Ads Annual Lecture: I'm Off to Philadelphia in the Morning: A Scotsman Looks at Dialect in America

SAN FRANCISCO, 5 JANUARY 2002. In his autobiography, the nineteenth-century English art critic John Ruskin remarks "though I have had kind invitations enough to visit America, I could not, even for a couple of months, live in a country so miserable as to possess no castles." When I came to the United States in 1965, I must admit that I wondered how I would survive in a country so miserable as to have no dialects. You may understand this if you look at what I found to be one of the most frequently cited examples of a dialect difference in the United States, Kurath's (1949, fig. 66) map showing the distribution of the words bucket and pail. On this map Kurath shows a general pattern of responses giving pail in the more northern parts and bucket in the more southern parts, though there are frequent exceptions. I invite you to compare Kurath's map with the corresponding map for pail from The Linguistic Atlas of Scotland (Mather and Speitel 1975, 51). Instead of the meager distinction between pail and bucket, which Mather and Speitel do not deign to map, there is the rich array of hannie, stoup, flaggon, scuttle, pitcher, cog or coggie, daffick, can, and luggie. Even more names are given in the atlas list (Mather and Speitel 1975, 184-86), though they are not deemed to merit mapping: fiddick, roosher, cuman, dirler, leglen, bowie, bakie, bine, and keggie. Eighteen names in addition to the two mapped by Kurath. How could Kurath dare talk about DIALECT boundaries? Clearly what he had got were two standard terms for which the respondents stated a preference. I felt that if this is what dialect means in America, then it did not excite me. Another Scottish example can be seen in Mather and Speitel's two maps for the different terms for playing truant (1975, 67-68). On these two maps no fewer than 26 items are displayed. Compare this with Kurath's maps (1949, figs. 157 and 158) showing just eight terms. Other examples in the first volume of The Linguistic Atlas of Scotland include over 40 different names given for 'broken pieces of china' (item 31), over 60 for 'earwig' (item 82), and 120 for 'sheep's dung' (item 63). At this point it may be relevant to remind you that the total population of Scotland is about five million, less than 2% of that of the United States, so diversity is not necessarily linked to size. My insular view of the limited nature of variety in the United States is not new. In 1781, John Witherspoon remarked that "There is a greater difference in dialect between one county and another in Britain, than there is between one state and another in America" (cited by Bailey 1991, 130).

In any social science the use of terminology is critical, so it is important to get our terms right. STANDARD LANGUAGE and VERNACULAR are such loaded terms that it is important that we should agree on their definitions both for our own work and also for communication with the world outside of linguistics. The furor over the Ebonics controversy is a cautionary tale of the difficulties that linguists can run into in trying to make clear their distinctions to a lay public that is confused to begin with. I believe that one of the sources for confusion in the debate, even to the extent that it was conducted on rational grounds, lay in the uncertainty that some linguists displayed in connection with the notion of DIALECT. It is the definition of dialect that I would like to examine now.

There are so many definitions of the term dialect that it would be easy to take extreme examples and show how problematic...