Search

This paper provides a preliminary semantic description of the verbal category referred to in modern grammar books of Moroccan Arabic as the “medio-passive” (e.g. Harrell 1962). This category is distinguished by the medio-passive marker ‘t-’ that is prefixed to a verb stem to express a variety of grammatical meanings. The origin of this prefix can be traced back to a similar morpheme in Forms V, VI and VII of the verb in Classical Arabic associated with reflexive, reciprocal, resultative and similar meanings. Most of these meanings still survive in the MA medio-passive form, which is also used to express the passive. Following the classification suggested by Kemmer (1993), this paper discusses four major uses of the medio-passive in MA. These are the passive (e.g. t-hǝrrǝs ‘to be broken’), the reciprocal (e.g. t-fahǝm ‘to get to mutual understanding’), the reflexive (e.g. t-ʕǝllǝm ‘to learn/teach oneself), and the middle (e.g. t-ʕǝʒʒǝb ‘to be surprised’) uses. The description is intended to serve as a basis for future comparison of the medio-passive in MA with its counterparts in other languages and the role diachronic change plays in shaping this grammatical category.

This paper investigates emphatic interpretations of object marking in various Bantu languages. We focus on Lubukusu (spoken in Kenya) and Cinyungwe (spoken in Mozambique) in particular, but we also report initial evidence from other Bantu languages (mainly Tiriki, Wanga, and Ikalanga). In these languages, OM-doubling—the co-occurrence of an object marker with a lexical DP object—is infelicitous in neutral pragmatic contexts. However, we show that certain contexts make OM-doubling possible. In these contexts, OM-doubling constructions receive particular emphatic interpretations that are very different from interpretations of non-doubling object marking constructions. We identify at least four types of these interpretations: verum, mirativity, exhaustivity, and intensity. We show that emphatic interpretations of OM-doubling are widespread among Bantu languages, and we provide strategies for identifying and analyzing them. While this paper is primarily descriptive, we discuss a possible analysis of these interpretations as conventional implicatures, influenced by recent work on a similar range of emphatic interpretations arising from focus fronting in Indo-European languages.

The Gǀui and Gǁana speech communities belong to the Khoe-Kwadi family of Khoisan languages which have not experienced development like other languages in Southern Africa, with the possible exceptions of Nama and Ju|’hoan in Namibia. When the Botswana Government decided to implement the Languages in Education Policy, there was a need to establish the vitality of all local languages with the view to include those that hitherto were not used as a medium of instruction in primary schools. Using historical sources and qualitative analysis of the vitality survey, the paper discusses the challenges faced in developing orthographies for Khoisan languages in general and Gǀui and Gǁana in particular. The paper argues that Gǀui and Gǁana should adopt a common orthographic convention and that there is a need to harmonize such.

This paper analyses borrowing from a sociolinguistic and translanguaging perspective through which it is viewed as a site for linguistic contestation, struggle, and power where some languages benefit while others lose out. The paper problematizes the process of borrowing and presents it as both empowering and disempowering. Empowerment is realized when borrowing is viewed as a translanguaging strategy through which loanwords are seen as representing creativity, dynamism, and the free movement of linguistic elements across language boundaries. Disempowerment is seen in linguistic prestige which causes restrictions in the movement of linguistic elements, language loss, lack of documentation, and insufficient etymologization which all lead to failure to acknowledge the contribution and existence of minority languages in the development of dominant languages.

The infinitive has been quite topical in recent Bantu literature. The majority of Bantuists who have written on this topic sweepingly refer to all ku- forms as the ‘infinitive’ and contend that the infinitive has dual characteristics – clausal and nominal. This paper aims to investigate the types of ku-categories in Ikalanga and to determine whether they all display the same syntactic properties. Using tests from the literature, the paper established that ku-categories which are sweepingly termed ínfinitives’ actually belong to different categories – nominal, poss-ing, and infinitives themselves. The paper also determined that while all three categories identified in Ikalanga display clausal properties, only nominal and poss-ing types display nominal properties. In addition, the paper also demonstrated that while both nominal and poss-ing display nominal properties, they differ in some ways. The findings in this paper suggest that to refer to all ku-categories sweepingly as infinitives is misleading.

This paper follows the analysis of Vansina in analyzing colonization as an attempt to destroy and replace Africa’s autonomous cultural systems. It shows that in Botswana, this has been only partially successful, due to clever forms of resistance. Botswana has been able to keep part of its autonomy intact. This helps to explain the relative success the country has had. However, in the educational field, the country is now also one of the first to be confronted with the limitations that are inherent in the colonial education system. The paper argues that a gradual transition to using indigenous languages as medium of instruction is practically possible and will become inescapable if the country wishes to reach the goals it has set for itself. In this, special attention is needed for the speakers of Khoisan languages.

