First, they are not diphthongs. Those are digraphs (using two characters typically associated with vowels) representing single vowels, not diphthongs. Diphthongs are two vowels working as a nucleus of a single syllable (regardless of how they are written down), that’s not what happens here.
An example of a diphthong in Irish would be
/ai/ in
adhmad /aiməd/ ‘wood, timber’ (even though it is written using consonant characters: ⟨adh⟩). There are no
/oi/ or
/ui/ diphthongs in Irish. The diphthongs you’ll find in Irish are:
/əi/,
/ai/,
/əu/,
/au/,
/ou/,
/iə/,
/uə/,
/ia/…
Second, yes, there are some rules, but they aren’t obvious and they vary by dialect. The ones given by Wikipedia seems to be OK for Connacht (but I don’t know Connacht Irish that well to really assess them). You can look at
this thread on IrishLanguageForum.com, and also at examples on
fuaimeanna.ie.
As for dialectal variation fuaimeanna.ie eg. gives Ulster
/kom’/, Connacht
/kim’/, Munster
/kiːm’/ for
coim. Or
scoil /skol’/ for both Ulster and Munster but
/skel’/ for Connacht.
One can try to get some rules from Diarmuid Ó Sé’s
Gaeilge Chorca Dhuibhne (English translation below):
Gaeilge Chorca Dhuibhne, § 29, p. 38 wrote:Dáileadh na ngutaí
29 Tá dáileadh na bpéirí gutaí gearra i, u agus e, o go mór faoi réir ag cáilíocht na gconsan. Beidh sé áisiúil idirdhealú a dhéanamh idir ‘comharsanacht chaol’ (nuair nach bhfuil aon chonsan leathan in aice leis an nguta), ‘comharsanacht leathan’ (nuair nach bhfuil aon chonsan caol in aice leis an nguta) agus ‘comharsanacht mheasctha’ (nuair atá consan leathan ar thaobh amháin den ghuta agus consan caol ar an taobh eile). Maidir leis na gutaí arda, is i a bhíonn i gcomharsanacht chaol, m.sh. t´iɡ´ tigh, in´ɪʃ inis, k´i cith; agus is u a bhíonn i gcomharsanacht leathan, m.sh. tuɡən tugann, uməd iomad, ɡu guth. Tá roinnt eisceachtaí ann, áfach: t´r´u triuf, li luich, ɡirt´ goirt (iorla). Tá rogha idir u ~ i in uməl ~ iməl imeall, ru ~ ri rith. Tá an dá fhuaimniú ri ~ ru ag roth chomh maith. Bíonn i idir consan leathan agus consal caol, m.sh. ɡib´ goib, kid´ cuid, liɡ´ luigh; agus idir consan caol agus déadach nó r leathan, m.sh. ɡ´itə giota, f´is fios, b´ir bior, m´in mion. Ach is u a bhíonn idir consan caol agus liopach, coguasach nó l leathan, m.sh. ɡ´ubəl giobal, t´uv tiubh, p´r´uk prioc, f´ulər fiolar; tá i coitianta in t´r´ibəˈloːd´ trioblóid áfach. Tá an fuaimniú g´ibəl ag roinnt daoine áfach. Maidir leis na gutaí e, o, is e a bhíonn i gcomharsanacht chaol, m.sh. m´el´ meil, el´ɪ eile, t´e te; agus is o a bhíonn i gcomharsanacht leathan, m.sh. mol mol, olə ola, do do. Eisceacht is ea ʃo seo. Bíonn e idir déadach nó r, l leathan agus consan caol, m.sh. den´əs doimhneas, sev´ɪr´ saibhir, tev´ toibh, lev´ loibh, rev´ raibh; ach o in tol´ toil. Bíonn o idir consain leathana eile agus consan caol, m.sh. sɡol´ scoil, kod´ɪr´əv coidreamh; agus idir consan caol agus consan coguasach leathan, m.sh. b´oɡ beag, d´ox deoch. Ní miste a lua gur líonmhaire go mór i ná e idir consan leathan agus consan caol. De bharr na n-eisceachtaí éagsúla (agus ar chúiseanna eile, b’fhéidir) ba dheacair a áiteamh gur baill d’aon fhóinéim amháin iad i agus u, ná e agus o. Tá srianta láidre ar dháileadh na ngutaí sin áfach, agus is iad na srianta sin is cúis le go leor de na malartaithe gutaí a luaitear thíos, m.sh. muk muc, mik´ɪ muice; obɪr´ obair, eb´ɪr´ɪ oibre.
