Ulster-Scots Urban Speech in Ulster: A Phonological Study of the Regional Standard English of Larne, County Antrim*

Robert J. Gregg

PART THREE: The Gaelic Background

3.0 The task of estimating the degree to which the US speech of town and country may have been influenced by Gaelic is beset with special difficulties since the local Irish spoken until recently in the Glens of Antrim (GA) and on Rathlin Island (RI) is now virtually extinct. We are, however, fortunate in having two studies in county Antrim Gaelic written by the Swedish Celticist Nils M. Holmer,(43) who during the thirties investigated the language of the last remaining native Irish speakers in the two areas mentioned. It is largely on these studies that the following general observations are based, although some check is provided by the writer's personal observations of the native Irish speech of Fanad, Glenvar, Urris and other parts of county Donegal.(44) The Urris dialect has recently been classified along with East Ulster by Professor Heinrich Wagner,(45) who also describes the two former as sharing 'a few typical East Ulster features'(46).

3.1 Irish is a language in which phonologically speaking the consonants have primacy, a fact obscured by the normal orthography, which uses vocalic digraphs and trigraphs to indicate consonantal quality.(47) In order to keep in line, however, with the treatment of the US speech above, the Irish vowels will be dealt with first.

3.2 It is convenient for our purposes here that Holmer has in every instance made comparisons between the county Antrim Gaelic sounds and those of the local county Antrim English. He does not actually set out to give a phonemic analysis of the different dialects, but on the basis of his comments it is easy enough to see which sounds are conditioned by their phonetic environment and which are likely to have phonemic status.

3.3 Holmer's general impression is that the GA and RI vowels are 'the same as' or 'identical with' the corresponding sounds in Antrim English.(48) The following list shows the vocalic range in both Gaelic dialects, Holmer's symbols being equated with those used earlier in this study, and the numbers of the Larne and Glenoe vowels being added for reference:

HolmerGreggGlenoeLarne
No.1ii[ixˊ](ith)'eat'No.1No.1
No. 2i:i:[śi:s](síos)'down'No.1No.1
No. 3ɪɪ̈[tɪ̈gˊ](tuig)'understand'No. 3No. 2
No. 4ɛæ̈[bæ̈g](beag)'small'No. 4(non-standard Larne)
No. 5ee[dˊexˊ](deich)'ten'No. 5No. 3
No. 6e:e:[fe:m](féidhm)'need'No. 5No. 3
No. 7ɛ:ɛ:[fɛ:r](féar)'grass'No. 6No. 4
No. 8ɑɑ[ˈbɑtə](bata)'stick'No. 7No. 5
No. 9ɑ:ɑ:[ˈbɑ:tə](báta)'boat'No. 7No. 5
No. 10ɔ:ɔ:[bɔ:](bó)'cow'No. 8No. 6
No. 11ɔo̤/ʌ?[ko̤s](cos)'foot'No. 9No. 7
No. 12o:o:[mo:r](mór)'big'No. 10No. 8
No. 13ʎÜ[kÜm](cum)'peat mould'No. 11No. 9
No. 14ʎ:ü:[kü:l](cúl)'back'No. 11No. 9
No. 15əə[ˈsɑ:stə](sásta)'satisfied'No. 12No. 10

3.4 The general parallelism between the vowels of the two Gaelic dialects on the one hand and those of the US styles of speech on the other needs no emphasizing. In spite of his using the same symbol for his No. 4 and No. 7, Holmer comments on the 'considerable difference in quality between the short and long ɛ'.(49) He adds that the short version ranges 'from an open 'i' ... to a broad (short) a-sound (as in 'hat' or even as in French la; phonetically æ, a) ...' .(50) This is the sound he hears in the county Antrim pronunciation of big, did, hill, pig, in other words the Glenoe vowel transcribed above as [æ̈]. Holmer's No. 7, the long [ɛ:], as in GA and RI féar, seems to be identical with Larne and Glenoe [ɛ:], as in fair.

3.5 The exact quality of Holmer's short [ɔ] (No. 11) is a little more doubtful. He hears it in GA and RI cos 'foot' and among other questionable examples also in the local English cut, trouble, which would suggest that it was identical with Belfast [o̤], and the equivalent of Larne and Glenoe [ʌ], which is similar in tongue position but unrounded. That these latter sounds are close in articulation to long [ɔ:] is shown by the fact that her (Glenoe [hʌr]) when pronounced with a lengthened vowel, as it is in non-standard Larne, becomes [hɔ:r] and in the same way does, which is [dʌz] in Larne, becomes [dɔ:z], a pronunciation frequently heard in Belfast. Holmer claims that his long [ɔ:] is 'almost exactly the same' as the short one (GA), that is, practically the short sound sustained (RI), and that in any case it is identical with the local county Antrim English vowel in God, Cushendall, cost.(51) This long vowel agrees therefore with Glenoe and Larne [ɔ:].