Verbs of separation and material disintegration are a category of verbs commonly referred to as ‘CUT and BREAK’ (C&B) verbs (Guerssel et al., 1985). This class of verbs has sparked a lot of arguments and discussions in the literature (cf. Guerssel et al., 1985; Levin, 1993; Essegbey, 2007; Lüpke, 2007; Majid et al., 2007; 2008; Hsiao, 2015; Agyepong, 2017). In Nzema, pɛ ‘to cut’ is the prototypical CUT- verb which describes actions that result in object separation/disintegration; usually carried out by using bladed instruments. This paper discusses the semantic extensions of pɛ ‘to cut’ in Nzema discourse. The paper employs the Cognitive Approach of Semantic Change (Sweetser, 1990) in exploring the semantic properties of pɛ, based on the cognitive-conceptual, environmental, and socio-cultural experiences of the people of Nzema. Data were obtained from primary and secondary sources. The paper demonstrates that the various extended usages have a close relationship with the original meaning of the verb. The paper also finds that the primary sense of pɛ ‘to cut’ evokes several contextual interpretations, such as pɛ awolɛ ‘cut childbearing’ (to stop procreation), pɛ azule ‘cut river’ (to travel overseas), pɛ ɔ nli ‘cut his/her mother’ (to insult one’s mother) among others. We note that the various meaning extensions of the verb are dependent on the context of use and/or communication.

This paper investigates linear and non-linear morphology in verbal inflection in Ebira, a Benue-Congo language of Nigeria. Non-linear morphology does not appear on the root, but rather occurs within pre-verbal units called STAMP markers or STAMP portmanteaux, a mnemonic for subject agreement, tense, aspect, mood, and polarity. Based on the previous description of Ebira morphosyntax, this paper decomposes its STAMP markers into several sub-STAMP morphs. Specifically, inflectional categories for subject agreement (1S, 2, 3S, 1P, 3P), aspect/mood (HABITUAL, COMPLETIVE, SUBJUNCTIVE, CONTINUOUS, PERFECT), polarity (NEGATIVE), and clause-level meanings (INTERROGATIVE, 'IF', 'WHEN') are decomposed into various floating tones, floating moras, and unassociated segments. By decomposing STAMPs in this way, it makes testable predictions for future Ebira studies, as not all logically possible category combinations have been investigated yet. This paper concludes by discussing the areality of STAMP markers in West and Central Africa (the so-called Macro-Sudan Belt), which is complicated by analytic vs. synthetic transcription practices by linguists. In order to circumvent issues of transcription, I propose that the areality of STAMP markers in Africa be based on whether the individual STAMP categories form a constituent before and to the exclusion of the verb root, requiring explicit evidence independent from transcription practices.

This paper examines how number is marked in Dagara (Gur, Nige-Congo) and French. It argues that number occupies a syntactic position in Dagara and French and that the difference between the two languages resides in the syntactic operation through which number is marked. Following previous studies, the paper mentions that the status of number in Dagara and French can be assumed to be similar to that in English and that the difference among the three languages is the operation involved in number marking and the syntactic position of number in the structure of nominal phrases. This difference is responsible for the morphological manifestation of number on all the constituents of a nominal phrase in French but on only one constituent in Dagara and English. Besides, the paper claims that the difference between Dagara and English in number marking is in the syntactic position of number and that this difference explains why the plural marker is always on nouns but not on adjectives in English. In contrast, nouns do not take the plural marker in the presence of an adjective in Dagara.

Abstract: The paper describes Leteh nominal morphology within the framework of Basic Linguistic Theory (Dixon 2010; Dryer 2006). The nominal morphology is described in the context of two phenomena: number marking and noun classification. Leteh is a South-Guan language from the Niger-Congo family of languages. The morphology of Leteh is largely agglutinative. Güldemann and Fiedler (2019) argue that current analyses of gender systems are heavily influenced by those in Bantu languages and not cross-linguistically applicable. They propose an alternative analysis that includes the notions agreement class and nominal form class. In this paper I adopt the notion of nominal form class to classify nouns in Leteh. The nouns are grouped into four major classes based on the plural morphemes that they take. These classes are subdivided based on the singular forms with which they are paired. Keywords: Leteh, nouns, morphology, classification, nominal form class, affixes Note: Changes were made to the title and abstract of this article after publication, on 9/20/2021.

Musical surrogate languages like talking drums remain understudied in the linguistics literature, despite their close connection with the phonetics and phonology of the spoken language. African surrogate languages tend to be based on tone, making them a unique angle for studying a language’s tonal system. This paper looks at the encoding of Akan tone in three instrumental surrogate languages: the atumpan drums, the seperewa harp, and the abɛntia horn trumpet. Each instrument presents different organological constraints that could shape how the tone system is transposed to musical form. Drawing on novel data elicited with musicians in Ghana, we show that all three systems are built on a two-tone foundation mirroring the Akan tone system, but with subtle differences in the treatment of downstep and intonational effects like phrase-final lowering and lax question intonation.

Abstract This paper has the primary objective of establishing the pragmatic basis for concluding on a present perfect sentence as currently or continuously relevant and the simple past sentence as currently not relevant. This is an attempt to make the concept current relevance as used in relation to present perfect and simple past sentences in Akan determinate for Akan. Three to four formal tests were conducted to reach the following conclusions: The one primary focus or motivation for making a simple past or present perfect sentence in Akan is to indicate the validity/persistence or non-validity/non-persistence of a past event context as at the moment of speech; context is defined here as non-temporal in the form of some evidence of a past event in some space, and/or as temporal such as the period of time within which an event remains/is valid. In summary, in Akan a past event is currently relevant when context(s) of it in focus is/are current (i.e. persists or is valid as at the moment of speech) - present perfect; and currently not relevant when context(s) of it in focus is/are non-current (i.e. not persistent or valid as at the moment of speech) – simple past.

