Moving beyond the modeling of boundaries, using a combination of linguistic filtering and spatial clustering a methodology is proposed for detecting language-internal transitions and boundaries in a probabilistic way, thus offering a novel solution that may complement conventional, visual methods of boundary delineation. Trend surface and regression analysis are used to quantitatively compare spatial variation of syntactic variables and to account for correspondence with theoretical models, concluding the importance of local analysis besides global.
Furthermore, travel time measures (for years 2000, 19) better reflect syntactic spatial variation than Euclidean distance, and older travel times account for contemporary syntactic spatial variation better than recent ones. Correlation analysis at different spatial granularities between linguistic and geographic distance matrices provides further evidence that geographic distance is responsible for, and thus explains, the majority of the variance found in Swiss German syntax. The main contribution of this thesis lies in introducing new methodologies in dialectology and dialectometry based on methods that already proved their power in GIScience and spatial analysis. The SADS is characterised by multiple responses at each survey site, distinguishing it from most other dialect surveys. All studies included in the thesis use data from the Syntactic Atlas of German-speaking Switzerland (SADS), a survey-based dialect atlas database focusing on (morpho)-syntactic phenomena. The focus of this thesis is on establishing methodologies in two key research topics: the role of geographic distances in dialectal variation and the quantitative assessment of language-internal boundaries and transitions. The research in this thesis aims to contribute to an improved understanding of the role of geographic space in dialectology by harnessing the potential of methods of GIScience. The inclusion of spatial, quantitative methods has therefore often been advocated in linguistics, yet spatially explicit methodologies have rarely been exploited so far. Although dialectology has thoroughly studied the formation of dialect areas, quantitative modelling of the spatial transition between different linguistic areas stays on the agenda. Digital linguistic data are being produced at an ever-increasing rate, paralleled by advances in computational methods development, including in Geographic Information Science (GIScience), leading to a great potential for collaboration between the computational sciences and linguistics. The variation inherent to linguistic and dialectal data is an important field of research due to the identities created through the perceived linguistic differences and similarities. Also check out my former colleagues' work at
0 kommentar(er)
