How to Read a DARE Map
by Craig M. Carver
Figure 1. The DARE map of the United States with a conventional map for comparison.
The DARE map is designed to illustrate visually the regional distributions of words and phrases elicited by the questionnaire. Though it resembles the conventional areal map of the United States, it appears distorted because it displays population density rather than land area (see figure 1).1 This makes Nevada, for example, a tiny wedge of a state, while New York, in reality half the size of Nevada, is a large elongated shape one quarter of which is the bulb-like protrusion of New York City. Despite their abstracted shapes, the states have the same general spatial relationship to each other on the DARE map as they do in reality, with a few exceptions, such as Maine’s boundary with Massachusetts and the missing boundary between Wisconsin and Iowa. Also, Alaska and Hawaii are joined to the mainland.
Figure 2. DARE’s 1002 communities on the DARE map.
The overall area of the map is almost completely filled in by the 1002 DARE communities, which are usually symbolized as large dots (figure 2), each dot representing a single community where one questionnaire was completed. Each community is assigned a position on the map defined by a set of x-y coordinates, which the computer uses to locate and plot the individual informant responses.
Figure 3. DARE’s 1002 communities on a conventional map.
Figure 3 displays the communities on a conventional map and, when compared with figure 2, shows how the DARE map is distorted. On the conventional map, the communities are concentrated in the most populous areas of the country, so much so that to fit them all on the map we have had to enlarge the scale of the northeast. At the same time the communities of the western states are widely scattered. The DARE map, by contrast, compresses the western states while expanding the more populous eastern states, creating a relatively uniform distribution of communities across the map. This makes it easier to see the clustering of the communities where a given response is recorded.
Figure 4. Informant numbers for the DARE communities.
Figure 4 shows, in their relative positions, the informant numbers for all 1002 communities where questionnaires were answered. Although there is only one questionnaire per community, more than one informant was usually involved, each answering different sections. For example, Colorado informants 11–16 each answered parts of the questionnaire for Durango. The informants’ numbers were assigned arbitrarily as the questionnaires were returned and have no other significance.
In some rural areas more than one village or post office address constitutes a single community. For example, informants 34 and 35 in Maryland, who answered different sections of the same questionnaire, were from the towns of Still Pond and Chestertown in Kent County. At the same time, large cities constitute many communities. New York City, for example, has 22 DARE communities and thus 22 questionnaires (see figure 9).
Because most of the map’s area is composed of juxtaposed communities, the space that normally exists between towns is absent. Durango, for example, is sandwiched between Moab, Utah (UT11–15) to the west and Trinidad, Colorado (CO22–27) to the east (see figure 4), when in reality they are hundreds of miles apart. Not every space on the DARE map, however, indicates a community. There are empty areas that do not correspond to any actual geographical feature, such as the gap to the south of Durango. These gaps are created by the inherent difficulty of superimposing a horizontal-vertical arrangement of communities on a map shaped to resemble the United States. With the exceptions of the small areas in southwestern Ohio and to the east of Tennessee and the stylized portions of Maine and Massachusetts, the largest empty spaces on the map occur in the West, where this difficulty is increased by the higher ratio of land area to population. Although the empty areas are apparent when the map is filled, as in figure 2, they are virtually invisible on the maps in the body of the Dictionary (which are never completely filled), and they are not large enough to interfere with the regional distributions.
Though the DARE map roughly retains the relative locations of the communities, there is some inevitable deformation. Figure 5 gives an idea of the degree to which some communities have been displaced from their actual locations. The numbers on the DARE inset in figure 5 are not informant numbers but place markers for the 28 communities in Virginia and the 5 in the District of Columbia and its suburbs; they correspond to the numbers on the geographical map.
Figure 5. The relative positions of DARE communities in Virginia and Washington, D.C. The numbers correspond to the same communities on both the DARE map and the conventional map.
Although the DARE map is generally oriented according to the compass points, the specific spatial relationships among the communities are occasionally skewed. For example, in the DARE version of Virginia in figure 5, Charlottesville (9) is east of Buckingham (8), whereas in reality it is almost directly north of Buckingham. Perhaps the greatest distortion occurs in the relationship between Fredericksburg (11) and Capron (12), which on the DARE map are displaced east and west of each other separated by Washington, D.C., when in actuality Fredericksburg is more than a hundred miles north of Capron.