This paper presents data from Gã (Kwa, Niger-Congo language spoken in Ghana) that show that controlled subjects of non-finite predicates must be overt in this language. The presence of an overt pronominal subject in a non-finite embedded clause is surprising from the perspective of languages such as English and French, where such subjects must be covert (PRO). We provide evidence that the overt pronoun in Gã patterns with obligatorily controlled PRO (Hornstein 1999; Landau 2013) and argue for the minimal pronoun analysis of Kratzer (2009).

Abstract This paper provides a descriptive analysis of some issues concerning the morphology of Yorùbá comparatives and superlatives. Data were obtained from the researcher who is a native speaker of the language as well as from two adults (A man and a woman) in their fifties who are also native speakers of the language. I was able to describe and translate the data using my intuition as a native speaker of the language. The data were grouped into three categories, monosyllabic, disyllabic and trisyllabic adjectives. The implication of the analysis shows that in Yoruba, tone is a distinct marker of the comparative and the superlative. Also, the comparative and the superlative suffix reflects inflectional morphology, and can attach to any tone pattern of base adjectives that are monosyllabic, disyllabic, or trisyllabic. In addition, within the framework of construction morphology, the study provides a word construction schema to account for the comparative and the superlative forms.

Rukiga (Bantu, JE14) generally does not have a dedicated morphological system to express mirativity, that is, speaker’s surprise at an unexpected occurrence. Nevertheless, using elicited data, I show that there are various non-dedicated linguistic tools that the language employs to express the speaker’s surprise. This paper discusses object marker doubling, predicate doubling, and the particles -o and ku, which convey mirativity as a secondary function because they have other various roles they play in the language. Hence, mirative interpretation is context-dependent. The findings indicate that the four strategies discussed can be combined for a reinforced mirative interpretation.

This paper argues for the recognition of mirative marking in two Bantu languages: Gĩkũyũ and Kiswahili. It shows that the two languages use lexical particles to indicate mirativity. Gĩkũyũ uses kaĩ, githĩ, anga, ni, and otho, and Kiswahili has kwani, mbona, kumbe, and si. These particles indicate surprise, unexpectedness, counter-expectation and new realizations, among other attitudes. Mirative marking in the two languages depends on the availability of direct evidence in a context that may be supported by inference. Miratives in Gĩkũyũ and Kiswahili share features with exclamative and interrogative moods. However, the questions are not content questions; they are either polar or rhetorical ones. The interrogative features are more subdued than the exclamative characteristics particularly because miratives and exclamatives share the surprise property. At the end, it is suggested that more research on the pragmatics of the mirative particles and the connections of mirativity, evidentiality and epistemic modality is suggested.

Of Chebanne’s many scholarly contributions, I have been inspired by his work on language use among ethnic minorities. Based on my anthropological research among the G|ui and G||ana, this paper examines the social situations in which their languages are imitated. Through this analysis, I show that not only recording imitable grammar and vocabulary but also maintaining activities and places to use such linguistic resources contributes greatly to the reproduction of the speech community. Furthermore, I will argue that the key to promoting the regeneration of a speech community is to activate language practices not only through formal education but also through informal socialization and nonformal education.

Several homophones –a occur in Naro. In this paper, an attempt is made to disambiguate the functional properties of –a, and also of the stand-alone a. The different uses will be shown, together with examples of each. A distinction is made between uses of–a carrying syntactic meaning, and occurrences where no syntactic meaning of –a can be identified. Differences between the –a suffixes and the a words are shown, followed by some statistical information, plus a transcribed Naro text to show the environments in which these morphemes occur.

The paper builds on the early detection by Carl Meinhof of one or two Arabic loanwords in Nama (Khoekhoe, Khoe), and explores the possibility of other borrowings, from not only from Arabic but also languages of the Cushitic family. A context in which such borrowings could have occurred is provided through the presentation of historical evidence that suggests, for the first time, an actual connection between speakers of Khoekhoe languages – and some of the early Shona states involved in the early Indian Ocean trade along the east coast of Africa, where foreign partners included speakers not only of Arabic but also Persian, Hindi, and Malay. Two sets of additional borrowings that have been found (from Arabic, and Cushitic) are then presented and discussed. Different distribution patterns in the case of borrowings shared with various Ntu languages are argued to suggest more than one stage in the emergence of the modern Khoe languages, while a particular connection is identified, for the first time, with languages specifically of the Mozambican and Madagascan coasts (as opposed to the Swahili coast). This evidence suggests that the earliest interactions behind some of the borrowings occurred during the older phase associated with the southerly hub of the Indian Ocean trade – which allows us to propose, in turn, that the people who came to be known as the ‘Khoi’ were forged by the constantly changing social and economic dynamics of an early African—yet cosmopolitan—world.

The aim of the study was to propose the Automatic Speech Recognition (ASR) technology as a means of improving access to televised and radio broadcasted government communications in English language for people who are deaf and hard of hearing (DHH) in Botswana. To gain better insight into possibilities of using ASR for information accessibility for people who are DHH, this paper evaluated the accuracy rate of the ASR technology in terms of Word Error Rate (WER) by comparing transcriptions generated by ASR software and transcriptions generated and edited by the researcher. The results of our study showed that all the audio-visual documents transcribed scored a WER of less than 15%, indicating that transcriptions were of a very high quality. Some documents even scored an impressive WER of less than 5%, meaning that the transcriptions generated could be used and understood without any human intervention. The study concludes that freely available ASR technology is adapted as a means of improving and diversifying accessibility modes to televised and radio broadcasted government communications in English language for people who are DHH in Botswana.

