The Ulster-Scots Dialect Boundaries in the Province of Ulster*
Robert J. Gregg
Historical and Geographical Factors affecting Dialect Distribution: Historical background
Much has been written and many are the controversies involved in the history of the Ulster plantation. For our present purposes, however, an outline including only the salient features of this modern Völkerwanderung is sufficient.
Up to the end of the reign of Elizabeth I the province of Ulster was overwhelmingly Gaelic-speaking, whatever Anglo-Norman settlers remained having become 'more Irish than the Irish themselves'. Even the Scottish Highlanders under the Macdonnells of the Isles, who had acquired and been confirmed in the possession of large parts of county Antrim, spoke a type of Gaelic which must have been very similar to the local Ulster Gaelic, because of the unbroken contacts and the population movements back and forth across the North Channel since the Roman withdrawal from Britain in the 5th century. Only the walled town of Carrickfergus and the town of Newry with its immediate surroundings were English-speaking.
The 300,000 acres of Antrim territory in the Route and the Glynnes (stretching from Larne to Coleraine) which King James granted to Randal Macdonnell by patent deed in 1603 was largely uninhabited at that time. Macdonnell, although a Roman Catholic himself, found it expedient and profitable to people his vast estate with Protestant Lowland Scots. The Scots penetrated beyond the bounds of the Macdonnell territory into lands in S Antrim that had been granted to the English undertakers and occupied the county approximately as far south as the upper reaches of the Six Mile Water.
County Down had a somewhat different settlement history. There, by an unfortunate twist of fate Con O'Neill lost a great portion of his huge Clannaboy estate to two enterprising Scots: Hugh Montgomery, Laird of Braidstane, and James Hamilton of Dunlop, both Ayrshire men, who brought in thousands of settlers through the ports of Donaghadee and Bangor respectively. O'Neill finally lost what was left of his land to Sir Moyses Hill, an English undertaker who peopled his territory with English planters.
Apart from these three grants, the Barony of Lecale (which had come under the control of an Englishman, Sir Nicholas Bagenal), and the southern end of the Ards Peninsula (which remained under the sway of the Savages, a Norman family that had 'gone Irish'), the rest of the county remained under native Irish control.
Unlike Antrim and Down, which were settled by private enterprise, Derry was included in the official Plantation plans. Its settlement was the prerogative of the London companies, which had little luck in the enterprise.(12) The Lowland Scots, because of their closer bases, were able to take over a good portion of the NE corner of the county and penetrate loosely the rest of Derry and Tyrone.
It was also as part of the official Plantation plans that Scots settlers were brought over to Donegal from 1610 onwards by the Ayrshire families of Cunningham and Stewart. They were settled in the northern parts of the low-lying east Donegal region known as the Laggan. The southern parts of the Laggan were English-settled and most of the remainder of the county was left in the hands of the native Irish.
Thus, the Scottish settlers came to occupy solidly a wide band of coastal territory extending from the Ards in NE Down northwards, taking in most of county Antrim, the NE corner of Derry and the northern portion of the Laggan district in E Donegal.
As for their place of origin, it is well established now through denizenship papers and other historical sources that most of the Scots settlers came from Ayrshire and the adjacent counties of Renfrew and Dumbarton and from Galloway in general.(13) The relatively small group of English settlers brought over by Chichester came mainly from Devon, but the predominant English group hailed from S Lancashire, Cheshire and the general hinterland of the N and W Midlands.
One common misconception about the plantation of Ulster is that it consisted of one rapidly executed transfer of population. In effect, there was a continuous but fluctuating movement of settlers across the North Channel and the Irish Sea throughout the 17th century and beyond, with an occasional reflux following civil disturbances such as the uprising of 1641. From about 1720, however, as a result of legislation oppressive to non-conformists, a new migratory movement began — this time mainly of the Ulster-Scots across the Atlantic to N. America.(14) It is noteworthy that, in spite of this mass exodus, the US districts of Ulster were never depopulated and in fact have maintained their own particular character down to the present day.
