On Agree without agreement as a source of reflexive voice constructions

Switch-reference has recently been argued to be the result of clausal functional heads entering into Agree with two nearby noun phrases, creating pointers to those noun phrases but not actually copying their morphosyntactic features. Instead, the semantic component interprets the pointers as referential dependency holding between pointed-to noun phrases. This article applies this analysis to reflexive voice constructions in which a feature-invariant affix appears on the verb to indicate that the (highest, direct) object is referentially dependent on the (thematic) subject of the same clause. First it surveys the properties that such constructions should have if reflexive voice is maximally like switch-reference. Then it argues that the Bantu language Lubukusu has just such a construction, the verbal affix i partnering with the overt anaphor omweene to create reflexive clauses. Dravidian reflexive voices are presented as another possible case. Finally, it turns to reflexive and reciprocal voice constructions in Shipibo (Panoan), which seem to have a detransitivizing effect. However, no major detransitivizing account fits all the facts. Rather, reflexive voice in Shipibo is like Lubukusu, except that the anaphor is phonologically null and deficient in phi-features, failing to trigger ergative case on the subject for that reason. True detransitivization may happen in some languages with reflexive voice, but not in all, and it will take considerable care to sort out which are which.

Verbal Extensions and Valency in Limbum

The objective of this paper is to examine the extent to which verbal extensions in Limbum affect valency. Limbum is a Grassfield Bantu language of the Northern group, spoken by the population who occupy a greater part of the Nkambe plateau in Donga-Mantung Division of the North West Region of Cameroon Binan Bikoi (ed) (2012). To attain my set objective, I carry out an analysis of those affixes (in the case of Limbum, they are suffixes), which are usually attached to verbs and the effect the addition of these suffixes has on the number of arguments in the sentence. Some of these suffixes have a valency decreasing effect, while some have a valency increasing effect on the verbs. The orientation of the discussions here centres on the description of the morpho-syntactic structure of the Limbum verb. In this regard, the analysis herein draws inspiration from the theory of Valency as proposed by Tesnière in 1959 and his followers and the Structuralist Framework as propounded by De Saussure and his disciples who hold that linguistic unit: words, phrases and sentences are perceived as a concatenation of smaller units which hold a close relationship between them. The structure of the Limbum sentence containing verbal extensions that express aspectual meanings have three consequences on the number of arguments that the verb takes: the discussions here show that, while the morphemes -ri, -Si, and -se marking the attenuative, the pluractional, and the distributive aspects respectively have no effect on the number of arguments taken by the verb to which they are suffixed, the causative morpheme -si, has a valency increasing effect on the verb to which it is affixed. In the same light, the reciprocative -ni, the separative -ti and the iterative -Nger, when suffixed to a verb, have the tendency of increasing the number of arguments that the verb takes. In a bid to clarify the structural cartography of verbal extensions in Limbum, the last part of this paper is dedicated to a presentation of some suffixes like -ri and -si, which has, with the evolution of the language, fossilized with the verb root to the extent that they have become an integral part of the verb in a way that they cannot be detached from each other. Conclusively, the paper shows that verbal extensions in Limbum are, for the most part, suffixal morphemes. While some of these suffixes have no effect on the number of arguments the verb subcategorizes for, some have a valency decreasing effect on the verb while others, on the other hand, have a valency increasing effect. Others have outrightly merged with the verb root.

Final Vowel Loss in Lower Kasai Bantu (drc) as a Contact-Induced Change

In this article, we present a qualitative and quantitative comparative account of Final Vowel Loss (fvl) in the Bantu languages of the Lower Kasai region of the Democratic Republic of the Congo. We argue that this diachronic sound shift rose relatively late in Bantu language history as a contact-induced change and affected adjacent West-Coastal and Central-Western Bantu languages belonging to different phylogenetic clusters. We account for its emergence and spread by resorting to two successive processes of language contact: (1) substrate influence from extinct hunter-gatherer languages in the center of innovation consisting of Bantu B80 languages, and (2) dialectal diffusion towards certain peripheral Bantu B70, C80, H40 and L10 languages.

