Search

In Mabaan, a Western Nilotic language, indirect speech is characterised by the following three features. Firstly, verbs are inflected in a special way, having an indirect mode as opposed to the direct mode of direct speech. Secondly, in pronominal morphemes a distinction is made between forms that refer to a third person reported speaker (third person forms) and forms which have other third person referents (fourth person forms), a distinction which is coextensive with the use of the indirect mode. Thirdly, third person pronominal forms are identical to the third person pronominal forms used in direct speech, while fourth person pronominal forms are different and, thus, anti-logophoric. Taken together, these features make Mabaan typologically unusual.

In Dinka, a Western Nilotic language, most adnominal modifiers follow the head noun. Before most of these modifiers, the head noun is in one of two construct states. One construct state, CS1, occurs before, among others, demonstratives, nominal possessors and relative clauses as CS1-modifiers. The other construct state, CS2, which is morphologically more complex, occurs before, among others, possessive pronouns, a few numerals and a diminutivizer as CS2-modifiers. When a construct-state triggering modifier is added to a CS2-modifier, the latter itself gets construct state marking, and the head noun changes from CS2 to CS1. Some CS1-modifiers also get construct state marking when followed by a construct-state triggering modifier. Multiple adnominal modification in Dinka may thus result in a chain of construct states, which is similar to what is found in Iranian languages with so-called ezafe marking.

The case-number suffixes of the Western Nilotic language Nuer (Frank 1999) display a remarkable combination of formal simplicity and distributional complexity, which is manifested in: (i) a seemingly erratic form-function mapping that precludes attributing a consistent meaning to the suffixes, and (ii) a wealth of inflection classes only barely differentiated from each other. The suffixes looks as if they were rule-generated, but behave as if they were memorized. I advance a model of inflection combining principal parts, implicational rules, and default inheritance, in which the bulk of the complexity is attributed to the lexical stem, revealing the underlying systematicity behind suffix assignment.

In Mabaan, a Western Nilotic language, there are several singular and plural number suffixes. However, many singular nouns have no segmentally distinct suffix, but have a root-final nasal which corresponds to a root-final homorganic plosive in the plural. The plosive is the original alternant, as can be determined on Mabaan-internal grounds as well as by a comparison of Mabaan with other Western Nilotic languages. The nasal alternant of a root-final plosive is a reflex of a former singular suffix with an alveolar nasal. As evidenced by Mayak, another Western Nilotic language, this suffix was an innovation, and it came to be used both in nouns that had no number suffix originally and as an additional suffix in nouns with a singulative suffix. A similar development seems to have taken place in plural nouns, in that a velar plosive suffix is now found in many plural nouns that were presumably morphologically unmarked originally. There is evidence that the new suffixes are Stage III articles. Thus, there are two historical layers of number inflection in Mabaan, a phenomenon somewhat reminiscent of the situation in the Southern Nilotic languages.

Dinka, a Western Nilotic language, has a class of auxiliary verbs which is remarkable in the following four respects: (i) It is unusually large, comprising some 20 members; (ii) it is grammatically homogeneous in terms of both morphology and syntax; (iii) most of the auxiliary verbs correspond to adverbs in languages like English, while the rest are tense-aspect markers; and (iv) it is possible to combine several auxiliary verbs in a single clause. For some of the auxiliary verbs there are extant full verbs from which they have evolved. To some extent, therefore, it is possible to observe what semantic changes the auxiliary verbs have undergone or are undergoing during the process of grammaticalization.

Jumjum, a Western Nilotic language, has an eight-vowel system divided into two sets by the feature [ATR] (Advanced Tongue Root), which is the basis of vowel harmony. A comparison with other Western Nilotic languages shows that (i) this vowel system goes back to a ten-vowel system in Proto-Western Nilotic (PWN), (ii) PWN high [−ATR] vowels have become high [+ATR] vowels in Jumjum, and (iii) conversely, PWN high [+ATR] vowels have become high [−ATR] vowels in Jumjum. The sequence of changes that resulted in this [ATR] reversal in Jumjum relative to PWN provides a historical explanation of synchronically odd, grammatically conditioned vowel-quality alternations in this language.

Pari, a Western Nilotic language, has a terraced-level tone system with total downstep. Although Pari could be analyzed as having three basic tone levels and automatic downstep, there is morphological evidence that it has two basic tone levels and non-automatic downstep. Furthermore, there is evidence that downstep is the manifestation of a floating high tone. Floating tones thus behave differently from tones of deleted vowels. In spite of many surface differences between Pari and Luo, a related language, a single tone change accounts for their underlying differences.