Apart from these historical population movements, certain geographical factors played an important role in the actual location of settlers. Mountainous areas in general remained in the hands of the native Irish. According to Canon Hume,(15) the best arable lands were allocated to English settlers and other cultivable lands to the Scots, who often had to clear forest and scrub-land before farming operations could begin. Boggy areas were either uninhabited or peopled sparsely by the native Irish for centuries. These bogs posed — like the mountains — natural barriers which were at the same time linguistic boundaries, and only in recent times with the progress of land drainage schemes have these areas been opened up to settlement. Thus the US-UHE boundary in the S Ards Peninsula coincides with the limits of the former boggy tract cutting off the Portaferry end of the Ards. The Magilligan area in NE Derry likewise was opened up just a century ago by a drainage scheme and settled by Lagganeers from E Donegal, a fact still apparent in the current dialectal subdivisions. The SW corner of Antrim, N of Lough Neagh, suffered also in earlier times from flooding and bad drainage. In the last century with the help of a reclamation scheme this area, which must formerly have set limits to the westward expansion of the Scots, was once more opened to settlement. This time, people moved in from different directions with the result that we have nowadays a somewhat mixed dialect zone.
A Sketch of the Present Distribution of Dialects in Ulster
The present distribution of dialects in Ulster reflects to a large extent the settlement patterns laid down in the 17th century.
Regions occupied predominantly by English planters at that time still exhibit many characteristics of the dialects of the N and W Midlands, the original home of most of these settlers. The modern dialects spoken by their descendants may be referred to in general terms as Ulster Hiberno-English (UHE) in contact with Hiberno-English proper, which is generally used as a designation for the Dublin or southern Irish type of English. UHE of one type or another may, then, be assumed to be the dialect in all areas outside the US- and Gaelic-speaking districts, although it would be easy (and wiser) to recognise a sub-variety which we might call Gaelic-English in which the Gaelic polarity is predominant. Gaelic-English arises in all bilingual areas where the acquired tongue, English, is heavily coloured phonologically (θ>t, ð>d, r>r etc.) as well as syntactically ('It does be stormy'),(16) lexically, and semantically by the native Gaelic speech. This creolized type of English may be self-perpetuating in a few areas after the disappearance of Gaelic, although its ultimate fate would seem to be a levelling under the pressure of UHE. The English spoken in Glenvar, in the Fanad Peninsula further north, in the mountainous zone around Churchill (and in fact all points west of the Laggan) and in N. Inishowen could, in most cases, be classified as Gaelic English. A similar type of Gaelic English crops up in the Sperrins (county Derry) at Magheraneany.(17) The Devon-type of English which must have been spoken earlier by Chichester's planters around Carrickfergus seems to have been submerged by the more generalized UHE, but in the barony of Lecale a very distinctive Hiberno-English dialect has survived (probably from the time of Bagenal's settlement) which is in marked contrast with the neighbouring UHE. Speakers of the latter from the Portaferry end of the Ards as well as from the Killyleagh area are very conscious of the distinctiveness of this Strangford-Downpatrick speech, especially in its handling of the diphthongal reflex of ME i ̄. The UHE speakers along the fringes of the US area all have two reflexes, a broad [ɑe] and a narrow [əi], distributed partly on an allophonic, partly on a contrastive basis, as in the US dialects themselves, but with a different incidence.(18) The Lecale dialect, on the other hand, has only one reflex, viz., [ɑe], which stands out in sharp dialectal opposition to Portaferry or Killyleagh [əi] in such words as eye, die, why, etc.
The Lowland Scots dialects have likewise been preserved in the areas of intensive Scottish settlement. That typically Scots lexical items (as distinct from the full-blown historical-phonological system) are found everywhere in Ulster reflects the fact that many small groups of Lowlanders pushed far beyond the limits of the homogeneously Scots-settled areas and in time assimilated to the surrounding UHE speech, but not before bequeathing many expressive items to the vocabulary of their neighbours.(19)
The US dialects have a distribution that is largely coastal. They dominate the whole of the NE corner of county Down, i.e., the Ards Peninsula (excluding the southern end, around Portaferry) and the area west of Strangford Lough from a point north of Killyleagh, sweeping west almost to Hillsborough, and north, skirting the boundaries of Belfast itself.
