- Rotuman and Fijian Case-Marking Strategies and Their Historical Development1
Although the Fijian languages and Rotuman are thought to be closely related genetically, and are all accusative languages, considerable differences are observed in their case-marking strategies. This paper describes the various strategies to be found in these languages, and discusses how they could have developed from a single protosystem. It is argued that in Rotuman, where it is word order alone that marks nouns as Nominative or Accusative, the preverbal position of clitic pronouns was generalized to become also the position of full noun phrases. On the other hand, in Fijian languages, different strategies resulted in the original clitic pronouns either remaining as clitics, or becoming grammaticalized as agreement features on the verb.
1.1 An Overview.
The Fijian languages and Rotuman are considered to belong to the Central Pacific language family and to be closely related to one another.2 Although each of these languages shows an accusative case-marking system, considerable differences are found in their case-marking strategies, that is, how nouns are formally marked (or unmarked) as Accusative and Nominative. The Rotuman case-marking system is relatively simple. As described in section 2, it is the word order that nominatively and accusatively marks nouns. There is no [End Page 85] morphological marking on the core nouns of a sentence.3 On the other hand, a variety of case-marking strategies are found in the Fijian languages. In some, nouns are never casemarked either morphologically or by their position. The "doer" and the "recipient of the action" are understood through agreement on the verb, or other nonformal aspects, such as the meaning of the verb, and/or the context. In other Fijian languages, whether a noun is casemarked or not depends on the nature of the noun, namely, whether it is pronominal or nonpronominal, and when nonpronominal, whether it is proper or nonproper. The different case-marking strategies observed in the Fijian languages are described in section 3. How Rotuman and the Fijian languages have developed these different systems from a single protolanguage is of interest from both theoretical and historical points of view, and is discussed in section 4. It is proposed that Rotuman has developed its Nominative position from what originally was the clitic pronoun position, while the Fijian languages have developed the original clitic pronoun position into the verb agreement system. Conditions and motivations that enabled the various changes to take place are also discussed. Section 5 is a summary.
1.2 Terms and Definitions.
The definitions of the basic notions and the terms that appear in this paper are described in this subsection. The original analysis was conducted applying a version of the Word and Paradigm model, or "whole-word morphology,"4 and Lexicase Dependency Grammar,5 and some of the terms used in this paper follow those of Lexicase.
The judgment of the transitivity of a verb in this study is based solely on the number of core noun phrases a verb may take. The core noun phrases in Rotuman and Fijian sentences appear as "bare noun phrases," that is, without being preceded by determiners or prepositions that could assign case.6 When a verb may take two of these, it is considered to be transitive, while when a verb can take only one, it is considered to be intransitive. An intransitive verb may take a complement noun phrase preceded by a case-assigning preposition in addition to the single bare noun phrase.7 The traditional transitive and intransitive classifications of Rotuman and Fijian verbs are often based on verb morphology, and do not always match the judgment of syntactic transitivity used in this study.8 The differences between the traditional analyses and the one in this study are mentioned in each section when it is relevant.
1.2.2 Case forms.
Case forms are labels that describe the morphological and/or position marking on nouns and prepositions. Nouns or prepositions with the same [End Page 86] marking are always labeled in the same way. The total number of case forms in each language differs, depending on the language. Those that...