'Supposition' as an Evidential Category in Wanano and Other Eastern Tucano Languages

Kristine Stenzel

Abstract for paper presented at
XVII International Congress of Linguistics
Prague
July 24-29, 2003

The study of 'evidentials'--grammatical markers whose primary function is to indicate the speaker's source of information for a given statement--has been greatly enriched by the study of lowland Amazonian languages, particularly those of the Tucano family, which code up to five evidential categories and are recognized as displaying some of the most complex evidential systems analyzed to date.

This paper examines a rare evidential category found in Wanano and other Eastern Tucano (ET) languages, that of coding SUPPOSITION, and has three objectives: to present 1) the semantic territory covered by the category in the Wanano system;
2) a cross-linguistic comparison of the category throughout the Tucano family; and 3) the theoretical questions that arise from the investigation of this category and its diverse functions.

Like all ET languages, Wanano requires evidential coding as the final suffix on the final verb in the matrix clause of a realis statement, the minimum obligatory constituents of a matrix verb being Verb root + evidential suffix. There are five categories in the system: VISUAL, NON-VISUAL, INFERENCE, HEARSAY, and SUPPOSITION.

The SUPPOSITION evidential category in ET languages leads to some interesting questions for investigators of evidentiality. First, we find overlapping semantic functions coded by this category throughout the family. In Wanano, SUPPOSITION markers indicate that the speaker's assessment of a situation is based not on accessible first- or secondhand evidence, but on internal evidence: the speaker's cultural, historical, or physical knowledge of the world or the speaker's own emotions, bodily sensations, and cognitive processes.

Second, these very semantics lead to a questioning of the 'evidential' nature and epistemic value of SUPPOSITION. The category is viewed here as a rare example of the grammaticalization of a specific type of inference, that involving conclusions based on association rather than on observable results (alternately coded by the INFERENCE evidential category). The existence of such a category not only reflects the scalar nature of the semantics involved, but also suggests the possibility of 'gray areas' between evidential categories prototypically viewed as reflecting REALIS and IRREALIS.

Kristine Stenzel is a PhD student in the Department of Linguistics at the University of Colorado and can be reached at Stenzel@Colorado.EDU.

Colorado Research in Linguistics - Volume 17, Issue 1 - June 2004