Antrim, however, is the heartland of the US dialects. Apart from the English-settled parishes in the southern part, the formerly boggy corner between the Bann River and the north shores of Lough Neagh and the mountainous Glens in the NE, the whole county is dominated by various forms of US, among which it is possible to distinguish a northern sub-dialect with /e/ instead of /ɪ̈/ as reflex of Early Scots/Middle Scots ǖ. Within the latter it is possible to distinguish a distinctive coastal dialect in which /æ̈/ has largely replaced /ʌ/ in a wide range of words.(20)
NE Derry US is really an overspill of the northern sub-dialect of Antrim which straddles the Bann almost as far upstream as the Rasharkin-Kilrea crossing. The Magilligan area as mentioned above was settled as late as the last century by Lagganeers from E Donegal who brought with them their distinctive version of US with its typical forms such as [əˈbin] above, [din] done, [gid] good, etc.
The US of Donegal is centred in the Laggan district extending northwards to the foothills of the mountain zone of the Inishowen Peninsula and to the Knockalla range, which cuts across the Fanad Peninsula. The mountains of central Donegal form the western limits of US, the watershed between the rivers Finn and Deele the southern limits, and the Foyle itself the eastern bounds of the dialect.
As far as the linguistic nature of the dialectal borders is concerned, US marches for the most part with various forms of UHE, but in Fanad (along the southern boundary of Glenvar) and to the west of the Laggan the boundary is formed by Gaelic.
The US dialect boundary stretching E-W across the S Ards is with a variety of standard English which in many ways resembles US urban, in other words, the type of English used by educated bilingual speakers of US rural or used in the towns of the US areas. On the west side of Strangford Lough US borders on the typical county Down version of UHE, all the way west from Killyleagh to near Hillsborough and then northward touching upon the outskirts of Belfast, which has, of course, its own particular urban version of UHE,(21) nowadays reaching out eastwards all the way along Belfast Lough's southern shore to Bangor and Groomsport.
All these county Down US boundaries are quite sharply defined over against UHE, although the division at the eastern — Cloughey — end of the S Ards border is less clearly marked than at the Ardkeen end.(22) The typical Scots forms have strayed much further south on the east coast than on the west, which may be explained partly by the poor communications on the west at an earlier date, before the drainage scheme had opened up the Ardkeen area.
Belfast UHE also stretches along the northern shores of the Lough as far as Whitehead and forms a sharp boundary with E Antrim and, partly, with Mid Antrim US. Further west, a more rural type of UHE provides an equally clear-cut border running towards Lough Neagh and passing north of Antrim town. The rather mixed dialect of the area north of Lough Neagh and west to the Bann forms a vaguer boundary for the one hundred per cent US of W Mid Antrim, but the UHE of the Glens in the NE corner of the county stands out in sharp contrast to the N Antrim and Braid versions of US.
On the Derry side of the Bann the local UHE, S and W of the US area, is again sharply contrastive, and the same holds good for the situation in Inishowen and the S Laggan in Donegal. The border with Gaelic to the W and N of the Laggan is, of course, the sharpest possible contrast, as it is part of one of the major European linguistic boundaries - that between Germanic and Celtic.(23) Yet in spite of this apparently sudden transition, the US dialect of the Laggan has actually intruded across the line as the second language of most of the Gaelic speakers living near the border.
A Brief Analysis of Heartland US Rural Dialect: The US dialect of Glenoe, East Antrim
A lifetime's collection of materials for a detailed study of the East Antrim dialect of Glenoe was the basis for the research project on the US dialect boundaries in Ulster. With the help of data from this bulky corpus, two articles were written for Orbis, the first being a brief analysis of the synchronic phonology of Glenoe with a phonetic description followed by a phonemic analysis, and the second dealing with the diachronic phonology, starting, however, not with ME or OE, but working back from the current dialectal sounds to the older forms of which they are the reflexes.(24) The whole approach was thus dialect-centred.
