Many years ago, in a book review in these pages, J. C. Wells remarked that ‘[w]e English phoneticians have shamefully neglected the Celtic languages lying on our doorstep: we have left their study to our colleagues from the Celtic countries themselves or from overseas’ (Wells Reference Wells1975: 43). This neglect was unfortunately abetted from the other side of the fence. The scientific study of spoken Irish was slow to develop in Ireland itself. In the first book-length study of the dialects of Irish, Thomas F. O'Rahilly (Reference O'Rahilly1932: 12–13) observed:

The Academy was of course essentially part of the British academic system. Sommerfelt's paper, in common with other works on spoken Irish in the early decades of the last century, made liberal use of non-IPA symbols and diacritics. In the 1940s the Dublin Institute for Advanced Studies published a number of phonetic studies of Irish dialects which were almost entirely in conformity with the IPA except for the use of an offset acute accent to represent palatalization, e.g. [t´]. However, Heinrich Wagner's Linguistic atlas and survey of Irish dialects (Wagner Reference Wagner1958–Reference Wagner1969), which Laver (Reference Laver1994: 73) rightly calls ‘a monumental survey of the dialects of Irish’, reverted to an in-house style of transcription, in which, for instance, [λ:] indicates a high back unrounded vowel (IPA [ɯː]) and [R] indicates not a uvular r-sound but an alveolar trill; the lambda symbol appeared as [ʎ:] (inverted ‘y’) in other works in this tradition. Those who approached such sources from outside the immediate discipline may have felt inclined to agree with Ladefoged (Reference Ladefoged1971: 2) that ‘published phonetic descriptions are often impossible to interpret accurately’.

More recently a number of works have appeared which offer transcriptions of spoken Irish which are fully in conformity with the IPA (e.g. Ní Chasaide Reference Ní Chasaide1999, Ní Chiosáin & Padgett Reference Ní Chiosáin and Padgett2012). These have, however, offered only a limited number of examples. The work under review here is described on the title page (and I translate from the original Irish) as ‘a basic course in phonetics and phonology with a description of the consonants and vowels of three dialects of Irish’, one situated on the Dingle Peninsula in the southwestern county of Kerry, another in the area west of Galway city in the midwest, and the third in north Donegal in the northwest. The first three chapters comprise the basic course, and they describe the organs of speech, phonetic transcription, the articulation of consonants, secondary articulation, the cardinal vowels and so on in a way which is broadly similar to textbooks on other languages, as one would expect; exercises accompany these chapters. Readers of this journal will doubtless be mainly interested in the detailed descriptions of dialect phonetics in the following chapters. An experienced phonetician who perused a copy of this book would be able to obtain valuable information because of the clarity of presentation, even if, as likely, they knew not a word of Irish. For instance, in the description of the sounds of the Irish of the Ceathrú Rua area west of Galway city the palatalized coronals are set out on page 58. A phonemic transcription, e.g. /tʲ/, appears at the left margin; slightly to its right a corresponding phonetic transcription is given, e.g. [tʲʰ], and a brief description follows. If you know the conventions of the IPA you will be able to interpret the symbols at least. However, all of this information, and more, can be found on the bilingual website www.fuaimeanna.ie, which accompanies this book. The numerous transcriptions and associated pronunciations can be searched in three ways. On the homepage one may click on the box ‘Explore the sounds of Irish in three dialects’ and then follow the prompt ‘Select a feature’, which scrolls down a list in alphabetical order from ‘affricate’ to ‘vowel’. Alternatively one may click on ‘Listen to sample recordings by professional broadcasters’, which gives two prompts. One of these, ‘Filter recordings by sound’, opens a list of phonemes (given unfortunately in traditional non-IPA transcriptions such as [t´], mentioned above). The other, ‘Find recordings by text’, offers a box into which one may type Irish words and find three dialectal pronunciations for them (assuming of course that one selects a word which is in the corpus). I have made extensive use of this website and can vouch that it really does work. Prompts open readily to lists and sound samples come to life instantly when clicked on. This technical excellence makes it possible for phoneticians anywhere in the world to explore the phonetics of Irish in some depth.

