The Mapping of Ulster-Scots

Philip Robinson

When Robert J. Gregg first published the results of his ground-breaking research project into linguistic boundaries in Ulster, its significance was recognised immediately by historical and cultural geographers(1). The 'Gregg Map' of the 'Ulster-Scots Speaking Areas' (as he later insisted on calling them) identified the precise spatial distribution of Ulster-Scots using linguistic criteria for the first time. His task was to distinguish this from Ulster-English (i.e. Northern Hiberno-English). The mapping of this cultural trait was regarded as only one marker or indicator of the areas primarily settled by Scottish plantation settlers in the 17th century. These Ulster-Scots areas stand in contrast to areas settled more by the English, or indeed those uncolonised 'native Irish' areas.

The extent of Irish-speaking areas in Ulster (particularly west Donegal, the Glens of northeast Antrim and the Sperrin Mountains in Tyrone) was already known from 19th-century and 20th-century census information, and Gregg also plotted these Gaeltachts to illustrate their respective territories. Of course, he realised the linguistic significance of the interfaces in counties Antrim and Donegal of these two traditional but declining languages. However, his primary mission was to distinguish Ulster-Scots from its closely related and dominant sister tongue, Ulster-English. In this, his linguistic expertise enabled him to develop criteria for distinguishing the survival of Ulster-Scots using local informants. He was aware that this task was complicated by the fact that Ulster-English contains a host of lexical and phonological items of Lowland Scottish origin.

Figure 1: Scotch-Irish (Ulster-Scots) Dialect Boundaries in Ulster

In the context of the contemporary Ulster-Scots renaissance (of which Gregg was himself a champion), his mapping of Ulster-Scots speaking areas, and his definition of the 'marker' criteria for identifying Ulster-Scots speakers, together represent a substantial legacy. His survey of 1960-63 provides a base-line against which future numerical decline or geographic contraction can be measured.

Gregg's published map (Figure 1) included the location of individual informants, all identified in his published thesis: 64 in county Down, 23 in county Antrim, 4 in county Londonderry, and 34 in county Donegal(2). Such an irregular distribution raises a number of questions about Gregg's cartographic methodology. Although unconventional, as we shall see, the concentration of Gregg's informants in particular areas was justified in terms of what was already known to him and what remained to be established by focused 'boundary tracking'.

Linguistic Criteria as an Index of Settlement and Cultural History

In Ulster there are three basic linguistic groups that contribute to traditional speech patterns today: Irish Gaelic, Ulster-Scots, and Northern Hiberno-English (Ulster-English). Ulster-Scots and Ulster-English are not only closely related to each other linguistically, but also are both considered to have originated from the 17th-century dialects of south-west Scotland and the north-west midlands of England respectively(3). The historical and settlement event which occasioned this phenomenon was the Ulster Plantation of the 17th century.

The assumption that a single index such as language or religious denomination can be taken to have ethnic, cultural and political implications may be grossly misleading. While religious denomination has always been a crucial factor in determining an individual's political and cultural allegiances, its role as the major factor in the process of determining identity has lost ground recently to language and other aspects of cultural tradition.

The identification of the Irish language with Catholic and Irish Nationalist identity can be traced back over a hundred years to the beginnings of the Gaelic Revival(4). The surviving Gaeltacht regions (as evidenced from the Irish language census data from 1911 and before) and the areas where the local population are Roman Catholic (also evidenced from the detailed census data of 1911 and before) are the usual indices to be mapped when the present Ulster landscape is examined to discover the extent (by obverse relationship) of plantation and other British settlement survival.

The Comisiún na Gaeltachta produced a mapped survey of Irish language survival in 1927 based on their 1925 survey. Areas where more than 10% of the population were found to be native-Irish speaking in this survey provide an identification of the 'traditional' Gaeltachts. Unfortunately, within Northern Ireland, these native-speaking communities in Antrim and Tyrone did not survive into and beyond the 1960s. A striking feature of the mapped extent of these Gaeltachts is its coincidence with areas where more than 50% of local surnames are of Irish origin and more than 50% of the local population are (or were in 1911) Roman Catholic (Figure 2)(5).