For the purposes of this study a brief summary of the relevant features of the Glenoe phonological system is presented at this point:
Glenoe has a range of fourteen vocalic phonemes:
No 1 /i/ (close, front) is best regarded as basically short with a long allophone in final open syllables, in hiatus, and before voiced fricatives, including [r]:
No 2 /e/ (half-close, front) is realized as a long vowel in syllables with primary or secondary stress but has a short variant in weakly stressed positions:
|[ə ˈbɛ:tərˈde:]||'a better day'|
|[ə ˈsɛ:tərde]||'a Saturday'|
|[ˈsʌm ˌde:]||'some day'|
Some speakers use a diphthongal variant with a schwa-glide in pre-consonantal positions:
A diphthong with an even closer starting point, [iə], may crop up in such words.
No 3 /ɛ/ (half-open, front) is generally realized as a long vowel:
No 4 /ü/ has four positional variants. The main member (close, front-central, rounded) occurs in open syllables, in hiatus, and before the voiced fricatives [v] and [z]. It is always long:
A short, opener allophone [Ü] occurs in closed syllables generally, and in dissyllables with at least one medial consonant other than [v], [z] and [r], and with strongly stressed first syllable:
The third allophone is long and rounded, but is lowered to the half close position and occurs only before [r]:
Fourthly, an advanced allophone, [ÿ], occurs following a yod, a palatalized consonant or an affricate, which is often the result of an earlier palatalization:
|[bjÿ:]||'blue' (earlier [blˊÿ:]|
|[pjÿ:]||'plough' (earlier [plˊÿ:]|
It must be noted that morphology plays a part in the distribution of these allophones. For example, monosyllables such as [smü:ð] 'smooth' keep [ü:] in inflected or derivative forms: [smü:ðz] 'smooths', [ˈsmü:ðər] 'smoother'. In these forms [ü:] points to a morphological juncture following [ð], whereas in non-derivative forms such as [ˈpÜðər] 'powder' and [ˈsˊÜðər] 'shoulder' [Ü] points to a solid morpheme. The occurrence of [ÿ:] in [ˈbjÿ:te] 'beauty' and [ˈdzˊÿ:te] 'duty' over against [Ü] in [ˈbÜte] 'booty', [ˈfÜte] 'small, trifling', etc., seems to be due to the preceding palatal elements (the yod in [ˈbjÿ:te] and [dzˊ] < earlier [dˊ], in [ˈdzˊÿ:te], coupled with another, following, yod element, for [e] is undoubtedly from earlier [i].
No 5 /ɪ̈/ (between close and half-close, between front and central) has one slightly lowered variant [ë], which occurs before [r] in
The main member occurs in
No 6 /æ̈/ (open, between front and central) occurs with primary stress in most cases:
but also with secondary stress in
|[ˈbɛ:dˌ tæ̈k]||'bed tick'||[əˈræ̈θməˌtæ̈k]||'arithmetic'|
No 7 /o/ (between close and half-close, back, overrounded) is realized mostly as a long vowel:
but a short variant occurs marginally in a few words, including:
|[spok]||'spoke' (sb.)||[pok]||'small bag'|
which are thus in phonological contrast with:
|[spo:k]||'spoke' (vb.)||[po:k]||'poke' (vb)|
No 8 /ʌ/ (half-open, fully back, i.e., further back than SSB [ʌ]) is always short and occurs in:
No 9 /ɔ/ (half-open, back, rounded) is always long and occurs in:
No 10 /ɑ/ (open, back) is always long and occurs in:
|[ˈbɑ:tl̩]||'bottle (of straw, etc.)'|
For comparative and historical purposes it is convenient to treat the four diphthongal vocalic nuclei of Glenoe as distinct phonological units:
No 11 /əi/ as in
|[fəif] 'fife'||[məin] 'mine' (sb)||[pəi] 'pay'|
No 12 /ɑe/ as in
|[fɑev] 'five'||[mɑen] 'mine' (adj.)||[pɑe] 'pie'|
|[ˈtɑeər] 'tire'; 'tyre'|
No 13 /ɔe/ as in
|[bɔe] 'boy'||[ˈmɔele] 'hornless'||[stɔex] 'stench'|
No 14 /əü/, as in
|[nəü] 'knoll'||[rəül] 'roll'||[ˈfəüər] 'four'|