This paper presents efforts to evaluate the representativeness of the Setswana corpus data with measures that are independent of corpora. Two frequency measures were used: one sourced via a subjective frequency rating survey and another from a corpus of Setswana. Strong correlations (r =.75; p

The endangered language Ts’ixa (Kalahari Khoe), spoken on the eastern fringe of Botswana’s Okavango Delta, is generally assumed to be a member of the Shua dialect cluster. In this paper, I reconsider this classification by taking a closer look at the interconnected domains of nominal gender and the marking of grammatical relations. Using doculect data assembled from both published and unpublished sources, I explore formal and functional aspects of Shua pronominal paradigms and a set of associated clitics commonly referred to as Person-Gender-Number (PGN) markers. While both Ts’ixa and Shua may represent nominal referents by a set of corresponding pronouns, the two languages differ in whether [- animate] nouns possess inherent grammatical gender which is optionally marked on the noun by a PGN clitic: Ts’ixa marks ~70% of noun phrases for gender and number, while in Shua, [- animate] nouns generally do not possess inherent grammatical gender. In both Ts’ixa and Shua, a subset of pronouns and PGNs have two paradigms which are associated with different grammatical roles: while Shua encodes a two-case opposition between subject/object and clausal dependents, Ts’ixa pronouns and PGN clitics display accusative alignment, distinguishing between subject/clausal dependents and objects. In both languages, the object of the clause may further be marked by a postposition (ʔ)à that occurs across Kalahari Khoe and probably grammaticalized from a copula. An additional strategy, indexing of pronominal objects on the verb, is exclusively found in the Deti dialect of Shua. Taken together, the commonalities between Ts’ixa and Shua rarely transcend cross-Khoe patterns, while the differing alignment patterns support a well-delimited genealogical split. Further comparison with Khwe, a Western Kalahari Khoe language spoken to the north and west of Shua-speaking territory reveals a closer link between Khwe and Shua than previously assumed, with Ts’ixa remaining clearly differentiated from both.

Yeyi (Bantu, R41) is an endangered language spoken in northwestern Botswana and northeastern Namibia. Yeyi exhibits two peculiar processes of regressive vowel harmony. The first changes a high front vowel /i/ to a back vowel /u/ when followed by a syllable containing a back vowel /u/, as in ʃi-púndi > [ʃùpúndì] ‘brat’, or /o/, as in ʃi-bowuma > [ʃu-bowuma] ‘kind of snake’, or the glide /w/, as in ʃi-hweta > [ʃuhweta] ‘conversation’. This paper analyzes these two vowel harmony processes in Yeyi, using data from a wide variety of published sources on different Yeyi regiolects. I will show that the use of vowel harmony differs between regional varieties of Yeyi, with certain varieties using vowel harmony in more phonological contexts than others. The diachronic functioning of vowel harmony is also discussed, comparing vowel harmony involving affixes to vowel harmony involving only lexical roots. Finally, a comparative perspective is taken, showing that regressive vowel harmony as used in Yeyi is rarely seen in Bantu languages of Southern Africa, but occurs sporadically in Khoe languages, suggesting that regressive vowel harmony in Botswana may be an areal phenomenon.

In this brief paper we show that the Bantu language Tiania (a variety of Kimeru spoken in Kenya) has a curious gap in the expression of the passive. In the anterior tenses that take an /-eet-ɛ/ inflectional ending, it is not possible to use the otherwise general -w passive suffix (*-eet-w-ɛ). Instead, the inflectional final vowel /-i/ occurs as an anterior-passive “portmanteau”. We decribe and exemplify this gap in the verb system and speculate that *-i was originally a stative that was extended as a “rule of referral” to the passive function it now has with transitive verbs.

This article describes a type of non-verbal predication (NVP) with a single term called "deictic identification" (for example, Abdù nee 'it’s Abdu') and which, in its basic function, is used to identify a referent present in the immediate spatial environment of the speaker. The paper shows that the one-term sentences must be distinguished from ordinary two-term specificational or equative sentences (for example, ùban Muusaa Abdù nèe 'Musa’s father is Abdu' and wannàn Abdù nee 'this is Abdu'). Indeed, the paper in particular shows that when the two types of constructions are used in non-assertive contexts, they can select two different replacive copulas. The paper also shows how the basic one-term deictic constructions acquired extended, non-deictic uses, including uses in focus-fronting constructions, where the immediate external environment is not always relevant. The paper proposes that copula nee/ cee is the sentence predicate in one-term deictic identification and is hence comparable to other one-term non-verbal predicates in Hausa, such as the presentational gàa (for example, gàa Abdù ‘here is Abdu’).