Figure 2: Irish/British Ethnic and Cultural Survival Areas in 20th-Century Ulster

Figure 2 illustrates the striking coincidence of these three indices of Irish ethnic and cultural survival in Ulster into the beginning of the 20th century — Irish language, Irish surnames and Catholicism. The areas at the core in this map (where all three indices are present) are in fact the areas of Irish language survival. Indeed, in these core areas the proportion of the local population that is Roman Catholic is invariably in excess of 75%.

Conversely, the areas shown in Figure 2 which contain none of the indices of 20th-century Irish cultural and ethnic survival represent the end-product of centuries of English and Scottish plantation, immigration, internal migration and acculturation. These areas are 'Protestant/Unionist' (in today's terms) and contain a population that is essentially of English or Scottish ethnic origin.

If the criteria that Gregg used to distinguish the Ulster-Scots speaking areas from Ulster-English dialect areas are accepted, his map can be used to differentiate between the 'English' and 'Scottish' ethnic and cultural areas which have persisted since the 17th century. In terms of religious denomination, the Ulster-Scots communities have long been associated with Presbyterianism, in contrast to the Anglican Protestantism that dominates the areas of south and mid Ulster that were settled by the English. The two main Protestant denominations in Ulster (Presbyterianism and Church of Ireland) have been broadly associated with Scottish and English settlement respectively. However, while Presbyterianism may have been a dominant philosophy among Scots settlers, it only became established as a separate and distinct denomination in the second half of the 17th century. It then spread rapidly throughout established areas of Scottish settlement. This spread was mapped by Alan Gailey in 1975(6).

While language and religious adherence have both been subject to processes of assimilation, surnames provide a more stable and representative index of national origin. The problems associated with the use of surnames in this context are not those of transmission between English and Scottish settlers, but rather of identifying which surnames are of English, and which of Scottish origin. A consensus of the surname, language and religious criteria which are markers of Scottish settlement, especially when mapped (Figure 3)(7), provides just as clear an indication of the centrality of language survival to the core areas of Scottish settlement as Irish does to the core areas of Irish settlement.

Figure 3: Scottish/English Settlement and Cultural Areas in 20th-Century Ulster

The core areas of Ulster-Scots settlement as shown in Figure 3 reveal a remarkable consistency, with more than 75% of the Ulster-Scots speaking population identified as Presbyterian within the Ulster-Scots speaking areas. Of course, the Ulster-Scots areas contrast with the areas of Ulster-English settlement in terms of their proximity to the coasts and the primary communication routes with Scotland and England respectively. This was a function of the historical processes of colonisation beginning in the early 1600s, when English and Lowland Scots migrated to Ulster in a territorially competitive fashion(8).

One legacy of this Ulster-Scots/Ulster-English interface that is well illustrated by Figure 3 is the 1798 Rebellion. In the same decade of the 1790s, the Ulster-Scots areas 'turned out' in the United Irishmen Rebellion of 1798, while the Orange Order was founded in 1795 at Loughgall in north Armagh, in the heartland of Ulster-English (and Episcopalian) settlement. Some of the government militias involved in suppressing the revolt were 'Orange' militias, and the principal engagements at Ballynahinch and Saintfield in county Down and at Antrim Town and Donegore Hill in county Antrim were all along the eastern frontier between these two cultural zones(9).

In this historical, geographical and cultural context, the significance of Gregg's map of Ulster-Scots speaking areas can be equated to that of the 1925 Comisiún na Gaeltachta in respect of the Irish language. The reality of Gregg's mapped area as the core area of Ulster-Scots settlement with its own distinctive cultural traits has, since Gregg's survey, been confirmed by ethnologists and cultural geographers.

