Spatial reference in Sumu-Mayangna, Nicaraguan Spanish, and Barcelona Spanish

This dissertation examines linguistic spatial frame of reference (FoR) usage across three cohorts, detailing the lexical and structural realization of particular spatial FoR classes within each linguistic community, as well as which linguistic and nonlinguistic factors are predictors of spatial FoR class usage. This study was designed to determine the degree to which language and environment are predictive factors of spatial FoR performance. Sumu-Mayangna (Misumalpan; Nicaragua) and Nicaraguan Spanish were chosen because while their speakers share a local environment, the languages are genetically unrelated. On the other hand, Nicaraguan Spanish and Barcelona Spanish were chosen because, apart from speakers\u27 access to similar linguistic resources, speakers of these two varieties occupy rather different environments. Over thirty-two hours of discourse data from three referential communication tasks were analyzed using generalized logit mixed models. Descriptive and predictive statistical analyses indicate that a constellation of factors are significant in terms of the production of linguistic spatial FoR preference across cohorts and tasks. Cohort, language, and environment were found to be predictive factors, with language identified as the most common predictor of spatial FoR usage. In contrast, age and gender were found to be marginally predictive. A subsequent qualitative analysis elaborates upon the quantitative results, showing that between speakers of the two Spanish varieties, there is evidence of an asymmetry in form-to-concept mapping with regard to spatial relators. In the same vein, spatial relators in Sumu-Mayangna are discussed in the context of meronymy. The relative productivity and geometric organization of the Sumu-Mayangna meronymy system is discussed. Additionally, evidence found to be supportive of Bohnemeyer\u27s meronymy hypothesis is presented, which provides some explanation as to the non-preference of relative spatial FoR classes in the Sumu-Mayangna case. This study concludes that a combination of factors may ultimately predict spatial FoR usage, with language, and specifically, its attendant community-specific discourse practices, playing a unique role