Fig 1.
Example calculation of LMI components, demonstrating the three main steps of the process: Linguistic distance calculation, exposure weights and relational weights.
The linguistic distance values are multiplied by the exposure weights and by the relational weights, and then the results are summed.
Table 1.
Example for the calculation of a linguistic distance between two localities based on data in the Sprachatlas der deutschen Schweiz (SDS).
A1, B1, B2 and C1 are variant categories, while A1.1, B1.1, B2.1, C1.1 and C1.2 are subvariants within the respective variant categories.
Table 2.
Composition of the four LMI prototypes.
Fig 2.
Distribution of LMI and dialect change with regards to age.
In Panels (A)-(D) the distribution of standardised LMIA, LMIB, LMIC and LMID values are shown (y-axes) against the age of the speakers (x-axes). Panel (E) shows the distribution of standardised dialect change rates against the age of the speakers. Each point represents a speaker (n = 500). Point colour represents educational level and shape represents sex. The blue concentrical lines show the density of speakers.
Table 3.
The ten lexical items used in the evaluation study.
Table 4.
The variables entered in the mixed-effects models of the evaluation.
Fig 3.
Dialect change rate for the ten lexical items by age cohort, sex and educational background.
Fig 4.
The relation between the four LMI prototypes and the dialect change rate.
LMI values and dialect change rates are standardised. The panels also show the numerical results of the linear regression models. Linear (red) and second-order polynomial regression lines (green) show the major trends. The slope of the lines shows the positive correlation.
Table 5.
Output of the mixed-effects models involving the survey criteria of SDATS as independent variables.
Table 6.
Addition of an interaction term between age cohort and the minimal LMI prototype.
Table 7.
Spatial characteristics of the dialect change rate and the LMI prototypes, using the Kruskal-Wallis test and measuring the spatial autocorrelation using Moran’s I.
Fig 5.
Spatial patterns of dialect change rates and LMID.
Panels (A) and (C): The average rates of dialect change are shown (on a scale of 0 to 1) in the polygons representing SDATS localities. Panels (B) and (D): The standardised LMID values of each SDATS locality are shown with a different colour scale. The darker blue and green colours (respectively) mean a lower value, while the darker red and purple colours mean higher values. In each map, those SDATS survey localities qualifying for the city rank in Switzerland (10000< inhabitants) are highlighted with magenta edges.