In using cultural phenomena to ascertain the national origin of the population in a particular area, a principal obstacle is the fact that many cultural traits (which may be of English or Scottish origin) were disseminated rapidly throughout populations of mixed origin. Despite this, the pattern of a range of distinctively Scots folk-cultural traits (in addition to language, religion and surnames) has been shown to be consistent with the Ulster-Scots cultural zone shown in Figure 3. For example, the distribution of a (presumed) Scots New Year custom of distributing straw wisps was mapped by Alan Gailey in 1972 and shown to be confined to the Ulster-Scots zone in county Down(10). A host of other Scots calendar customs has been similarly identified in Antrim and Down, and the distribution of 18th- and 19th-century Ulster-Scots poets is also co-extensive with the Ulster-Scots speaking area mapped by Gregg in the mid-20th century (Figure 4)(11).

Figure 4: Ulster-Scots Areas in Ulster, showing the distribution of Ulster-Scots poets

Among these 'folk-life' criteria, those relating to vernacular building traditions are least prone to territorial spread by acculturation, presumably because the vernacular buildings themselves are immobile structures, fixed in the local landscape. In Alan Gailey's seminal work. Rural Houses of the North of Ireland, the results of his 20-year survey into the relative distributions of the two principal vernacular house types were mapped in detail for the first time(12). This map, particularly with respect to the boundaries in counties Antrim and Down, almost exactly traces the Ulster-Scots/Ulster-English linguistic boundary as mapped by Gregg. Historically, the house-type features mapped by Gailey can be shown to have originated in the different 17th-century vernacular house types that characterised Scots (and Irish) settlements on the one hand and English settlements on the other. Other specific folk-cultural criteria have been mapped in some detail in south Antrim, mid Ulster, and the Ards Peninsula in east Down(13). Here again the Gregg map is validated as confirming the historicity, the holistic nature and the persistence of the core areas of Ulster-Scots.

Historical Development of Scots-Settled Areas in Ulster

Although the focus of attention for any study of the origins of Scottish settlement in Ireland is inevitably the plantation of English and lowland Scots in the 1600s, it must be remembered that the movement of peoples between north-east Ireland and south-west Scotland has been a constant factor throughout the history and pre-history of these islands.

The Ulster Plantation was a governmental scheme for the colonisation of the six Ulster counties of Armagh, Donegal, Cavan, Fermanagh, Tyrone and Londonderry during the reign of James I (1603-1625). However, in practice, the colonisation extended throughout the 17th century, and over a much greater area, significantly in the 'unofficial' plantation of Antrim and Down(14).

By 1610, a plantation scheme had been approved by James I which involved granting one new county (Londonderry — formerly county Coleraine with some additions) to the twelve principal London Guild Companies. In county Londonderry each company got its own estate (such as the Drapers, Grocers, Goldsmiths, etc.), and, acting together, they were to be responsible for developing the two new city ports of Londonderry and Coleraine. Because of the proximity of these ports to Scotland, they actually became dominated by Scots merchants and artisans by 1630, despite their London-English ownership and control. Similarly, those London Companies with estates in the north of the county between Londonderry and Coleraine employed Scottish land agents to manage and 'plant' the lands, so here too the actual population was soon more Scots than English in origin.

The other five counties of the official scheme were divided into 'baronies', each barony being reserved for the use of one particular type of major land grantee. The different classes of grantees were to be groups of English Undertakers, Scottish Undertakers, English Servitors and Irish Natives. However, it was to be the Undertakers, in separate groups of English and Scots, who were to be responsible for the main implementation of the scheme for colonisation. In return for their estates, the grantees 'undertook' to plant 24 adult men, representing at least ten families of English or 'inland' (lowland) Scots on every 1000 acres.

The Undertakers were grouped in exclusively English or Scottish consorts for each barony under a principal Undertaker. Some of these consorts were of family groups, or of groups from adjacent areas in England or Scotland. The Scottish Undertakers were drawn from quite a restricted area within Scotland. The vast majority came from the central lowland belt. However, under the plantation allocations of land circa 1610, the lands allocated to Scottish Undertakers were (with the exception of north-east Donegal) well away from the ultimately Ulster-Scots zones in east and north Ulster.

A succession of governmental surveys in 1611, 1613, 1617, 1619 and 1622 were carried out to establish if these Undertakers had fulfilled their obligations to 'plant' settlers.

Figure 5: British plantation settlement in Ulster based on the plantation surveys of 1622.

