Move back one step  Move forward one step  Display the start-up screen
Larger font Smaller font

Dialectology


Techniques
Attitudes and stigma

The impetus for dialect geography It may be surprising that the major step toward studying dialects systematically begins in the latter half of the nineteenth century, although there is a long history of observation of dialect differences prior to this time. In France, for example, the primary dialect division between the north and the south was characterised as early as 1284 by the poet Bernat d’Auriac. Here the forms of the key word ‘yes’ are essential and have even resulted in the names of two large parts of France, Languedoc and Languedoeil, the former referring to the region south of the Loire (the source of modern Provençal), the latter to that north of the Loire (which later developed into modern French).

In England, John Trevisa described a dialect continuum from north to south in 1387, and this has been supported by the systematic studies that began more than five centuries after he wrote. The first attempts to systematise were initiated by the striking advances in philology and language studies in general which were made at the end of the 18th and the beginning of the 19th century.

The Neogrammarians (historical linguists of the late 19th century) were the first to search for general principles of language change. One principle of their research was ‘Verner’s Law’ (named after the Dane Karl Verner who discovered it in 1875) which eliminated the largest set of apparent exceptions to ‘Grimm’s Law’ (called after Jakob Grimm who formulated it in his Deutsche Grammatik of 1816 and 1821) by showing that all sound changes are rule-governed. The relevance of this hypothesis is seen in the development of dialect geography. The first results of dialect geography seemed to disprove the theoretical stances of the Neogrammarians. As a consequence, from the first studies to the most recent, dialect geography has scarcely involved itself at all with linguistic theory. Only recently has there been a rapprochement between the different positions.

An outline history of dialect geography The pioneering work of dialect geography was carried out in Germany by Georg Wenker. In 1876, he began sending out questionnaires to schoolmasters in the north of Germany asking them to provide equivalents of standard words in their local dialect. It took him ten years to cover the entire nation. He sent his list of 40 sentences written in standard German to nearly 50,000 schoolmasters and received completed questionnaires from about 45,000 of them. The amount of data, enabled by the postal questionnaire method, forced Wenker to limit his analysis to the variants for certain words. He ended up making two sets of maps, which were then published under the title Sprachatlas des Deutschen Reiches in 1881 which covered north and central Germany. Wenker carried on gathering questionnaires so that in 1926 the first volume of the Deutscher Sprachatlas which was based largely on Wenker’s data was published under the editorship of Ferdinand Wrede.

Although it was possible to accumulate a large amount of data by sending out questionnaires, this method had its limitations mainly that dialect pronunciations could not be accurately recorded. Therefore, in 1896, Jules Gilliéron came up with an alternative. He sent out trained fieldworkers to conduct interviews and record the data in a consistent phonetic notation. One of Gilliéron’s fieldworkers, Edmond Edmont, who was famous for his good hearing, went around in France from 1896 to 1900 and recorded no less than 700 interviews. The results of his observations, together with those results from Gilliéron and his other assistants, were subsequently published between 1902 and 1910 under the title Atlas linguistique de la France. Other projects followed, as for example in 1930 by Hans Kurath who served as the director for the first region to be surveyed in the United States, the New England States. More surveys followed so that in 1949 Kurath’s Word Geography of the Eastern United States appeared.

The Survey of English Dialects (SED) was undertaken between 1959 and 1961 by Eugen Dieth and Harold Orton and published between 1962 and 1978. The Basic Material was published as a compendium of four volumes including each informant’s response to each question in the interview. The SED further published interpretative volumes like the Phonological Atlas of the Northern Regions by Edouard Kolb in 1964, A Word Geography of England by Orton and Nathalia Wright in 1974, and The Linguistic Atlas of England by Orton, Stewart Sanderson and John Widdowson in 1978.

The techniques of traditional dialectology


The methods of dialect geography: The questionnaire In so far as dialect geography seeks to provide an empirical basis for conclusions about linguistic varieties occurring in a certain area, its methodology is much the same as in other branches of linguistics. However, there are some aspects which are uniquely associated with dialect geography, like the questionnaire.

Within the guidelines established by a questionnaire, fieldworkers try to elicit a common core of data. The advantage of a questionnaire is thus to ensure that the results of all the interviews will be comparable. The way to elicit the data by a questionnaire can be different. Its use can be either direct or indirect. An example for the direct use may be the following question ‘What do you call a cup?’. The other way is to make use of indirect questions like ‘What is this?’ holding up a cup. The advantage of this kind of question is to encourage the informants to give more natural responses.

Another criterion for differentiating questionnaires might be formality. In one case the forms of the questions are free, and in the other case the form is supplied in advance. Within the use of indirect forms there are various possibilities to frame questions. The basic types are naming and completing. Naming questions involves quizzing the informant like ‘What do you say to a caller if you want him to enter?’ (come in). A subtype of naming is formed by asking questions which elicit more than one word like ‘What can you make from milk?’ (butter, cheese). The so-called reverse question is an attempt to elicit a particular word by getting the informant to talk about it at some length by questions like ‘What’s the barn for, and where is it?’. The second basic type, completing questions, can be represented by questions like ‘You sweeten tea with ...?’ (sugar) leaving a position blank. A subtype of completing questions is called converting questions which requires completing sentences like ‘A tailor is a man who ... suits.’ (makes).