The evolution of the noun class system of Ngwi (West-Coastal Bantu, B861, DRC)

In this paper we offer a first systematic account of the noun class system of Ngwi, a West-Coastal Bantu language spoken in the Democratic Republic of the Congo. First, we describe the synchronic system of noun class prefixes and the agreement patterns they trigger on constituents of the noun phrase and the verb. Second, we provide a diachronic analysis of the innovations the synchronic Ngwi noun class system underwent with respect to the noun class system reconstructed for the most recent common ancestor of all Narrow Bantu languages. Finally, we compare the morphological innovations found in the Ngwi noun class system with those identified in the noun class systems of other West-Coastal Bantu varieties and assess whether some of these could be diagnostic for internal classification within this western Bantu branch.

CARP Affix Ordering: Problematic in Lubukusu

Hyman and Mchombo (1992), Babye (1985), (Myers 1987), (Rice 2000), and Hyman (2002 & 2003) have shown that there is affix ordering in Proto-Bantu languages that obeys the 'CARP' (Causative-Applicative-Reciprocal-Passive template). Drawing data from Lubukusu, a Bantu language, the current study analyzes affix ordering of class-changing morphemes, arguing against the templatic morphology that most researchers have shown to be dominant in Bantu languages. The current study uses Bybee’s (1985) principle of iconicity (principle of relevance), where it is proposed that affixes closer to the verb stem are more 'relevant' to the verb than to the rest of the sentence and those affixes further away are less relevant. Based on Baybee’s relevance principle, the study argues that there are various affix ordering orders in Lubukusu, which are semantically motivated. The data that are used in the analyses are self-generated and verified by three native Lubukusu speakers who are competent in the language. Findings show that as much as Lubukusu obeys the templatic morphology, the same is violated in various morpho-semantic contexts. The study recommends more studies on affix ordering in the Lubukusu language based on other existing frameworks that have been tested on languages rather than those from the proto Bantu family.

Sequences of high tones across word boundaries in Tswana

The article analyses violations of the Obligatory Contour Principle (OCP) above the word level in Tswana, a Southern Bantu language, by investigating the realization of adjacent lexical high tones across word boundaries. The results show that across word boundaries downstep (i.e. a lowering of the second in a series of adjacent high tones) only takes place within a phonological phrase. A phonological phrase break blocks downstep, even when the necessary tonal configuration is met. A phrase-based account is adopted in order to account for the occurrence of downstep. Our study confirms a pattern previously reported for the closely related language Southern Sotho and provides controlled, empirical data from Tswana, based on read speech of twelve speakers which has been analysed auditorily by two annotators as well as acoustically.

The robustness of Botne and Kershner aspectual classes in Nyamwezi

The framework proposed in the works of Robert Botne and Tiffany Kershner has been widely used to classify verbs in Bantu languages. In this framework, verbs encode events which consist of maximally three phases: onset (represents the coming-to-be phase), nucleus (represents the state change itself; can also be represented as a coming-to-be phase if the verb lacks an onset) and coda (represents the result-state phase). Hence, verbs are defined depending on which phases they encode and whether particular phases are punctual or durative. The phasal structures of verbs can be diagnosed using various tests. The application of these diagnostics to Nyamwezi (a Tanzanian Bantu language, [nym]) produces three significant variations. First, Botne and Kershner’s conception of statives as events with no phasal structure is not tenable in Nyamwezi. The tests show that in Nyamwezi, statives have structure. Second, some classes described in Botne and Kershner do not occur in Nyamwezi. Third, in Botne and Kershner’s works, classes are described depending on whether particular phases are punctual or durative. In addition to this characteristic, the classes in Nyamwezi can also be described depending on whether particular phases are dynamic or static, and whether the result state is permanent or reversible.