It may be helpful to show how this website can be used to explore one very important aspect of the phonetics and phonology of Irish. There is an almost pervasive phonological opposition in the consonant system between ‘broad’ consonants (velars along with coronals and labials which are velarized to varying degrees) and ‘slender’ consonants (post-palatals or prevelars along with coronals and labials which are palatalized to varying degrees). This has suggested comparison with Russian with good reason (e.g. Ní Chiosáin & Padgett Reference Ní Chiosáin and Padgett2012: 172), but there are significant differences between the two languages. In Russian, the phonemes /kʲ gʲ xʲ/ ‘are rare before non-front vowels and occur in this position mainly in loanwords and foreign names’ (Yanushkevskaya & Bunčić Reference Yanushkevskaya and Bunčić2015: 223). The palatalization contrast might therefore be said to be weak for the velars of Russian, from the point of view of phonology. In Irish, on the other hand, /kʲ gʲ xʲ ɣʲ/ occur freely in combination with front and back vowels, as may be seen by calling up the examples on this website. Irish, however, shows a weakness regarding the palatalization of labials from a phonetic point of view. Initial ‘slender’ labials tend not to be phonetically palatalized before front vowels but rather show coarticulation with them; this of course equates to true palatalization before /iː/ or /i/. This can be seen by calling up the examples of /bʲ/ in the way described above. The first three examples are bé ‘muse’ (noun), béal ‘mouth’ and bean ‘woman’. These are transcribed phonemically as /bʲeː/, /bʲeː ˠ/ and /bʲa ˠ/, respectively, and justifiably so since the neutral initial labial is in complementary distribution with palatalized labials in medial and final positions, but none of these words contains a palatalized consonant in the sense of having an accompanying high front gesture. Suitable phonetic transcriptions would be [b eː], [b eː ˠ] and [b a ˠ] (leaving aside the varying pronunciations of the long mid vowel). The initial CV in the second of these words sounds quite different from that of Russian белый bely ‘white’, which tends to have a distinct high front gesture after the labial, something which Irish never has before /eː/ or /e/. The ‘slender’ labials of Irish are characterized by lip spreading and this contributes to the contrast where true palatalization is lacking; in addition the ‘broad’ labials are velarized. The only example here of final /bʲ/ following a back vowel is in the genitive form ciúib of the borrowed word ciúb ‘cube’ but it contains a palatalized labial in the true sense, with an accompanying [ɪ] gesture, as may be heard clearly in the sound clips. There are dialect differences with regard to ‘slender’ labials before back vowels. The Galway speaker pronounces beo ‘alive’ with a quite neutral labial consonant followed by a complex [ oː] nucleus, whereas the Donegal speaker has a clear high front gesture which might justify a phonetic transcription [bʲɔː] as given here, but the high front transition sounds rather prominent to me and [b jɔː], with a separate [j] segment, might be preferable. The latter is the traditional view in Irish-language studies (e.g. McKenna Reference McKenna, hUiginn and Catháin2001: 156), and is also assumed for Scottish Gaelic. It is interesting though that Ladefoged et al. (Reference Ladefoged, Ladefoged, Turk, Hind and Skilton1998) attribute palatalized labials to the speech of Greater Bernera on the west coast of Lewis. Indeed the spectrogram which they give of /pʲa nˠəx k/ beannachd ‘blessing’ (p. 14, solidi supplied here) suggests that that is right, for the area in question at any rate. A more extensive study of the pronunciation of words such as beo ‘alive’ in Donegal Irish would be desirable.

There are other dialectal differences relating to palatalization. In words such as cuid ‘part’ and deas ‘nice’ the Galway and Donegal speakers have laminal postalveolar or alveolo-palatal plosives, differing in the slight affrication which usually accompanies them; that of the Galway speaker has a [z] effect whereas one hears more of a [ʑ] from the Donegal speaker (and indeed in some northern dialects /dʲ/ has gone further and become [dʑ]). In cuid ‘part’ the Kerry speaker has a palatalized alveolar (apical, with some laminal contact also perhaps), roughly the articulation shown in a sagital cross-section of a palatalized voiced alveolar stop [dʲ] in Laver (Reference Laver1994: 323). In deas ‘nice’, however, the Kerry speaker does not have a particularly high front transition between the plosive and the low vowel. In general, palatalization effects are less prominent for coronals in Munster dialects.