The basic organisation of the questionnaires generally refers to semantic fields, i.e. semantically similar items clustered into groups. For the rural areas, the semantic fields may include such areas as farming techniques, the weather, etc. to get the appropriate data. Furthermore, it is necessary to take several conditions under which the interview is conducted into account to get the representative data one requires. Some elicitation techniques may lead to misleading results as the informants answer in a relatively formal or careful style. The success of the interviews is therefore often dependent on the technique used to elicit the information.

Linguistic maps Linguistic maps can either take the form of display maps or interpretative maps. Display maps, which are more common, transfer the results of each of the items indicative of dialect variation onto a map and thus reveal a geographical perspective. Each utterance of an informant is associated with a distinct symbol. Therefore, display maps give detailed information about the entire survey.

Interpretative maps, are often based on display maps or other comparable dialect geography projects as their primary source. The elicited data is used to make more general statements in terms of picking out responses for a particular item that predominates in various regions. Interpretative maps simplify display maps as they represent the relevant trends and their distribution, with the rare items omitted while the very frequent items indicate a trend.

The selection of informants The most typical feature the major projects in dialect geography have in common is the type of informant selected. Nearly all informants fit into the category of non-mobile, older, rural males. Only 60 of 700 informants in the French survey by Edmont were women and about 200 of the informants were educated; all of them came from rural areas. Kurath attempted to select a broader base of informants by establishing three different types of informants (3 categories and 2 sub-categories).

Type I: Little formal education, little reading, and restricted formal contacts.
Type II: Some formal education, usually high school; wider reading, and more social contacts.
Type III: Superior education, usually university; wide reading, and extensive social contacts.
Type A: Aged, or regarded as old-fashioned.
Type B: Middle-aged, or regarded as more modern.

All of the informants were non-mobile and the majority conformed to the criteria of the typical informant mentioned above.

The motivation for consistent choice of informants is clear: non-mobile informants ensure that characteristic features of speech appear in uncontaminated form. Older people reflect the variants of a bygone era. Rural people were preferred because urban people are too mobile and males as a rule use vernacular speech more than women do.

Dialectology and philology As mentioned above, dialect geography originated in response to a theoretical claim by the Neogrammarians and Wenker’s original work was motivated in part by the claim that sound change was regular. The significance of this claim is that if a sound change takes place, it will take place in all cases which had the sound in question, or at least in such cases in which the sound occurs in a particular environment - i.e. sound change is rule-governed and exceptionless. In his studies, Wenker found proof for this law, e.g. the change of word-initial /t/ (as in English tide) to /ts/ in German (as in Zeit).

Although the suggestion about the rules of sound change is substantially correct, the situation seems to be more complex and reveals that sound changes are not really exceptionless. Wenker, for instance, investigated a change of medieval German /u:/ to modern German /au/, but this diphthongisation did not take place in the entire area, i.e. some areas remained unaffected by the change. In other words, the German-speaking area was divided into parts which had the original /u:/ and those which had the newer /au/.

Attitude and stigma


Overt stigmatisation One obvious indication that a variable is a marker rather than an indicator (see above) is that it is the subject of unfavourable comment. Variables which are often mentioned in the Norwich community include (h), (t) and (ng), all of which are markers. Why are these variables subject to overt criticism while others are not? One of the main reasons seem to lie in the divergence between pronunciation and orthography. The low prestige variants of the three Norwich variables - ø, [ʔ] and [n] - can all be, and often are, characterised as ‘dropping your h´s, t´s and g´s’. These characterisations are commonly given by schoolteachers, but they are also given by other members of the community as well. The fact that laypeople have an expression for these features shows that they are highly consciousness of them.

Linguistic change Overt stigmatisation alone does not account for all linguistic variables which become markers. Another factor in a variable becoming a marker is that the variable is involved in an ongoing linguistic change, since speakers are more aware of the social significance of forms if the variable is subject to a linguistic change. By contrast, indicators appear to be relatively stable. Even variables which are not subject to any overt comment but which are definitely markers can thus be accounted for.

Phonological contrast Studies of urban dialects show, however, that there are still other markers left which can neither be related to overt comment nor linguistic change. One example is the Norwich variable (yu) involving the vowel sound in words like tune, view or news. Historically, words like rule and rude were pronounced [rju:l] and [rju:d]. In Modern English, however, [j] no longer occurs after [r] (in an area of eastern England which includes Norwich, the loss of [j] before [u:] has been extended still further to include environments following any consonant). Therefore, the variable (yu) has two variants: [ju:] as in RP view, and [u:] as in [vu:] ‘view’. One therefore requires an additional explanation to account for the status of (yu) as a marker. One possibility arises from the fact that (yu) has variants which are phonological rather than merely phonetic. Minimal pairs depend for their differentiation upon the presence of [j]. It seems reasonable that the involvement of (yu) in a phonological contrast may draw more attention to it than to variables which are simply phonetic.

Stereotypes Awareness by speakers of linguistic variables obviously admits degrees of ‘more or less’ and can change in the course of time. Thus, linguistic variables can change from the category of an indicator to the category of a marker and vice versa. Variables may, for instance, start as an indicator following a linguistic change in a particular social group. These changes can take place relatively unobserved. Changes of this type are called changes from below the level of conscious awareness. Subsequently, as usage of the new variable increases, awareness of class differentiation will also increase and the indicator may well become a marker.

At another stage, awareness of particular variants becomes even higher and speakers become especially conscious of them. The aristocratic English pronunciation off /ɔ:f/ rather than /ɒf/, for instance, can be referred to as a stereotype, since its social and regional connotations have become part of common knowledge. Changes of this kind stem from above the level of conscious awareness and can thus become part of a stereotype of a class.