Bare nouns in languages without articles can be semantically ambiguous between definite and indefinite interpretations. It is here assumed that speakers of such languages can still signal to the hearer when they refer to unique and identifiable referents. This paper contributes to the long-standing cross-linguistic question of how bare nouns are interpreted and what means languages without articles have to disambiguate between definite and indefinite readings. This question is largely unexplored for Bantu languages. The answer is sought in the use of different word orders and morphosyntactic constructions, with a focus on the existential in this paper. In many languages of the world, there is a restriction on definites as pivots in existential constructions, serving as a motivation for exploring these constructions in Xhosa. Xhosa makes use of a non-verbal copula in prototypical existentials as well as predicate locatives, to express the existence or presence of a referent. The paper argues that the existential is used for inactive referents and the predicate locative for (semi-) active referents. The inactive referents of the existential are mainly indefinite referential or non-referential. The active referents of the predicate locative are referential indefinite or definite. There is no absolute definiteness effect in the existential. A further motivation for this study is the occurrence of this copula in a short and a long form, giving rise to four different structures. The paper reveals an unexpected analogy between the use of the short and long form and the use of the so called conjoint and disjoint forms in Xhosa tense-aspect paradigms.

This paper presents some of the most prominent properties of Eegimaa, a Jóola/Diola2 language spoken in the Basse-Casamance (Southern Senegal). The phonological features examined include [ATR] vowel harmony, backness harmony, lenition, and Eegimaa’s typologically unusual geminate consonants. Most of the paper, however, focuses on Eegimaa morphology. My analysis of the noun class system separates morphological classes from agreement classes (genders), and presents the most important principles of semantic categorization, including shape encoding. I also show that Eegimaa classifies nouns and verbs by the same overt linguistic means, namely, noun class prefixes. I argue that this overt classification of nouns and verbs reflects parallel semantic categorization of entities and events. Other prominent typological features include associative plural marking and nominal TAM marking with the inactualis suffix, which also expresses alienability contrasts.

This paper serves as an introduction to the special issue of Studies in African Linguistics devoted to conditional constructions in African languages. I first describe the motivation for this volume and the common terminological conventions used in the papers, before discussing some of the more influential attempts to categorize conditional constructions together with some of the functions of conditional constructions. I then present an overview of conditional constructions in African languages, noting the various kinds of conditional meanings that are distinguished grammatically in different languages, and types of isomorphism between conditional constructions and other categories. I conclude with a note on concessive conditionals.

Cross-linguistically, verbs have combinatorial patterns. When the semantics of a verb combines with the semantics of its internal arguments, different interpretations are derived. These interpretations can be literal or non-literal (Ameka 2019; Spalek 2015; Rappaport Hovav 2014; Bobuafor 2013, 2018; Levin and Rappaport Hovav 2013; Ameka and Essegbey 2007; Levin 1993). Using data collected from written texts, video-stimuli descriptions, spontaneous utterances and native-speaker intuitions, this paper explores the potential combinatorial patterns of twá ‘to cut’ in Asante-Twi (Akan, Kwa-Niger Congo). I show that the verb combines with different types of internal arguments, in different argument structure constructions to derive multiple interpretations. Taking into consideration the fact that “natural language tries to minimize polysemy” (Levin and Rappaport Hovav 2013: 2), I propose a univocal lexical semantics for twá ‘to cut’ and show that its basic semantics is kept constant even in non-prototypical contexts. I argue that the use of twá ‘to cut’ in non-prototypical contexts represent contextual modulations of the verb’s single meaning. Following Spalek (2015) and Ameka (2019), I suggest that such contextual interpretations should be analysed compositionally, paying attention to the verb’s internal arguments as well as the argument structure constructions in which it occurs.

In this paper I examine eleven different processes of deverbal nominalization in Runyankore, a Lacustrine Bantu language spoken in Uganda. After establishing both general and Runyankore-specific properties that distinguish nouns from verbs, I test each of these nominalizations against 13 phonological, morphological, and syntactic criteria. Although all eleven nominalization constructions can take the determiner-like initial vowel “augment”, and all can be derived from verb bases that include derivational suffixes (“extensions”), e.g. causative, applicative, and reciprocal, only some of the nominalizations allow a pronominal object prefix or a following noun phrase object or adverbial. The various properties are tabulated to show that the different nominalizations vary along a cline, meeting all, some, or none of the nine most discriminating criteria in defining “noun” vs. “verb”.

In addition to the stem and noun class prefix, the structure of nouns in Bantu languages may contain an augment. This augment typically is a vowel, but some languages show a CV augment. Interestingly, the Bantu language Kinyakyusa shows nouns with a V as well as with a CV prefix, both of which have been analysed as augments (De Blois 1970). In this short paper we clarify the formal and functional properties of the ‘CV augment’ in Kinyakyusa. First we show that it does not behave like the V augment, but is a separate marker that is attached to the noun phrase. Second, we narrow down the previous analyses of the CV marker that describe it as ‘emphatic’ (De Blois 1970, Persohn 2020): On the basis of a range of focus tests, we argue that the CV marker functions as a marker of exhaustivity. This is remarkable, as exhaustive focus is in Bantu languages typically associated with marking in the clause and not on the noun itself.

How languages solve the grammatical problem of agreeing with conjoined arguments is a well-known area of cross-linguistic variation. This paper describes these patterns for Kuria (Bantu, Kenya), documenting a pattern of agreement that has not been previously reported. We show the relevant patterns involving a range of noun classes, showing that human noun classes trigger different effects than non-human noun classes. We also demonstrate distinctions in the grammar between subject marking and object marking: whereas subject marking allows for resolved agreement forms, object marking does not. The paper also includes a brief survey of notable patterns in other Bantu languages to put Kuria in a relevant context.

