Abstract
In preparation of his edition of the 14th century alliterative poem Piers Plowman, the 19th century philologist,Walter Skeat, was able to find forty-five manuscripts. These were used in two different ways. First, he studied these with respect to their dialects, which led to his identification of three versions of the poem, denoted as texts A, B, and C. Second, he published about ten lines from each extract to display the variability of spelling of Middle English. Today grouping texts can be thought of as a clustering problem, which requires measuring distances between pairs of strings. This is a well studied problem from information retrieval, and many methods such as the Levenshtein edit distance exist. This paper visualizes the distance matrix obtained from applying edit distance to the forty-five extracts that Skeat published by using multidimensional scaling, which reveals the geometric structure of the variability among his manuscripts.