Figure 5 shows the situation in 1622(15), with obvious concentrations in fertile regions of north Armagh, east and south Tyrone, mid Cavan and Fermanagh, the Foyle basin south-west of Londonderry, and across the north coast to Coleraine.

Figure 6: British settlement in Ulster, c.1659

These early surveys did not distinguish between English and Scots tenants, nor did they provide lists of surnames that might be used as an indication. Figure 6 provides a much broader picture of the settlement situation in 1659(16). This 'census' was based on poll tax returns and specifically distinguished 'Irish' inhabitants from 'English and Scotch'. Here we can see how the settlement on the official plantation estates sat in the context of the larger unofficial settlements in Antrim and Down. The distributional pattern of plantation settlement in the first half of the 17th century coincides substantially with the 20th-century situation (compare, for example, the detail of the British settlement distribution that is implied in Figure 2).

Figure 7: The Distribution of Protestants (Established Church and Dissenter) in Ulster, 1766

Most historians familiar with the settlement history of this period agree that the late 17th century witnessed a massive increase in British settlement in Ulster, and the levels reached in the 18th century (Figure 7)(17) represent the culmination of what had been more than a century of successive waves of incoming settlers. Although we can be confident of the detail and historicity of the predominantly 'British' settled regions of the rural Ulster landscape, none of the historic surveys of the 17th and 18th centuries enumerates English and Scottish settlers separately.

At the top social level of Undertaker and landlord, the precise original locations of the English and Scottish 'planter' families are known, but it is a fundamental error to assume that their tenants came from the same locations(18). Occasionally, contemporary records are specific. For example, in 1638 the Scottish Covenanters believed that there were 40,000 Scottish men in Ulster(19), while Sir Thomas Wentworth estimated in 1639 that there were 100,000 of the Scottish nation in Ireland(20). Scots in particular entered Ulster during the 1630s in increasing numbers. One report from Scotland in 1635 stated that in the preceding two years, 10,000 had passed through the port of Irvine on their way to Ireland(21).

Figure 8: The Spread of Presbyterianism in 17th-Century Ulster

The colonisation of Ulster was undertaken by the English and by lowland Scots settlers in a competitive manner which often enabled the two groups to retain their national identities in different districts(22). The spread of Presbyterianism in Ulster had by the late 17th century resulted in the formation of many Presbyterian congregations throughout the major areas of Scottish settlement (Figure 8)(23). The distribution of these earliest congregations conforms closely to the notion that there were four core areas where Scottish settlers were dominant (and even equate to the four distinct dialect areas of spoken Ulster-Scots today): north Down, east and mid Antrim, the 'Route' area of north Antrim and north-east Londonderry, and the 'Laggan' area of the Foyle basin in north-east Donegal and north-west Tyrone. These areas are precisely those of the first four presbyteries of Presbyterianism in Ireland: Antrim, Down, the Route, and the Laggan. From these four areas Presbyterianism continued to spread territorially during the late-17th century. By the time of the 1911 census, a clear territorial distinction was still evident between the two principal Protestant denominations, with a north-eastern crescent of Presbyterian ascendancy, contrasting with the Episcopalian areas of mid and south Ulster (Figure 9)(24).

Figure 9: Presbyterian/Episcopalian Areas in 20th-Century Ulster

Figure 10: Scottish/English Surname Areas in 20th-Century Ulster

Surnames not only enable the confirmation of the areas of early Scottish settlement, but they also provide positive identification of English-settled areas. The 20th-century 'surname landscape' of rural Ulster, as revealed in electoral lists of the 1960s, also provides (Figure 10) further confirmation of the continuity of English and Scottish settlement differentiation(25). There are two periods in the 17th century (the 1630s and the 1660s) for which it is possible to examine the surnames of the British inhabitants throughout most of Ulster. The 1630 Muster Rolls list the names of all the adult 'British' males capable of bearing arms on each large land-owner's estate, while the Hearth-Money Rolls of the 1660s provide the names of all hearth-owners in most of Ulster, townland by townland. Figure 11 shows which estates contained predominantly English or Scottish settlers in 1630(26) The areas of Scottish settlement indicated on this map coincide with the areas of Presbyterian dominance in the 17th century, although some areas in mid Ulster appear also to have had a majority of Scottish tenants. Areas of English settlement in county Londonderry, north Armagh, south-west Antrim and Fermanagh support the assumption that most non-Presbyterian (i.e. Episcopalian) British were of English stock. However, some of the plantation estates in mid Ulster were apparently 'mixed', with neither English nor Scots dominant.