Ik, a Kuliak (Nilo-Saharan) language spoken by roughly 7000 people in northeastern Uganda, has up to now been described as having a case system with seven cases (König 2002, 2008). However, evidence from morphophonological variations on case suffixes requires that an eighth case be posited. Based on the semantic roles the eighth case marker encodes, it can be called the ‘instrumental’ case. The present paper provides a brief description of this so-named instrumental case in Ik. It does so by first giving an overview of the allomorphy of all eight case suffixes to provide some analytic context for distinguishing the instrumental from other case markers, particularly the ablative. Then it gives examples of the six semantic case roles that the instrumental case suffix encodes. By introducing an eighth case, the paper shows that a supposed instance of case syncretism (between the ablative and the instrumental) should no longer be thought to obtain in this East African language.

Shekgalagadi is an endangered Bantu language of the Sotho cluster spoken in Botswana. While it is known that the language shows extensive regional variation, very little documentation exists of smaller, more remote varieties. This paper provides a first ever description of the northern-most Shekgalagadi variety known as Shetjhauba, spoken along the Okavango panhandle close to the Namibian border. Using original field data, I provide a grammatical sketch of Shetjhauba. First, the segmental phonology of Shetjhauba is described, providing an overview of its phonemic consonants and vowels, as well as a major morphophonological process that occurs in several morphological environments, referred to as “strengthening”. A striking difference between the phonology of Shetjhauba and that of previously described Shekgalagadi varieties is its extensive use of click phonemes. Secondly, the nominal and verbal morphology of Shetjhauba are discussed, giving insights into noun classes, nominal derivation, and various pronouns and agreeing modifiers. Shetjhauba also has an extensive verbal morphology, with various verbal derivational suffixes, and inflectional affixes marking tense, aspect, mood, negation, as well as subject and object.

The interaction of grammatical aspect and lexical aspect has been a topic of interest to linguists for many years, yet only recently has more attention been focused on this topic specifically in relation to Bantu languages. The difficulty of mapping Bantu actional types onto the commonly accepted categorizations has resulted in new frameworks being proposed to take these complex lexicalisations into account. In this paper, we use Croft’s (2012) two-dimensional diagrammatic representations of events to represent the aspectual contour encoded by certain predicates in the South Tanzanian Bantu language Bungu [wun]. We investigate the semantics of the Progressive, Anterior, and Resultative aspects in Bungu, and by identifying the phases within each aspectual contour that are profiled by each of these grammatical aspects, we show that the particular interpretation of each aspect depends on the type of predicate that it modifies. By using Croft’s diagrammatic representations of events to identify the semantic contribution made by different grammatical aspects in Bungu, we illustrate its potential as a descriptive tool.

The paper deals with conjoined subjects and verbal agreement resolution in Akebu and contributes to the typology of noun phrase coordination and agreement resolution in Niger-Congo. Akebu demonstrates both inclusory and compositional coordination with pronominal conjuncts, the former being typical for animate conjuncts, the latter being obligatory for inanimate conjuncts. Semantic person agreement resolution of the verb is possible, and semantic noun class agreement resolution is obligatory with animate conjoined 3rd person subjects, while inanimate conjoined 3rd person subjects trigger syntactic or mixed agreement resolution.

Bemba tonology has been described with respect to two prominent claims: H tone local spreading is binary, and is blocked by the OCP. These claims are based on Bemba, as spoken in Northern Zambia. This paper examines these two claims with respect to contemporary Bemba as it is spoken today in the Copperbelt province of Zambia. This paper shows that in Copperbelt Bemba (CB), these two aspects of H tone spreading are markedly different. In CB, local spreading is ternary, not binary, and a H will undergo binary spreading even if it causes an OCP violation. Ternary spread will be shown to follow from two rules: High Tone Doubling and Secondary High Doubling motivated by different constraints within CB tonology. In addition to documenting and describing the behavior of high tone in CB, a comparison to other cases of ternary spreading is also made.

This paper accounts for the strategies that Yorùbá adopts to mark plural. One way in which plural is marked syntactically is by certain plural words. The plural word can either interpret the noun as plural directly as in the case of àwọn and quantifying words such as púpọ̀ ‘many’ and méjì ‘two’; or it can be realized on a primitive adjective (in the form of COPY) or on a demonstrative (in the form of wọ̀n-). Such elements in turn make available the plural interpretation of the noun they modify. The paper proposes that these plural words possess a covert or an overt [PLURAL] feature, which percolates onto the NP. This analysis of plural marking predicts that there are two ways by which languages may (overtly) mark their nouns for plural cross-linguistically. Languages like Yorùbá, which do not show agreement, mark plural syntactically and make use of a plural feature percolation mechanism, while languages like English, which show agreement, mark plural morphologically and use a plural feature-matching mechanism. It further demonstrates that in Yorùbá, an NP can be freely interpreted as singular or plural in specific discourse context and proposes a general number analysis to account for this type of case. As to the syntax of these plural words, It is proposed that like other non-morphological plural marking languages (e.g., Halkomelem (British Columbia, Canada) as in Wiltschko 2008), Yorùbá plural words are adjuncts that are adjoined to the host head (noun or modifier/demonstrative).

