or ì, a w or very short u is heard sometimes in Arran. It is very slight in this case, however, is found only after l and n as in laogh, glaodh, lagh, naomh, naoi, and is perhaps to be regarded properly as the passage from the ‘broad’ sounds of those liquids to the narrow vowel. The w or u in the case of e for a has arisen no doubt from the effort to pass from a consonant, associated as it was with a broad vowel, to a now slender vowel, as it is found only before those attenuated vowels in Arran and in Kintyre.
The identity of the pronunciation of ao in aon, caomh, naomh, with that of aoi in aoine, maoin, naoi, etc., in Arran and Kintyre might lead to the supposition that there the oblique case, as in so many instances, has usurped the place of the nominative, and that the forms in which ao is sounded ì, are really aoin, caoimh, naoimh. The occurrence of the same pronunciation in caomhain, and of distinct pronunciations of oblique cases is, however, rather against the supposition. While naomh, for example, is ‘n’ìmh,’ or rather ‘n’ìf,’ in Arran, naoimh strangely enough is there ‘nûimh.’ Except in two instances, chaoidh and oidhche, in which oi generally has the sound of aoi, the vowels which get this ì sound are in contact with nasal liquids m, mh, or n, and are themselves nasalised. In other cases, though in contact with nasals, ao, however pronounced, is not nasalised. So in Argyllshire generally it is where nasalisation is found that ao gets the sound of û. What we have then is this: where ao, with or without i following, is nasalised, it has been changed from ao to û in the great part of Argyllshire, and from é to ì in Kintyre and in Arran. In other districts no apparent trace of such divergence is found; the sound of û nasal or not, as the case may be, is given to ao beside all consonants. Whether the Arran and Kintyre sounds (é and ì) are modifications of the Argyll sounds (ao and û), or whether the sounds in both cases are modifications of older sounds, may be a moot question, but there need be no doubt whatever that the starting point of the divergence from é to ì, and from ao to û has been in both cases nasalisation. The cause of the divergence is