[The distribution of the vowels
29 The distribution of the short vowels i, u and e, o is mostly governed by the quality of the consonants. It’ll be convenient to make a distinction between ‘slender neighbourhood’ (when there is no broad consonant next to the vowel), ‘broad neighbourhood’ (when there is no slender consonant next to the vowel) and ‘mixed neighbourhood’ (when there is a broad consonant on one side of the vowel and a slender consonant on the other side). Regarding the high vowels, it is i in a slender neighbourhood, eg. t´iɡ´ tigh, in´ɪʃ inis, k´i cith; and it is u in a broad neighbourhood, eg. tuɡən tugann, uməd iomad, ɡu guth. There are some exceptions, though: t´r´u triuf, li luich, ɡirt´ goirt (plural). There is a choice between u ~ i in uməl ~ iməl imeall, ru ~ ri rith. roth has the two sounds ri ~ ru as well. i happens between a broad consonant and a slender one, eg. ɡib´ goib, kid´ cuid, liɡ´ luigh; and between a slender consonand and a broad dental or r, m.sh. ɡ´itə giota, f´is fios, b´ir bior, m´in mion. But it is u between a slender consonant and a broad labial, velar or l, eg. ɡ´ubəl giobal, t´uv tiubh, p´r´uk prioc, f´ulər fiolar; It’s commonly i in t´r´ibəˈloːd´ trioblóid though. Some people have the pronunciation g´ibəl though. Regarding the vowels e, o, it is e in a slender neighbourhood, eg. m´el´ meil, el´ɪ eile, t´e te; and it is o in a broad neighbourhood, eg. mol mol, olə ola, do do. ʃo seo is an exception. It’s e between a broad dental or r, l and a slender consonant, eg. den´əs doimhneas, sev´ɪr´ saibhir, tev´ toibh, lev´ loibh, rev´ raibh; but o in tol´ toil. o is between other broad consonants and a slender consonant, eg. sɡol´ scoil, kod´ɪr´əv coidreamh; and between a slender consonant and a broad velar consonant, eg. b´oɡ beag, d´ox deoch. It should be mentioned that i is hugely more common than e between a broad consonant and a slender one. Because of various exceptions (and of other reasons, possibly) it would be difficult to argue that i and u are allophones of a single phoneme, nor e and o. There are strong constraints on the distribution of those vowels, though, and those constraints are the reason of a lot of the vowel alternations that are mentioned below, eg. muk muc, mik´ɪ muice; obɪr´ obair, eb´ɪr´ɪ oibre.]
The interesting parts are the ones about mixed neighbourhoods with a broad consonant at the beginning, and from reading that I’d say in Munster Irish:
- ui is generally always /i/,
- oi is /e/ after a dental consonant (d, t, n, s) or r or l (with the exception of toil /tol’/) and in the beginning of a word (like oibre /eb’ɪr’ɪ/),
- oi is /i/ generally in most other cases, especially when etymologically from /u/ (see below), eg. cnoic /knik’/,
- but oi is /o/ other cases, if it’s not /i/, eg. scoil /skol’/ (when etymologically from old /o/, see below).
The problem is that what is written as ⟨oi⟩ etymologically comes from two different vowels: either
/o/ (in
scoil, which historically has been always written either
scoil or
sgoil) or
/u/ (eg. in
cnoic, gen. and plural of
cnoc, which historically had been written
cnuic) – it seems to me that it’s
/i/ when it etymologically comes from old
/u/ (since
ui is always
/i/) but either
/o/ or
/e/ when it comes from older
/o/. But then Ó Sé gives eg.
kin´ɪ coinne also has
/i/, so I’d just assume
/i/ generally and if not, use
/e/ after a dental (or
/r/,
/l/) and
/o/ otherwise.
Also, it’ll always be
/ɪ ~ ə/ when unstressed, and in Munster that means eg. when the second syllable is long (eg.
coinín /kɪˈn’iːn’/).