What was historically a palatalized [sʲ] has become a ‘hushing’ sibilant in modern Irish and Scottish Gaelic, as may be heard in words such as seacht ‘seven’ and cúis ‘cause’ here. The English word whisk(e)y was borrowed in Scotland from uisce beatha ‘water of life’ (aqua vitae) before this change occurred. Greene (Reference Greene1966: 23) remarks that in Irish ‘uisce “water” is now pronounced as if it were isce and could only give English ishkey!’. This palatalized consonant has traditionally been transcribed as [ʃ] (or even /s´/) but the articulation tends to be quite alveolo-palatal for most dialects. Ní Chasaide (Reference Ní Chasaide1999) was to my knowledge the first to transcribe this sound as [ɕ] (for a Donegal speaker), and rightly so. The practice on this website is to transcribe [ɕ] in all positions for Donegal and Galway but [ʃ] in all positions for Kerry. However, the Kerry [ʃ] has a very clear quality compared to English [ʃ], and a more detailed transcription might be [ʃʲ] (so that [ɕ] in Irish will not simply be a transcriptional variant of [ʃʲ]). Indeed, I think that when this fricative is next to high vowels and front mid vowels at least it would best be transcribed [ɕ] for the Irish of Kerry also.

Palatalized ‘r’ has also undergone significant historical changes. It was depalatalized in initial position everywhere, so rí ‘king’ is [ɾˠiː]. A palatalized tap [ɾʲ] is found in all dialects in initial clusters, but the Kerry speaker here has accompanying friction in words such as cré ‘earth’ and creid ‘believe’. In intervocalic and final position he has a voiced fricative sound which is typical of Munster speakers nowadays. The Galway speaker has this fricative in fuair ‘got’ but not in cuir ‘put’ or generally elsewhere. It will be noticed that the Donegal speaker has a non-palatalized tap in fuair ‘got’ and other items. In her dialect palatalized ‘r’ has changed to an approximant [j], but speakers often replace this with a neutral tap or even approximant [ɹ] when they wish to avoid confusion, or in an attempt to standardize. It would have been preferable had she been advised to use her familiar [j] in these recordings.

This leads to an observation about the word sets chosen to illustrate the sounds. Under palatalized ‘r’ (accessed as /r´/) there appears a word cáireach ‘dirty’. This is an Ulster word, historically cáidheach [kaːjax], and the spelling cáireach is a false correction of the sound change [ɾʲ] > [j], mentioned above. This ahistorical form was then presented to the Galway and Kerry speakers and they pronounced it according to their dialects. Such a procedure risks creating ghost forms which will then be cited in the literature. Another problem is that the orthographical forms which appear at the left margin of each page conform to the state-sponsored standard written form of the language, but the sound clips are often at variance with this. For instance, the Kerry speaker pronounces gaineamh ‘sand’ with a lightly-palatalized final [vʲ], because in his dialect the prepositional-case form gainimh has ousted the basic form gaineamh. It would have been preferable to give both forms on the left margin in this instance. Users of the website who are familiar with written Irish are liable to be confused by such discrepancies. A more extreme example arises with oir ‘fit’ (verb), where the Galway speaker replaces the written prompt with his dialectal feil. On page 4 of the book the author acknowledges the existence of such mismatches, but it is unfortunate that they were allowed to occur. The speakers are all professional broadcasters and would have been able to adapt to some systematic way of dealing with them.

The great merit of this book and the website is that they present a very large number of Irish words in IPA transcription with sound clips which enable experienced phoneticians to make up their own minds about what they are listening to. Some aspects of the presentation may be faulted. There are probably too many neologisms at the expense of traditional vocabulary. For instance, I am surprised that the word áit ‘place’ was not included. It would have been a very good example of the different ways of saying the long low vowel and also of final /tʲ/. It should also be pointed out that for the Galway speaker historical short low vowels have been lengthened (as in bean ‘woman’), although this is not marked in the transcriptions on the website. And the mismatches mentioned above are often irksome. Nevertheless one cannot but applaud the fact that a large amount of spoken Irish has been made available on the internet in a format which will enable phoneticians throughout the world to explore the sounds of this language and form their own opinions about what they hear.