This paper presents an analysis of the verbal morphology and associated phonological processes in Asante Twi, a member of the Akan group of languages/dialects spoken in Ghana and Côte d’Ivoire, which belongs to the Nyo subgroup of the Kwa language family (Lewis 2009). There has been considerable interest in Akan in the theoretical literature, largely due to some peculiarities in the tense/aspect system which will be addressed later in this paper. However, the verbal morphology and phonology have been given relatively little attention. In this paper I show that the verbal morphology exhibits a number of interesting properties including tonal marking of tense/aspect categories – the latter having been largely ignored or misrepresented in the previous literature.

Separation events differ in lexicalisation patterns (Talmy 2000) and in argument realisation (Levin and Rappaport Hovav 2005) cross-linguistically. There are different types of separation events. “Cutting” and “breaking” events involve a non-reversible change in object integrity and have been systematically researched cross-linguistically in recent times (Guerssel et al 1985; Bohnemeyer 2008; Majid et al. 2008; Schaefer and Egbokhare 2012). In this paper, some of the generalisations that have been made concerning CUT and BREAK verbs are tested based on data from Tafi, a Ghana-Togo Mountain language. I investigate the morpho-syntactic properties of Tafi CUT and BREAK verbs in relation to a suggested generalisation by Guerssel et al. that BREAK verbs have a transitive/intransitive argument structure and participate in the causative/inchoative alternation; while CUT verbs are transitive and they are not expected to occur without their external argument. The types of events referred to by the CUT and BREAK verbs and the combinatorial capacity of the individual verbs are also explored. Based on an analysis of stimulus-elicitations and spontaneous language performances recorded in the field, I show that the Tafi verb bhui ‘cut’ can be used in an intransitive/resultative construction in which the theme, the internal argument, occurs as the subject. Drawing on the behaviour of bhui ‘cut’ I interrogate the explanations that have been offered in the literature with respect to such deviations from the generalisation. I argue that the verb argument alternation potential of a verb depends on the verb semantics as well as the type of (internal) argument it collocates with. Moreover, I explore the semantic interpretations of the verb when it combines with non-typical objects such as ‘water’. I show that such patterns and collocations such as ‘the water cut’ = ‘the water stopped running’, ‘cut a village’ = ‘establish a village’ are areal in nature (cf. Huttar et al. 2007).

This paper presents an analysis of the system of argument marking on the verb in Zaghawa. Zaghawa, also called Beria in the literature, is a Saharan language of the Nilo-Saharan language phylum spoken in the border region of Sudan and Chad. Like other Saharan languages, it has complex verbal morphology including person indexing. The primary aim of the study is descriptive in that it presents linguistic data of the underdescribed Wagi dialect which is mainly spoken in Sudan. First, the paradigm of bound verbal affixes and their morphology is described. Secondly, one of the functions of the final morpheme of the verb which has not yet been described in detail in previous studies on Zaghawa is analysed. This final morpheme interacts with the person indexes to mark plural participants, and it is exploited to mark a morphological category not yet recognized in the other dialects of Zaghawa: the exclusive/inclusive distinction in the 1st person plural. Therefore, the study provides new data on the Zaghawa verb system and contributes a further detail to our knowledge of the Nilo-Saharan language family.

While previous studies on Dàgáárè tone have looked at the nouns, this paper particularly examines tone in verbs, perfective vs imperfective forms. The verbal system has different patterns based on the form of the verb. There are three tone classes for Dàgáárè verbs and for each of the classes, the surface tone pattern it exhibits in the perfective is systematically different from the tone patterns in the imperfective. For the perfectives we have L, H and HL while the imperfectives have LH, HL and H!H, at least in the dialect under study. I treat tone as a combination of the features [±upper] and [±raised] which are connected to what is described as a Tone node (T-node). These Tone nodes in turn connect to the syllable. Under this system, I assume L is represented with the features [-upper] and [-raised] and H with the features [+upper] [+raised]. Underlying tonal melodies of the root morphemes are identical to the surface tones of the perfective forms whether these contain an overt suffix or not. For the imperfectives, the suffix comes with an unspecified underlying T-node. The grammar then chooses the features [±upper] and [±raised] to insert under the already existing T-node.

This paper presents some key findings of studies of actionality and the verbal grammar–lexicon interface in two Nguni Bantu languages of South Africa, Xhosa and Southern Ndebele. We describe interactions between grammatical tense marking (and other sentential bounding elements) and lexical verb types, arguing for the salience of inchoative verbs, which lexically encode a resultant state, and, in particular, a sub-class of inchoative verbs, biphasal verbs, which encode both a resultant state and the “coming-to-be” phase leading up to that state. We further discuss other important features of actional classes in Xhosa and Southern Ndebele, including topics such as the role of participant structure and the relative importance of cross-linguistically prominent distinctions such as that between Vendlerian activities and accomplishments. Although differences between Xhosa and Southern Ndebele are evident both in the behaviour of individual tense-aspect forms and in the interpretive possibilities of specific verbs, the general patterns are quite similar. This similarity suggests that the patterns are likely to extend to other Nguni languages, as well, and that cross-linguistic comparison of particular lexical items across these languages are both feasible and likely to bear fruit. Note: Changes were made to the title of this article after publication, on 9/23/2021.

This paper presents a descriptive overview of the phonology and morphology of the Lower Jubba dialect of Maay, a language of southern Somalia. The paper highlights several points of typological, dialectological, and theoretical interest in this language. For example, the nominal morphology exhibits a somewhat unusual pattern of plural marking that interacts in a complex way with the gender marking system. Where relevant, comparisons are made between this dialect and other dialects of Maay and Somali, and speculation is made as to the historical origin of some of the unusual phonological and morphological phenomena observed in this dialect.

