Linguistics Department Spring 2012 Lecture Series

Friday, February 3, 2012

Lattimore 513

3:30pm

How Many Tasmanian Languages Were There?

Claire Bowern, Yale University

The languages of Tasmania have been something of an enigma. The data are difficult to work with; they were recorded from multiple speakers in multiple locations, with little accompanying analysis. They are non-phonemic and recorded by non-speakers of the languages (and in the absence of other phonemically recorded data they are difficult to interpret). We are unsure currently of how many languages were spoken in Tasmania at the time of European settlement, how many groups (or families) they fall into, and how those groups should be subgrouped.

Previous work on Tasmanian languages has varied in the number of languages assigned to the island. Wurm (1972) suggests that there were just two languages, one in the Northern part of the Island and another in the south. Dixon and Crowley (1981) tentatively state that there were perhaps five, but perhaps as many as twelve. Horton (1994) has 6 areas, but the source of his information is not known. No previous work has made full use of Plomley's (1976) compilation of extant resources on the languages of Tasmania.

In this paper I present the results of a phylogenetic analysis of the languages in order to discover 1) how many languages are represented in the data, 2) how many families they are likely to belong to, and 3) how these categorizations relate to the more than 40 tribal and band groupings, which are fairly well described in the literature. I show how techniques from population biology can be used to classify wordlists of unknown origin.

Friday, February 24, 2012

Lattimore 513

3:30pm

Partly lost, but not forgotten: Non-prototypical noun class systems in the Cameroon-Nigeria borderlands and their role in understanding change in Niger-Congo

Jeff Good, SUNY Buffalo

An important problem of comparative Niger-Congo morphology is understanding the historical processes that relate word structures in languages of the isolating "Kwa" type to those of the agglutinating "Bantu" type. A salient subproblem of this larger morphological puzzle is charting the connection between the noun class systems of languages of the Kwa type which, at one extreme, can lack such classes entirely, against those of the Bantu type which, at the other extreme, are famously elaborated. Languages of the Cameroon-Nigeria borderlands can offer important insight into this issue since their noun class systems occupy intermediate points between these extremes and vary among each other in ways that help establish the extent to which different possible factors, such as sound change, class prefix realization, and class assignment, may play a role in the loss of noun class systems. A significant conclusion will be that noun class systems in Niger-Congo appear to be quite robust, in the sense that loss of one part of the system need not be correlated with loss of the other parts. This suggests that the level of noun class attrition found in Kwa languages was not a historically "natural" event. More generally, this study shows that languages not adhering clearly to typological prototypes may be especially instructive for understanding the nature of abstract grammatical systems such as those involved in noun classification.

Friday, March 2, 2012

Lattimore 513

3:30pm

Evidentiality and Clause Modality in Desano

Wilson Silva, University of Rochester

In this study, I investigate the complex and typological interesting evidential system of Desano, a highly endangered Eastern Tukanoan (ET) language. Evidentials in Desano are obligatory suffixes on the verb and agree with the subject in person, number and gender. According to Kaye (1970), Desano has three evidential markers: visual, non-visual and hearsay. He treats evidentiality and modality as two distinctive systems. In this paper, I propose a different analysis. I follow Stenzel's (2004, 2008) analysis for Wanano (also an ET language) and investigate evidentiality as a category of clause modality. I show that the evidential system of Desano is highly complex, having six evidential markers (and not only three as proposed by Kaye). In my analysis, the 'hearsay' evidential category can be broken into two morphologically distinct 'diffuse' and 'quotative' reportative forms, which can be added to the visual, non-visual, indirect (inferred), and internal (assumed) specifications to yield six distinctly coded evidential category. I conclude that in Desano, evidentiality (expressed by statements coded by evidentials) and modality (expressed by statements referring to irrealis situation and interrogatives) correlate as part of only one system that is sensitive to scalar values of speaker commitment.