Such distortions at the local level, however, become less important at the regional level. From DARE’s satellite view, the details recede and a new picture unique in linguistic geography emerges—a view of the nationwide distribution of American speech. It is at this level that the DARE map is intended to be used in conjunction with descriptive labels to give a reliable impression of the overall regionality of a word. For purposes that require a more exact picture, such as the drawing of isoglosses (lines showing boundaries formed by the distribution of particular features), community locations can be transferred to a conventional map with the aid of figure 4 and the List of Informants.
Figure 6. The distribution of mosquito hawk and skeeter hawk on the DARE map and a conventional map.
Figure 6 illustrates the general reliability of the DARE map by comparing the ways a regional expression is plotted on it and on a conventional map. Both show the distribution of informants who gave the response mosquito hawk (=dragonfly) and those who gave the response skeeter hawk to question R2. The distribution of this expression is a paradigm for the southern dialect region. The DARE map gives a concise visual statement of the overall clustering of responses. For example, it is just as easy, if not easier, to see on the DARE map as on the conventional map that the variant skeeter hawk is especially concentrated in the South Atlantic states from North Carolina to northern Florida. The DARE map is essentially a scatter diagram that economically illustrates degrees of clustering—that is, degrees of regionality.
The clustering of mosquito hawk is a notably tight one, but even here the map reveals some stray responses deep in Yankee territory. As the user of this Dictionary will soon realize, language refuses to stay within strict geographical boundaries and almost always ignores political or state boundaries. Some of the apparent anomalies here, however, can be explained using the field records.
Occasionally, when an informant was slow in responding to a question, the fieldworker would mention a number of possible responses. The fieldworker was instructed, in such cases, to mark “sugg” (suggested) beside the response in the questionnaire. The two anomalous instances of skeeter hawk in Ohio are both marked “sugg” in the questionnaires. The unexpected occurrence of mosquito hawk in Minnesota, Wisconsin, and Michigan is explained by the biographies of the informants: four of these informants frequently vacationed or wintered in Florida or other parts of the South; it was there, very probably, that they picked up this memorable name for the dragonfly. Such supplemental information is provided in the DARE entries when appropriate.
Figure 7. The distribution of mud wasp, mud dauber, and dirt dauber on the DARE map and a conventional map. When mud wasp and mud dauber occur in the same community, an open triangle is used. If dirt dauber occurs in the same community as either of the two other responses, only dirt dauber is indicated. On the conventional map, the largest square symbol represents four or more communities where mud wasp was recorded; the smaller square (as in Vermont) represents two or three communities.
Figure 7 shows the relationship between the two maps for a regionally more complex set of responses. It also illustrates a fairly common situation in linguistic geography: the complementary distribution of synonymous expressions. In this example the synonyms are folk names for any of various wasps that build a nest of mud: mud dauber, dirt dauber, and mud wasp (qu. R20). Each of these expressions is used in a more or less distinct region which forms the counterpart of the other two.2 In this case, mud wasp is primarily used in the North and North Midland; mud dauberprimarily in the Midland and West; and dirt dauber in the South and South Midland.
DARE handles complementary regionalisms with a combination of cross-references and contrast maps. In the entry for mud dauber, for example, both dirt dauber and mud wasp are cross-referenced. In addition, the reference “See Map and Map Section” directs the user not only to the map accompanying the entry but also to the Map Section of the final volume. That section of the Dictionary contains maps that plot two or three regionally contrasting terms like those in figure 7. It also contains maps that illustrate the regionality of a response in terms of one of the five social variables: community type, age, education, race, and sex.
Figure 8. Social map of the distribution of beau dollar.
Figure 8 maps the response for beau dollar (=a silver dollar) in terms of the race of the informants. It shows that this expression is used primarily by Black speakers, since 27 of the 34 informants were Black compared to the much smaller percentage of Blacks in the entire sample of DARE informants. But more important, it reveals that the spread of this term northward is the result of the migration of southern Blacks into the urban North. All of the informants in the North who gave the response beau dollar are Black and live in large cities, namely Chicago, Cincinnati, Philadelphia, and Trenton.
Figure 9. Locations of urban, coastal, and mountain DARE communities.
