The accidental intransitive split in the Cariban family
The Cariban language family consists of 30–60 languages spoken in lowland South America, most of which are chronically underdescribed (cf. Gildea 1998:3–14). For several of those which have received some attention, an analysis of the verb system as split-S (in Dixon’s 1979/1994 terminology) has been proposed (beginning with Hall 1988 for De’kwana, followed by Tavares 1994 for Wayana, Gildea 1994a for Katxuyana, Apalaí, Kari’ña,1 Hixkaryana, and maybe Tamanaku; more recently, Gildea (pc) has suggested this analysis for Panare; Gildea 1998:88–96 reconstructs the system back to Proto-Cariban).
The present study is a first attempt at a critical evaluation of these claims by means of a closer look at data from the following Cariban languages: Apalaí (Koehn & Koehn 1996; my field notes), Kari’ña (Hoff 1968, Mosonyi 1978), Tiriyó (my field notes), Wayana (field notes by Tavares and me), Makushi (Abbott 1991, Williams 1932), Katxuyana (field notes by Gildea and me) and Panare (Mattéi-Muller 1994). As we shall see, the semantics of the two intransitive verb classes do not seem to correlate with any of the factors that have been claimed to be relevant in the literature. The only clear correlation is that between the SA (or ‘active’) category and detransitivization: almost all verbs in the SA class are detransitivized forms of transitive verbs, either synchronically (with still exisiting transitive sources) or diachronically (with reconstructible but no longer existing transitive sources). The inflectional morphology associated with the SA verb class thus appears to be historically connected to reflexive morphology, in ways that are not entirely clear yet. The semantic anomalies of the ‘SA’ class become more understandable if seen as the result of the evolution of the original reflexive semantics into new areas (middle voice, passive, antipassive, etc.), extensions which are well-attested outside of the Cariban family (e.g. reflexives in Romance and Slavic). This leads to the conclusion that the modern split-S verb agreement pattern is wholly epiphenomenal, a casual consequence of the historical development of detransitivizing morphology in the family.
In what follows, the basic parameters of the split-S systems found in Cariban languages are described and exemplified, with data from Apalaí, Wayana, Tiriyó, and Kari’ña (Sec. 2). The intransitive verb classes are examined for semantic consistency, based on the different semantic parameters that have been proposed thus far (Sec. 3); after that, the connections between the split-S system and detransitivization are explored in detail (Sec. 4). The synchronic and diachronic consequences of detransitivization in Cariban are treated in Secs. 5 and 6.
