Introduction

Methods of computer mapping for linguistic data are heavily dependent upon the linguistic theory that the researcher brings to the task. The most familiar maps in language variation studies derive from a Neogrammarian approach; they assume the existence of areas where a ‘dialect’ might exist. For instance, Hans Kurath’s famous map of American dialect regions (Figure 3.1) shows clearly bounded regional dialect areas.

Such maps make a neat generalization from the often unruly distribution of the many variant forms for any given linguistic feature, such as the various designations for the dragonfly (Figure 3.2), perhaps the most famous dialect feature in American English. The different terms in use at the time of the survey to name this insect (only a few of the most frequent terms are shown) sometimes appear to be segregated by area, and in other cases appear to be mixed together.

For this reason, many language variationists now prefer other theoretical options, for example to chart just linguistic features instead of dialects, or to use statistical means to measure and represent how linguistic features occur in spatial patterns. Thus, this chapter cannot simply describe new ways of drawing isoglosses with computers, though it is indeed possible to do that by various means (cf., e.g., Light and Kretzschmar 1996). Instead, the purpose of this chapter is to suggest some basic ideas about the presentation of linguistic data on maps, and to illustrate a number of approaches to quantitative analysis and mapping of linguistic data. Discussion of some theoretical matters is necessary in order to show why different scholars might prefer different kinds of maps. Data on variation in English from the Linguistic Atlas Project will be featured here, especially data from the Linguistic Atlas of the Middle and South Atlantic States (LAMSAS, see Kretzschmar et al. 1993).