Figure 11: English and Scottish Settlement in Ulster (based on numbers and surnames recorded in the muster rolls), c.1630

Before 1630, we know from the contemporary records of the Hamilton and Montgomery 'plantations' in north Down that these lands (and similar areas in Antrim) were overwhelmingly 'Scotch' as early as 1605(27). This included the towns of Bangor, Holywood, Newtownards and Donaghadee, but only in 1630 (Figure 11) can we see this in context for the first time. According to the 1637 Customs Report, the English in Londonderry were 'weak and few in number ... the Scots being many in numbers, and twenty to one for the 'English'(28). Older, pre-plantation towns with an established English ascendancy such as Carrickfergus, Downpatrick and Armagh had developed 'English', 'Irish' and 'Scotch' quarters or streets. In mid Ulster, the town of Dungannon had its inhabitants listed as 'English, Scotch and Wealch' in the certificate provided for the 1622 plantation commissioners(29). Indeed, for most estates in counties Armagh and Tyrone, the 1622 'Undertaker's certificates' survive, with the actual names of settlers listed. These confirm the presence of Scots on many of the English-owned estates, and the totality of Scots at the port and town of Strabane in north-west Tyrone(30).

By the 1660s, however, the Scots had consolidated in north and east Ulster and largely withdrawn from mid and south Ulster. The 1622 Survey of Cavan relates that many of the Scottish tenants of Sir Henry Piers 'had left that land (as we are informed) and were gone to dwell in the Clandeboyes from whence they came'. The 'Clandeboyes' was the area of north Down first settled with Scots by Sir James Hamilton (later Viscount Clandeboy) in 1605.

The 1611 Plantation Survey provides evidence that some Undertakers brought 'followers' with them to four Scottish-owned estates in county Tyrone. Bodley's Survey of 1613 records 'English' tenants on some English-owned estates near Dungannon, but states that another English-owned estate in Clogher Barony contained 'inland Scots'(31).

By the 1790s the consolidation of north and east Ulster into an almost continuous Ulster-Scots settlement zone, and mid and south Ulster into an Ulster-English zone, was both well-defined and well-known. The frontiers between these two territories provided the theatre for conflict between the two sides in the '98 rising, and so the boundary between Scots and English cultural areas sharpened with increased political and religious significance in the local psyche. In the Ordnance Survey Memoirs of the 1830s, the Surveyors noted acute local awareness of this frontier down to the detail of particular townlands in south Antrim(32).

In the core areas of Ulster-Scots settlement in Antrim and Down, those parishes which were peopled with Scots were identified very clearly in the Memoirs. These Memoirs often included written descriptions of every parish to accompany the first edition of the 6-inch mapped Survey of Ireland, and these descriptions included the 'habits, customs, and amusements' of local inhabitants. The Surveyors preparing the Memoirs were English or Anglo-Irish officials with much greater hostility towards Ulster-Scots than to Irish. For example, in the Parish of Carnmoney to the north of Belfast we find the following:

In east Antrim, the home territory of Professor Gregg, the Memoirs record for the following parishes:

In north Antrim, perhaps the stereotypical Ulster-Scots area, the Memoirs tell much the same story as for east Antrim in the following parishes:

Elsewhere in mid Antrim and Londonderry, the Surveyors of the 1830s again confirm the settlement continuity underlying the linguistic survival of Ulster-Scots.

These Ordnance Survey Memoirs provide a link with the plantation settlement patterns and establish beyond doubt that during the 1830s the same areas that are 'Ulster-Scots' today were even more strikingly 'Scotch' to the surveyors and cartographers of those days.

