Social Labels in DARE

Social Labels in DARE George H. Goebel DARE distinguishes two main types of labels — bold-faced regional labels1 and italic social labels — which can be applied to entries, senses, or individual grammatical forms or pronunciations. What we broadly call social labels really comprise three distinct types, which we call "user labels," "currency labels," and "style labels." User labels identify social groups or types among whom the item in question is used more frequendy than in the population as a whole. Currency labels identify items that have gone out of use or are in the process of going out of use. Style labels are concerned with the tone or usage level of words. Because DARE presents the reader with the evidence, or a selection of the evidence, on which each entry is based, the function and scope of labeling is somewhat different than it is in a citationless dictionary . Our basic policy is to label items only when the evidence is sufficient both in quantity and quality to allow us to generalize with confidence . What constitutes sufficient evidence, however, is rarely a simple objective question. When we have statistical evidence from the DARE Questionnaire or other linguistic surveys, we are on more solid ground, though even then the evidence is seldom entirely unambiguous . When we take into account other evidence, varying widely in nature and value, the problem becomes very difficult indeed. Inevitably, we sometimes wonder in retrospect what made us so confident here, so cautious there. Also inevitably,judgments have changed in the 20 years since editorial work began; we have become more cautious about mak1On the subject of regional labels in DARE, see the article by Luanne von Schneidemesser in this issue Social Labels in DARE179 ing broad generalizations, more willing to reserve judgment and let the evidence speak for itself. But even then editorial judgment is involved , not only in the selection of the quotations to be printed but also in the decision as to what supplementary information about social context, not deducible from the quotations themselves, should be included . Because these matters cannot really be separated from one another , I will begin by surveying briefly the sources we have for social labeling and how they are presented in our entries. Types of sources and their treatment From the beginning of the DARE project the intention was to treat social variation on an equal footing with regional variation. Although the selection of informants for the DARE Questionnaire was frankly biased toward the people most likely to show markedly regional types of speech — especially older and more rural people — care was taken that the full range of major social variables would be represented and that the relevant data would be recorded for each informant . After the responses had been computerized, software was written that would not only display the geographical distribution of the informants giving a certain response or set of related responses but also the basic social data about them. The social variables and the categories set up within each one were the following: Community: 1 — urban (population in the 1,000,000s) 2 — large city (population in the 100,000s) 3 — small city (population in the 10,000s) 4 — village (population in the 1,000s) 5 — rural Age:Y — young (18-39) M — middle-aged (40-59) O — old (60+) Education: X — unknown L — less than fifth grade G — fifth grade or higher H — at least two years of high school C — at least two years ofcollege orvocational school 1 80George H. Goebel Race:W — White N — Negro2 0 — Oriental 1 — American Indian Sex:F — female M — male Clearly some of these categories are very broad, but given the relatively small number of informants, it would have been unrealistic to make finer distinctions.3 (In all, 1,002 Questionnaires were administered ; in many cases different sections were administered to different informants, so that while the total number of informants used was 2,777, the maximum total possible for any one question was 1,002.) As an example of the output of our mapping/statistical software , the following table shows the social data for the informants who gave forms of the verb harn in response to question N8...