3.2.1. Analysis of Geographical Patterns

Parent Previous Next


3.2.1. Analysis of Geographical Patterns

This type of analysis considers the distribution of the map data and tries to find a geographical pattern in order to reach a conclusion on how a language feature is related to the geographic location. A geographical pattern may possess one of three properties:

    1. a complete spatial randomness (CSR), where every point corresponding to a language feature occurs within a study area in a completely random fashion, e.g. locations of trees in a forest;
    2. more uniformed than CSR, e.g. the white and black boxes on a chess board are completely uniformly distributed;
    3. more clustered than CSR, e.g. the high population density in urban centers.

Such analysis can be useful in finding the spatial correlations between language data. In section 4 maps are prepared to show how the initial consonant of who is pronounced across the United Kingdom. Four different pronunciations, namely, [hw-], [w-], [h-], and [∅-] are plotted against geographic location. The maps in figure 14 and figure 24, show that such a distribution possesses the property of being “more clustered than CSR”, meaning that a particular pronunciation depends highly on geographic locations.

Created with the Personal Edition of HelpNDoc: Full-featured EPub generator