Gregg's Mapping of the Distribution and Boundaries of Ulster-Scots

Gregg began his fieldwork for a survey of Ulster-Scots in 1960 and completed it in 1963. His objective was 'the mapping of the distribution and boundaries of the Scotch-Irish [Ulster-Scots] dialects throughout the province', thereby 'tracing the external expansion of lowland Scots speech', which he observed in Ulster 'marches with Ulster Hiberno-English'(47). Gregg was aware that the term 'Scotch-Irish' which he used at the time of his first survey was understood in North America to cover 'things or persons of Ulster origin in general and, with reference to speech in particular, covering all types of Ulster dialect'(48).

The urban and rural dialects of Scotch-Irish examined by Gregg he defined specifically as the Lowland Scots dialects, which are 'still spoken in the areas that were most intensively peopled by lowland Scottish settlers during the 17th century'(49). To Gregg, the coincidence of Ulster-Scots settlement and language was a premise rather than a conclusion. He began with a good understanding of the settlement history of the province as a whole, but particularly in county Antrim where his local knowledge determined that fewer informants would be required (Figure 1).

With the benefit of the survey behind him, he could confirm that:

Gregg began his survey in the Laggan area of north Donegal, for which he had compiled a lexical questionnaire involving over 500 words. It soon became apparent, however, that many of the Scots words on his list had 'spread far beyond the original settlement area as described by the historians of the Laggan plantation, and that their rather haphazard distribution would not give a satisfactory linguistic demarcation line'(51).

An alternative questionnaire was then put into operation, 'based on some of the historical phonological differences that separate Scots from English dialects in general'(52). It was quickly established that these typically Scots phonological patterns had remained entrenched in the areas of intensive Scottish settlement. 'The opposition between the Lowland Scots forms and those of English provenance provided almost everywhere a sharply-defined border'(53).

Gregg's Definitive 'Markers' of Ulster-Scots Speech

For the purposes of his survey, Gregg developed a phonological questionnaire based on over thirty years of previous research on the Lowland Scots speech of his home district of Glenoe in east Antrim. As his dedication at the beginning of his published survey acknowledges, this was literally his 'mother tongue'. Dialectal differences within Ulster-Scots were to be found in some cases between, say, county Down, Donegal, north Antrim, and south-east Antrim (the four principal areas of Ulster-Scots settlement). However, these differences proved only to occur within the accepted Scots typologies and indicated (perhaps for the first time in any academic study) that there may be different dialects of Ulster-Scots. Gregg was aware that these differences could provide problems, for example, with attempts at spelling standardisation, but he was convinced of the ascendancy of Antrim Ulster-Scots, which he used as the yardstick in recording regional variation. Despite the range of Scots dialects found across the Ulster-Scots speaking areas, he stated, 'Antrim, however, is the heartland'(54). Outside the Ulster-Scots areas were the '... regions occupied predominantly by English planters', which he understood to '... still exhibit many characteristics of the dialects of the N and W Midlands, the original home of most of these settlers'(55). Ulster Hiberno-English was assumed to be the dialect spoken in the areas outside of his boundary, and one possible criticism is that he did not allow for the possibility of any 'outliers', i.e. small isolated areas of Scots survival outside of the main area.

Having established that the lexical questionnaire he began with in Donegal was unsuitable for distinguishing Ulster-Scots from Ulster Hiberno-English (because of the widespread adoption of many of the words with Scots etymologies into vernacular speech across Ulster), he devised a phonological questionnaire. This alternative was based on '... some of the historical differences that separate Scots from English dialects in general', and was distilled by survey testing until a Final Phonological Questionnaire was determined. This final questionnaire involved 665 items arranged in 14 typological lists and was used throughout the survey area.

'These lists proved to be very successful, for it was quickly established that the typical Scottish phonological patterns had remained entrenched in the areas of intensive 17th century Scots settlement … providing almost everywhere a sharply defined border'(56).