These diphthongs are all of the descending type, with a relatively strong stress on the first element and various glides making up the second; thus, the glide in No 11 reaches the close front position, in Nos 12 and 13, the half-close front position only, and in No 14, the close position, between front and central, rounded. The schwa as first element is quite short, especially in closed syllables ending with an unvoiced consonant:
Both elements of this diphthong may be slightly longer before voiced consonants or in open syllables and may be closely transcribed thus:
The [ɑ] and [ɔ] as first elements are, however, long and may even be over-long, which gives the impression of two syllables instead of one:
|[kɑ:e] or [ˈkɑ::e]||'cows'||[tɔ:e] or [ˈtɔ::e]||'toy'|
The distribution of /əi/ and /ɑe/ in Glenoe is typical of the US dialects. In this respect the latter are in contrast with UHE dialects, which have only one diphthong of this type. Traces of an allophonic distribution are clearly marked in Glenoe, in that [ɑe] occurs before voiced fricatives (including [r]), in hiatus, and in most final open syllables, whereas [əi] occurs in other situations, thus:
This simple picture has been disturbed, however, by various factors: the historical development of certain ME diphthongs, e.g. [əi] (<ME ei/ai), which in final open syllables is in contrast with the regular [ɑe] (< ME i ̄) and gives rise to many meaningful contrasts:
|< ME ei/ai||< ME i ̄|
Verb forms occasionally have 'irregular' diphthongs which can be ascribed to the conservative effect of analogy operating within the subset of strong verbs, thus:
probably owe their [əi] — for which [ɑe] would be expected —to the analogy with, say,
A weak verb such as dive (dived) on the other hand is always pronounced [dɑev].
A more obvious case involving analogy is heard in the singular/plural forms of:
|life:||[ləif] / [ləivz]||(i.e. NOT [lɑevz])|
|knife:||[nəif] / [nəivz]||(i.e. NOT [nɑevz])|
|wife\||[wəif] / [wəivz]||(i.e. NOT [wɑevz])|
Following /w/ or /ʍ/ the diphthong /əi/ is used to the exclusion of /ɑe/, regardless of what follows the diphthongs, thus:
|in an open syllable||[ʍəi]||'why'|
|before voiced fricative||[wəiz]||'wise'|
Dialect borrowing may explain the incidence of /əi/ in
|[sˊəi]||'shy' (the normal dialect word is [ble:t]).|
It is worth noting, in respect to this, that in a number of words which have /i/ in the dialect, US urban or regional standard speech has /əi/, presumably borrowed from local UHE dialects, thus:
|US dialect||UHE||US urban||meaning|
These US urban forms are thus phonemically opposed to
[lɑe] 'lie' (assume a recumbent position)
Note: Apart from the fourteen vocalic phonemes listed above (all of which occur with primary or secondary stress) there is a vowel quality, [ɪ], (between close and half-close, front) which, in syllables with relatively strong stress, is found only at morphological junctures, and which arises from a contextual shortening of [e:]. Examples are to be found in the fused form of verb plus negative particle, preposition or pronoun; preposition plus pronoun, etc., thus:
The same vowel quality also occurs in weakly stressed syllables:
A half-close, central vowel, [ə] — the schwa — is, of course, heard also in weakly-stressed syllables:
As mentioned elsewhere,(26) most of the Glenoe range of consonants correspond to those of SSB, but at the phonetic level Glenoe has (in addition to the various phones of the standard language) a series of breathed ejective or glottalized plosives, a full range of interdental and of alveolar-palatalized consonants and a breathed velar fricative.