Figure 9 shows the locations of three types of DARE communities: urban, coastal, and mountain, as well as selected geographical landmarks, relating the DARE map to the conventional map of the United States. For example, what appear on the DARE map to be coastal communities in Mississippi are not, and what appears to be inland territory is sometimes actually coastline, as in North Carolina. In addition, unified geographic features sometimes appear to be split, as is the case with the Missouri Ozarks, the South Carolina Appalachians, and New York’s Hudson Valley. Some cities—Boston, Charleston, New Orleans, St. Louis, and Los Angeles—are also divided on the map.
A DARE entry may present evidence of regionality in two ways: with a selection of quotations, which also illustrate history and usage; and with the findings of the fieldwork as represented in maps accompanying certain entries. The maps do this by plotting the locations of all the informants who gave the same or closely variant responses to the questionnaire as summarized in the DARE quotation.3 When their locations, represented by dots, cluster together or are clearly denser in one area than others, the response is considered regional. In this way the maps give a visual definition of the specific region of use.
DARE uses 37 basic labels to indicate regionality. In the Dictionary entries these regional designations appear in bold type. Because few regional expressions strictly confine themselves within sharp boundaries, DARE labels are often modified by such terms as chiefly, esp (especially), or a combination of formulas, such as widespread, but esp common in a particular region, orthroughout U.S., exc (except) in a particular region. These modifiers cannot be quantified precisely: they merely serve to indicate generally the degree to which a particular distribution of informants conforms to the regional designation. They can be defined only in relation to one another.
Chiefly means that most of the evidence, of which there is a significant amount, is confined to a particular area; consequently, on the maps the informant responses tend to cluster in a relatively dense pattern, and even the more scattered locations are generally near the area of high density. For example, mosquito hawk is labeled chiefly Sth.
Especially (abbreviated esp) indicates somewhat weaker regionality than chiefly, because there is less evidence confined to the given area. For the maps this means that the clustering is not as dense. The exceptions are often distant or widely scattered. Esp, however, is only occasionally used by itself to modify a label. More often it is used in formula with other modifiers to designate a subregion within a larger context, as chiefly Nth, esp NEast. It is also often used with widespread or scattered, as in widespread, but esp Sth, S Midl. Widespread means that an expression has considerable currency virtually everywhere in the United States. Scattered indicates wide dispersion but low density.
To designate subdivisions of states the following scheme is used:
North | ||||
West |
nw | cn | ne | East |
cw | c | ce | ||
sw | cs | se | ||
South |
Each state is thought of as having three divisions in each of the four compass directions, plus a central area. The state code is two capital letters, as used by the U.S. Postal Service; the subdivisions are in lower-case letters preceding the state code. For example, the subdivisions of Ohio can be referred to as nwOH, cnOH, neOH, and so on. These abbreviations are often used with the quotations, to name a speaker’s home area or to indicate the locale of the quotation. They are rarely used as regional generalizations for the whole entry. Because the states are not rectangular, the scheme is only approximate, but for purposes of DARE it takes the reader reasonably close. With a living language, and without knowing the responses of every speaker, it would be misleading to pretend to any higher degree of exactness.
The definitions of the regional labels themselves are primarily based on the organization and evidence of the Linguistic Atlas projects and on general regional patterns that commonly occur in the DARE maps. For example, Upper Midwest and North Central States are based on the organization of the Linguistic Atlas; North, North Midland, South Midland, Midland, and South (among others) are based both on DARE maps and Linguistic Atlas fieldwork; and South Atlantic, Appalachians, Mississippi Valley, and West Midland are based almost exclusively on the DARE maps. Settlement patterns and the geographical evidence from quotations are also used in defining areas.
For those regional labels which are inductively derived from fieldwork, the intent is to define broadly the geography of particular usages, not to discover the specific boundaries of dialect regions. In other words, the regional label in an entry is a description of the actual distribution of a usage (insofar as it is known) in terms of predefined sections of the country.
- To be more specific, the map is distorted to reflect the number of DARE informants in each state, this number being roughly proportional to the state populations as of the 1960 census.
- Although this situation occurs with some frequency, the reader should beware of assuming that for every expression used in only one part of the country there is a regional counterpart in other sections of the country. Often a nonregional term, one that is standard or widely known, is a counterpart.
- When the number of informants given is greater than the number of dots on the map, this is usually because some informants gave the same response to more than one question, and are thus counted more than once. On the map, however, the community of each informant is represented by only one dot.