3.6 One point of difference must be noted, namely that in the two Gaelic dialects vowel length is not conditioned by the phonetic environment, as it is almost exclusively in Larne and Glenoe.(52) In fact, although — as stated — Holmer does not set out to make a phonemic analysis of GA and RI and therefore has not sought out series of minimal pairs, yet using the material he provides we note that in both dialects, e.g., bata [ˈbɑtə] 'stick' is distinguished from báta [ˈbɑtə] 'boat', and in GA cuinne [ˈkʎnˊi] 'churn' from caoineadh [ˈkʎ:nˊi] 'crying',(53) merely by the length of the stressed vowel. Further, in GA ith [ixˊ], ithe [ixˊi] (eat) has a short [i], whereas oidhche [i:xˊə] 'night' has a long [i:] in similar phonetic surroundings.(54) A more detailed knowledge of these two dialects of Irish would undoubtedly reveal that a significant contrastive use of the suprasegmental feature of length was widespread in their vocalic systems. The same opposed vowel lengths also exist in Glenoe and Larne, but as already shown they function only at the sub-phonemic level and with one very marginal exception are not used to distinguish one word-form from another.

3.7 It is worth noting in connection with 1.3 (No. 8) above that the opposition [ɔ:] versus [o:] is not neutralised before [r] in GA, but is preserved and may be exploited, as in US in minimal pairs such as: bóthar [bɔ:r] 'road' versus bodhar [bo:r] 'deaf'.

3.8 Ranging further afield it is observed that similar correspondences to those described above characterise the relationship of Gaelic to the local English of county Donegal. The extremely open Glenoe vowel [æ̈] crops up as far away as the country to the south and west of Ardara in Gaelic words such as im [æ̈m] 'butter' sin [śæ̈n] 'that' as well as in the local English version of Jim, shin, etc. In north-east Donegal, the Glenvar English and Irish speech have the same fronted [ü:] and [Ü] typical of the county Antrim Gaelic and US dialects described above and have the same characteristic distribution of the two sounds in each of the languages under discussion. In other words Glenvar English uses [ü:] or [Ü] according to phonetic context, as does county Antrim US, whereas in Glenvar Irish as well as in that of county Antrim, although the same two vowels — identical in quality and quantity — occur, the length is not determined in this way. In Glenvar Irish, e.g., cúl 'back' is [kü:l], as it would also be in RI and GA, but cool is [kÜl] in Glenvar, just as in county Antrim, where the shorl [Ü] is used because the syllable is closed by [l], a consonant other than one of the voiced fricatives.(55)

3.9 Holmer lists the following consonants and semivowels for GA and RI:

Labials:pbmfv
Dentals:tdnslr
Prepalatals:(rˊ)
Palatals:ŋˊ
Velars:kgŋx(ɣ)
The Aspirate:h
Semi-vowels:jw

of these, [rˊ] has 'almost disappeared'(56) and [ɣ] is 'disappearing fast'.(56)

3.10 The old distinction between 'broad' and 'slender' seems to be lost in the GA labials, and Holmer equates them with the corresponding 'neutral' sounds in English. The dentals [t] [d] and [n] are to be identified with US [t] [d] [n] and are kept distinct from the alveolar version of these sounds used by GA speakers in loan-words from English. On the other hand [s], [l] and [r] are 'not plainly different in English and Irish'.(56) The pre-palatals [tˊ] and [dˊ] are with the younger people replaced by the affricates [tˊś] and [dˊź], which crop up also in the local English pronunciation (and, of course, in the Glenoe and Larne version) of words like tune and dew, in place of the SSB cluster [tj] and [dj]. The remaining GA pre-palatals [nˊ] [ś] and [lˊ] also occur in Larne and Glenoe, corresponding to SSB [nj] [ʃ] and [lj]. Of the GA palatals [xˊ] appears in US but only as an allophone of /x/. There may, however, be some evidence of the former existence in Larne of palatals [kˊ] [gˊ] and [ŋˊ] in the widespread fronting or Umlaut of [a] described above (57) in words such as:

back [bɛ:k]bag [bɛ:g]bang [bɛ:ŋ]

which in many other Ulster dialects would be pronounced as [bakˊ] [bagˊ] and [baŋˊ] respectively. If the palatals actually were the agents in this fronting, they must have been replaced by normal velars once the process was completed. The GA velars (apart from the obsolescent v̥) are all matched by the corresponding US sounds, as are the aspirate [h] and the two semi-vowels [j] and [w].

3.11 Summing up, it may be said that apart from the two obsolescent sounds and the three palatals discussed in the previous paragraph all the GA consonants are identical with their counterparts in the US systems at the phonetic level although different phonemic use of, for example, the dentals may be made in the two languages. Even the alternation of [h] with [x] or [xˊ] in medial positions is common to both Irish and US. On the other hand US has certain consonants which do not occur in the Gaelic dialects except in loan words, namely: [ʍ] [θ] [ð] [z] [ź] and the point-open frictionless continuant [r], although the frequently occurring flapped allophone of the latter, [r], is probably identical with what Holmer calls the GA 'plain r' or the RI 'soft, alveolar trill'.(58)

Introduction | Part 1 | Part 2 | Part 3 | Conclusion | Notes