phonologique du Gavar. Yaounde: Yaoundé I Maitrise en Linguistique Générale.Search in Google Scholar
Saint-Dizier, Patrick. 2006. Introduction to the syntax and semantics of prepositions. In Patrick Saint-Dizier (ed.), Syntax and semantics of prepositions, 1–25. Dordrecht, The Netherlands: Springer.10.1007/1-4020-3873-9_1Search in Google Scholar
Smith, David. 1969. The Kapsiki language. Ann Arbor, Michigan: Michigan State University Doctoral thesis.Search in Google Scholar
Stassen, Leon. 2000. AND-languages and WITH-languages. Linguistic Typology 4(1). 1–54. https://doi.org/10.1515/lity.2000.4.1.1.Search in Google Scholar
Stennes, Leslie. 1961. An introduction to Fulani syntax. Hartford: The Hartford Seminary Foundation.Search in Google Scholar
Stolz, Thomas. 2001. To be with X is to have X: Comitatives, instrumentals, locative, and predicative possession. Linguistics 39. 321–350. https://doi.org/10.1515/ling.2001.013.Search in Google Scholar
Stolz, Thomas, Cornelia Stroh & Aina Urdze. 2006. On comitatives and related categories: A typological study with special focus on the languages of Europe. Berlin & New York: De Gruyter.10.1515/9783110197648Search in Google Scholar
Stolz, Thomas, Cornelia Stroh & Aina Urdze. 2009. Varieties of comitative. In Andrej Malchukov & Andrew Spencer (eds.), The Oxford handbook of case, 601–608. Oxford: Oxford University Press.10.1093/oxfordhb/9780199206476.013.0042Search in Google Scholar
Tchikoua, Paul. 2006. Morphologie Nominale du Gavar. Yaounde: Yaoundé I Maitrise en Linguistique Générale.Search in Google Scholar
Van Valin, Robert. 2006. Functional relations. In Keith Brown (ed.), Encyclopedia of language and linguistics, 683–696. Amsterdam: Elsevier.10.1016/B0-08-044854-2/05275-5Search in Google Scholar
Viljoen, Melanie. 2013. A grammatical description of the Buwal language. Bundoora: La Trobe University PhD thesis.Search in Google Scholar
Viljoen, Melanie. 2017. A comparison of the structural and functional properties of the optional copula in Buwal and Gavar. In Proceedings of the 8th world congress of African linguistics Kyoto 2015, 149–168. Kyoto: Research Institute for Languages and Cultures of Asia and Africa, Tokyo University of Foreign Studies.Search in Google Scholar
Viljoen, Melanie. 2019. Gavar verb morphology. Topics in Chadic linguistics X: Papers from the 9th Biennial International Colloquium on the Chadic languages, Villejuif, September 7–8, 2017. Memoriam Russell G. Schuh. (Chadic Linguistics/Linguistique Tchadique/Tschadistik, 11), 257–274. Cologne: Rüdiger Köppe Verlag.Search in Google Scholar
© 2022 the author(s), published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.