On the aspectual system of Wushi (Babessi), a Ring Grassfields Bantu language of Cameroon

One of the characteristics of Bantu languages, including Grassfields Bantu languages, is their multiple time distinctions. Within the Ring Grassfields group, multiple tenses are also well attested. For example, Aghem has three past and two future tenses (Anderson 1979), Babanki has four past tenses and three future tenses (Akumbu & Fogwe 2012), as well as Lamnso’ (Yuka 2012). Oku has three past tenses and two future tenses (Nforbi 1993) and Babungo has four past and two future tenses (Schaub 1985). These tenses represent different degrees of remoteness in time such as hordienal, immediate, distant, etc. However, in spite of the indisputable lexical unity of Ring Grassfields Bantu languages (Stallcup 1980; Piron 1997), Wushi strikingly stands apart: it does not mark tense morphologically. As a matter of fact, the aspectual system of Wushi is based on five aspects: perfective, imperfective, retrospective or anterior, potential, and the distal or dissociative marker kə̀ that is analyzed in the light of Botne & Kershner (2008). This paper sets out to analyze these verb forms.

African elements in Saramaccan Maroon plant names in Suriname

The ancestors of the Saramaccan Maroons, who were brought as enslaved Africans to Suriname, used their ethnobotanical knowledge and native languages to name the flora in their new environment. Little is known about the influence of African languages on Saramaccan plant naming. We hypothesized that Saramaccan plant names were more influenced by Central African languages than found so far based on ethnobotanical research, because data of the Central African region was scarce. We compiled a new database on Saramaccan plant names and compared these names with an unpublished plant name database from the Democratic Republic of the Congo and the earlier published NATRAPLAND database on Afro-Surinamese plant names to find comparable plant names for botanically related species in Africa. We further analyzed form, meaning, function, and categories of Saramaccan plant name components by means of dictionaries and grammars. In total, 39% of the Saramaccan plant names had an African origin, of which 44% were African retentions, 54% were innovations and 2% were misidentifications with botanical links to Africa via other plant species. Most retentions were of Central African origin (62%). The Bantu language that contributed most to Saramaccan plant names was Kikongo, followed by West African Kwa languages. Plant names reveal important information on the African origin of the Saramaccans, and deserve more scientific attention.

Comprehension and production of Kinyarwanda verbs in the Discriminative Lexicon

How do speakers comprehend and produce complex words? In the theory ofthe Discriminative Lexicon this is hypothesized to be the results of mapping the phonology of whole word forms onto their semantics and vice versa, without recourse to morphemes. This raises the question whether this hypothesis also holds true in highly agglutinative languages, which are oǒten seen to exemplify the compositional nature of morphology. On the one hand, one could expect that the hypothesis for agglutinative languages is correct, since it remains unclear whether speakers are able to isolate the morphemes they need to achieve this. On the other hand, agglutinative languages have so many different words that it is not obvious how speakers can use their knowledge of words to comprehend and produce them.In this paper, we investigate comprehension and production of verbs in Kinyarwanda,an agglutinative Bantu language, by means of computational modeling within the theDiscriminative Lexicon, a theory of the mental lexicon, which is grounded in word andparadigm morphology, distributional semantics, error-driven learning, and uses insightsof psycholinguistic theories, and is implemented mathematically and computationallyas a shallow, two-layered network.In order to do this, we compiled a data set of 11528 verb forms and annotated for eachverb form its meaning and grammatical functions, and, additionally, we used our dataset to extract 573 verbs that are present in our full data set and for which meanings ofverbs are based on word embeddings. In order to assess comprehension and production of Kinyarwanda verbs, we fed both data sets into the Linear Discriminative Learningalgorithm, a two-layered, fully connected network. One layer represent the phonological form and the layer represents meaning. Comprehension is modeled as a mapping from phonology to meaning and production is modeled as a mapping from meaning to phonology. Both comprehension and production is learned with high accuracy in all data and in held-out data, both for the full data set, with manually annotated semantic features, and for the data set with meanings derived from word embeddings.Our findings provide support for the various hypotheses of the Discriminative Lexicon:Words are stored as wholes, meanings are a result of the distribution of words in utterances, comprehension and production can be successfully modeled from mappings from form to meaning and vice versa, which can be modeled in a shallow two-layered network, and these mappings are learned in by minimizing errors.