Perception of geographically conditioned linguistic variation Charlotte Gooskens, Renée van Bezooijen & John Nerbonne 1. Introduction Over the course of several decades, geographically conditioned linguistic variation in the Netherlands and the Dutch speaking part of Belgium has been investigated from many different perspectives. In our view, the various methodological approaches can be divided into two fundamentally different types. In the first type, data about the linguistic characteristics of dialects are usually gathered by means of questionnaires, sometimes on the basis of recordings of conversational speech, collecting information about what people actually say. The linguistic products, recorded in written or auditory form, are subsequently subjected to (simple or more complex) data analysis techniques. The final aim is to draw a linguistically based map showing the distribution and boundaries of dialect features and dialect areas. Some of the more well- known examples of these ‘objective’ techniques include the construction and tracing of isoglosses and isogloss bundles (e.g. Weijnen 1941), feature frequency counts and correlations (Hoppenbrouwers and Hoppenbrouwers 2001) and the calculation of Levenshtein distances between matched segment strings (e.g. Heeringa 2004). These traditional- dialectological and computational techniques have been discussed elsewhere (see the contributions by Niebaum & Taeldeman and by Heeringa & Nerbonne) in this book and will not discussed further here. In the second type of research, the basis of the investigation is formed not by objectively assessed linguistic elements (written or spoken dialect words, sentences or texts) but by the dialect speakers’ subjective awareness of linguistic variation. The global perceptions which dialect speakers hold of the degree of similarity or difference between their own dialect and neighboring dialects is used as input to draw dialect maps. Here, the metalinguistic knowledge and intuitions of dialect speakers rather than their linguistic competence form the point of departure to classify dialects. In other words, the emphasis is on what people think rather than what they do. This type of research is referred to as perceptual dialectology by Preston. We note that the sort of knowledge reflected in laymen’s opinions about dialects may be based not only on perception per se, but also on second- (or third-)hand information, perhaps not amounting to more than popular prejudice. It is also worth mentioning that, as “objective” techniques have become more sophisticated, resulting in ever more methods by which language varieties might be distinguished, many researchers have felt the need to “validate” objective methods by mean of subjective, behavioral tests, involving perspectives of the second type (Heeringa 2004; Heeringa et al. 2006). The term ‘perceptual dialectology’ was first used in Preston (1981) but developed fully in 1999. According to Preston, perceptual dialectology “represents the dialectologist’s- sociolinguist’s-variationist’s interest in folk linguistics. What do non-specialists have to say about variation? Where do they believe it comes from? Where do they believe it exists? What do they believe is its function?” (1999: xxv). When comparing scientific and folk characterizations of a dialect area, the dialectologist may discover, for example, “that there may be perceived folk dialect areas where there are none scientifically and vice versa” (ibidem). According to Preston the earliest systematic technique for determining folk perceptions of dialectal variation was developed by the Dutch dialectologist Weijnen, when