As the largest language phylum in the world and the most geographically widespread (Williamson & Blench 2000), Niger-Congo understandably exhibits some variation at all grammatical levels. Basic word order stands as no exception to this generalization, and there have been partisans for both an SOV and an SVO reconstructed word order. Gensler 1994 attempts to reconcile the two by claiming that neither proposal is correct; he suggests that both SOY and SVO are derived from Proto-Niger-Congo *S-AUX-O-V-Other. Because of the pattern's "quirkiness" (being found virtually nowhere else in the world) and because it is so widely attested in geographically widely separated Niger-Congo languages, the pattern should be reconstructed for all of Niger-Congo. One crucial piece of evidence for this claim comes from the Southern Atlantic language Kisi. This paper explores Kisi' s facts in further detail to show how central the structure is to the language. It then expands the investigation to other languages of Atlantic, finding that the pattern is much more widely attested than was previously realized, albeit in an attenuated form. The paper concludes by discussing the significance of the Atlantic facts to Niger-Congo in general.

This paper investigates the controversial morphemes that occur both at the clause level and within the DP in Bantu languages such as Ikalanga, which have been analyzed as "agreement morphology" by some (Baker 2002, Demuth and Harford (1999) and as resumptive/incorporated pronouns by others (Bresnan and Mchombo 1987, Zwart 1997). The paper proposes a unified analysis of this phenomenon, analyzing both clausal and DP-internal occurrences of these morphemes as agreement morphology which holds between the head of an XP and its predicate, or between the head of an XP and its modifiers. In both instances, the agreement relation is instantiated after movement of the relevant category from its base position to a specifier position which enables the moved category to enter into a checking relation (i.e. a spec-head configuration) with another category that has matching features (Chomsky 1995).

It is common knowledge that noun classes in Mabia (Gur) languages are primarily characterized by stems and affixes. Common to all studies on nominal classification in Kusaal is the observation that nouns that exhibit common morphological properties also share identical semantic features. Though this is true to some extent, the generalization breeds a lot of leakages because classifications based on semantic field alone is unable to explain the inclusion of nouns that share identical morphological and phonological features but different semantic features. Thus, this problem questions the assumption that noun classification in Kusaal is dependent on common semantic properties or coherence shared by all nouns in a group. The semantic classification of nouns, in this study, is composed based on the assumption that speakers of Kusaal put together nouns that are connected by identical semantic features and others that are linked by pragmatic associations into networks that define concepts and aspects of their survival as human beings. It is further observed that nouns within such groups also go through identical phonological rules or constraints. Nouns in this paper are classified based on their morphological features which are closely knit to their semantic networks as well as phonological constraints. The framework of Lexical-Phonology is employed in analyzing the morphophonological components of the nominal classification system of the language

This paper examines the imbrication process in Lunda. It is widely acknowledged in the literature that imbrication in Bantu languages is mainly triggered by the perfective or past tense suffix (Bastin 1983). In Lunda, however, this morphophonological process is not only conditioned by the remote past tense inflectional suffix but also caused by derivational extensions that include the applicative, the causative, and to a lesser extent the reciprocal affix when they are attached to a large number of verb bases. Imbrication typically occurs with verb bases containing what is known as an extension, whether productive or frozen. Unlike many Bantu languages, this phenomenon always results in the deletion of the base final consonant and the consonant of the suffix survives. This segmental loss leads to either vowel coalescence between the two adjacent vowels or gliding

This paper investigates the interaction of focus and negation in the Bantu language Zulu (Nguni; S42). I discuss four strategies that are used to negate transitive sentences in Zulu. The default strategy, in which an object marker is added to the negated verb, expresses polarity focus by dislocating the object-marked object from the VP-focus domain. In the second strategy, no object marker occurs, and focus falls on the object or the VP. I show that in this strategy, negation typically associates with the focus and is not part of the presupposition, and I argue that this is responsible for a (hitherto unexplained) additional contrastive inference that speakers report with this negation strategy. The third strategy, a cleft, is used to remove the focused object from the scope of negation; as a result, negation can associate with the presupposition. In the fourth strategy, the object noun loses its augment and is interpreted as a negative polarity item (NPI). Based on a proposal by Lahiri (1998), I argue that in negated sentences with NPI-objects, focus is placed on an implicit cardinality predicate which is associated with the semantic representation of the indefinite NPI-object.

This paper investigates the principles that govern subject marking in Awing (Grassfields Bantu). We observe that the subject marker (SM) that doubles the subject is sometimes obligatory, sometimes optional and sometimes prohibited. We argue that it is the referentiality of the subject that controls the distribution of the SM in Awing, rather than factors such as its morpho-syntactic features or its information structural status, which have been identified to govern argument doubling in a number of other languages with a similar phenomenon. The empirical evidence leads us to conclude that the SM is a pronominal element in Awing rather than an agreement marker. When it occurs, it functions as the argument of the verb and the associated subject NP is base-generated in the left periphery of the clause; when it is absent, the NP is the verbal argument. Awing thus qualifies as a pronominal argument language in the sense of Jelinek (1984); Bresnan and Mchombo (1987); Baker (1996).