With 125 informants, and 665 items in the questionnaire used with each informant, a theoretical 80,000 responses were anticipated by Gregg for his data-base. Although the full questionnaire was not completed for each informant, 300 pages of tabulated phonetic responses are provided in the published survey. In general, he only recorded the key sound in each word form, usually the stressed vowel. Gregg used mostly IPA symbols for his tabulation, although he describes some additional symbols he felt were necessary to use(57).

It is important to stress the importance of this body of tabulated information over and above the maps the data were used to produce. At each location of an identified informant, we have a unique historical record of a wide range of Scots phonological and lexical items elicited. Not only does Gregg use this information to map the geographical extent of Ulster-Scots, but in these tables he also records regional differences, defines the distinguishing characteristics of the language — the 'markers' — and provides us with an historical base-line against which present and future erosion can be measured.

The items in the phonological questionnaire were grouped into 14 lists, each list marking a group of words with similar features (usually the identity of the stressed vowel reflex as characterised in the present-day dialect).

It is useful to look at Gregg's first list as an illustration, for he considered the survival of the Germanic 'ch' sound in words like fecht (fight) as the 'most important consonant by far' for the purpose of discriminating between Ulster-Scots and Ulster Hiberno-English. This feature is /x/, 'the reflex of the velar fricative in O[ld] E[nglish] and M[iddle] E[nglish]'. It was (and is) well preserved in all the Ulster-Scots areas, not only in dialect words of a Scottish type such as sheugh, pegh, spraghle, etc., but in a large number of Standard English and Ulster Hiberno-English words where the feature is absent. List 1 contains 47 items where this feature was anticipated: bought, bright, brought, cough, daughter, dough, draught, eight, enough, fight, fought, height, high, laugh, light, might, neigh, night, ought (pronoun), right, rough, sigh, sight, sough, straight, thought, tight, tough, trough, weight, wright, wrought, ... and a further 16 items of dialect words such as sheugh, etc. Of these 47 items, Gregg selected six for mapping: dauchter, echt, eneuch, fecht, nicht and teuch (NB Gregg used only phonetics to represent these forms of 'daughter', 'eight', 'enough', 'fight', 'night' and 'tough'. In doing so he was able to record significant regional variation in Scots pronunciation).

The following list identifies which forms from each of the 14 lists Gregg selected for individual maps in his published survey. The Ulster-Scots forms are approximated here, rather than illustrate the phonetic variations recorded by Gregg.

Table 1. Items mapped individually by Gregg in his published survey. (Items marked* are those finally selected for confirmation of boundary in Figure 1, and these maps are presented below — Figures 12-63).

Gregg's Statistical and Cartographic Analysis

It had become apparent to Gregg at an early stage of his survey that some of the 665 features he had tabulated were 'not suitable for boundary drawing'(58). This was often because a particular feature was only preserved in part of the survey area (e.g. mave for 'move' was largely restricted to Down, and absent in Antrim), or the feature had become virtually obsolescent throughout the survey area, or had an equivalent form in Ulster Hiberno-English (usually a lexical item).

In his published survey, Gregg took 88 of these features and mapped them individually, selecting them on the basis of their apparent role as universal markers of the boundary. These maps, or rather the data-base used to compile them, was then subjected to a rigorous statistical analysis to further eliminate items which deviated from a 'perfect' correlation with the final mapped boundary. These deviations are explained in terms of the factors mentioned above, but with considerable detail given which is, of itself, of academic interest with potential for much further research.

The final boundary map as published in 1972 and 1985 (Figure 1), was compiled using a statistically-refined residual total of 52 of the 88 mapped features. The 'best fit' features (with virtually no deviation for all informants from a perfect correlation with the final boundary), were the three maps and data for the features COW (Figure 12), ABOUT (Figure 13) and FIND (Figure 14). The full set of these 52 maps is presented here (Figures 12 - 63) in order of increasing deviation, so that Figure 63 (where een is found for 'EYES') has the most deviation among informants in the final selection of features. The 52 features selected statistically by Gregg in his final sift are identified in the above Table (Table 1) by asterisks.

Figs. 12-23 | Figs. 24-33 | Figs. 34-43 | Figs. 44-53 | Figs. 54-63