This range of consonants is not, however, confined to Glenoe and the US dialects, but is found in UHE as well. Hence in seeking among the consonants differential criteria that would separate the two groups of Ulster English dialects we are thrown back on distributional contrasts. It is, for example, not the velar fricative /x/ in itself that distinguishes US from UHE but its incidence. This sound may even be common to all or most of the dialects of whatever origin in certain words of clearly Scottish provenance such as [pɛ:x] 'pant', [ˈsprɑ:xl̩] 'sprawl', etc., where the Scottish phonology was not in obvious conflict with that of a corresponding standard word. In many words with orthographic -gh(-), however, US (like Scots) has preserved [x], which in UHE (as in SSB) has become [f] or zero. This feature was found to be by far the most valuable consonantal criterion for separating US from the other dialects and provided the basis for the first part of List 1 of the phonological questionnaire.
At the morphological level, Glenoe has a few features of Scottish type which mark it off from the standard language and from UHE in particular. The plural allomorph /-n/ survives in the words:
|shoe||[śe:] or [śü: ]||[śɪ̈n]|
as well as the Umlaut plural [kɑe] 'cows'.
The morphology of some of the verbs also shows typical Scottish developments:
An analysis of Glenoe at the syntactical level provides some interesting material, but none of it offers possibilities of contrastive opposition with non-Scottish dialects. In fact the syntactical peculiarities of the Ulster dialects are frequently common to all of them and are of particular interest to the Celticists, as they very often exhibit structural patterns calqued on Gaelic originals, which are mostly the same whether it is a question of Ulster Gaelic or Scots Gaelic, for example:
|I have a thirst on me||may be compared with|
|Tá taort orm (Irish)||or Tha tart orm (Scots Gaelic)|
Influence of the US urban speech of Larne on Glenoe
Although the phonetic repertoire of Larne is almost identical with that of Glenoe the two are, in effect, different languages phonologically by virtue of the wide range of discrepancies in the incidence of the sounds in question, particularly as regards the vowels.
The importance of Larne is, of course, that it is the 'other' language of bilingual dialect speakers and that, as a local version of the standard language, it enjoys special prestige and exercises considerable influence over Glenoe; when a markedly Scots element in Glenoe has fallen into disuse, it has usually been replaced by the corresponding Larne form.
Other varieties of US compared with Glenoe
A general comparison of Glenoe with the other forms of US reveals a very close parallelism throughout all these dialects. They form, in fact, a dialectal unity which is still very close to the speech of SW Scotland although generally more archaic in character than the latter.
The similarities are naturally most marked between Glenoe and the dialects of Mid-Antrim although those of N Antrim, which extend into NE Derry (at points L1, L2 and L3) are still very close. The speech of Magilligan, county Derry (L4) represents a 19th century expansion of the US dialect of the Laggan in E Donegal. Laggan itself diverges most widely from the Glenoe norm, especially in preserving some apparent archaisms and in certain peculiarities related undoubtedly to the local Gaelic substratum, which must have differed in a few details from that of NE Ulster.(27) Laggan links up in some typical features with the Mid Ards and a large part of the W Strangford US dialect, not, obviously, because of geographical contiguity but because county Down US has also in many areas preserved the same archaisms as Donegal.
It must be emphasised that there are an overwhelming number of points of agreement among the US dialects with regard to their reflexes of ME sounds. The main criteria by which they may be separated are the various developments of ME ọ̄, which gave the subdivisions adumbrated above thus:
Reflexes of ME ọ̄ or Early Scots / Middle Scots ǖ
|Laggan||Mid Antrim||North Antrim||Meaning|
|Mid Ards||North Ards||Northeast Derry|
North Antrim and Northeast Derry have also a special treatment for the plural houses, which appears in these areas with medial [z] and the long close variant of /ü/, thus: [ˈhü:zəz], over against Glenoe-type [ˈhÜsəz] and the intermediate form [ˈhÜzəz], found in various places to the south of the [ˈhü:zəz] area.