Irish pronunciation

[Ciarán12]

I'm obviously not very good at explaining what I mean, but bare with me. Here is a chart of certain sounds that I have found to occur in what I think of as palatalised , velarised and "normal" forms. I'm sure you'll tell me that I'm wrong in assigning those terms to those groups of words, but can you deduce whatever phenomenon it is I am actually talking about from the chart? I definitely pronounce the sounds differently in each group...

[kevin]

..., but bare with me.

Sorry for the off-topic, but isn't it "bear"?

[Ciarán12]

..., but bare with me.

Sorry for the off-topic, but isn't it "bear"?

Is it? I'm terrible at spelling (I'm slightly dyslexic)...

[linguoboy]

I'm obviously not very good at explaining what I mean, but bare with me. Here is a chart of certain sounds that I have found to occur in what I think of as palatalised , velarised and "normal" forms. I'm sure you'll tell me that I'm wrong in assigning those terms to those groups of words, but can you deduce whatever phenomenon it is I am actually talking about from the chart?

Sorry, no--there's no consistency to the columns. In general, both the second and third contain velarised (plus labialised) consonants, but then you have ní in the "plain" column when it's clearly palatalised. I can't figure out how your pronunciation of the /nʲ/ could differ significantly from the /nʲ/ in neart.

To reiterate, when you produce a palatalised consonant, the middle of your tongue is raised a bit towards the hard palate. When the consonant is velarised, the back of your tongue is raised towards the soft palate (or "velum") at the back of your mouth; if the consonant is labial, then the lips will be rounded as well, resulting in labialisation.

When the following vowel has the same quality, then the transition to it is smooth. That is, a high front vowel like [iː] has the middle of the tongue near the hard palate, so in a word like ní, the chief adjustment is just the opening of the mouth. (Frequently, some air continues to escape through the nose, resulting in partial or complete nasalisation of the vowel, i.e. [nʲĩː].)

However, when the vowel has a different quality, then you have to adjust the tongue position between the moment of release and the peak of sonority which forms the nucleus of the vowel. It is during this transition that you are likely to hear some sort of glide--and the greater the distance between the two tongue positions, the more distinct the glide.

So with a word like naoi you're going from velarisation (a high back tongue position) to a palatal vowel (high front). Air keeps escaping over your tongue as it's shifting and it's this phenomenon which is often represented with [ɰ̯]. Of course, all glide symbols like this (or [w] or [j]) are broad approximations. The vowel space is two dimensional, and there are an infinite number of points of transition between any two positions.

Instructors often tell you that if you get the quality of the consonants correct, then the transitions will take care of themselves, and that's very true. But this can also be turned on its head: inserting the transitions can often help you to get the qualities correct. This is what I was getting at when I told kevin:

What you say is true, kevin, but one approach the proper pronunciation is to concentrate of producing the artefacts and then try to pare back their expression until you're really only modifying the consonant itself.

The downside to this approach is if you get stuck at the stage of producting the artefact but not the effect on the consonant itself. That is, I wonder if with neart you might actually be producing an unpalatalised consonant followed by a palatal glide. This is possible to do--as I say, it's been observed in Russian speech--and would result in the /nʹ/ in /nʹart/ sounding different from the /nʹ/ in /nʹiː/ (and more like the /m/ in /mə/ or the /d/ in /dow̃ənʹ/

Try to pay attention to where you're holding your tongue during each stage in the pronunciation of these words. What similarities and differences do you notice there? When does it feel to you like it starts to adjust for the following vowel? Does it adjust at all? Thinking about these questions might give you a better idea what is actually going on with your pronunciation.

[Llawygath]

Sorry to go back to the earlier topic.

First thing that leaps out at me: an is unstressed. So all those occurrences of /an/ should really be /ən/.

But I thought it was phonemically /an/. Ciaran seemed to be saying it was. I'm not marking stress because I'm not entirely clear on which dialect does what and I'd probably make a mess of it.

(Or even /ənˠ/, unless you're assuming a class of "neutral" consonants who take their broad/slender quality from the adjoining segment--which will make it tricky to deal with cases like an ionga [ɪˈnʲʊ̟ŋə].) Alternatively, you could indicate stress (which can otherwise be assigned mechanically) and assume a rule whereby unstressed vowels reduce to shwa.

I'll have to assume neutral consonants. I'm not marking any stress yet as explained earlier, so no luck there.

An tsiopa should be /ən(ʲ)ˈtʲupˠə/. (That is, however you decided to deal with stress and assimilation, there's no phonemic glide and the stressed vowel is /u/.)

Whoops. I wasn't sure if it was a glide or palatalization. Why is it /u/? Is there a rule for that?

In codladh, the /d/ assimilates. The only argument for showing it in transcription is that some varieties show metathesis, e.g. Ballymacoda [ˈkɔl̪ˠd̪ˠə].

Okay then, I won't show it.

I'm not sure where you got the idea that Irish /ʃ/ is alveolo-palatal rather than palato-alveolar.

Everything else I've read says it's alveolo-palatal. Where did you get the idea that it was postalveolar?

On the other topic: whatever other errors Ciaran made in his chart, there can't be any such thing as a distinction between broad and plain /g/, /k/ and /x/. No way. What would that even mean? Nobody could pronounce that. If you did have a distinction it would have to consist of some sort of glide after the consonant, which wouldn't really count.

[Ciarán12]

On the other topic: whatever other errors Ciaran made in his chart, there can't be any such thing as a distinction between broad and plain /g/, /k/ and /x/. No way. What would that even mean? Nobody could pronounce that. If you did have a distinction it would have to consist of some sort of glide after the consonant, which wouldn't really count.

I refer you back to the earlier confusion on my part about what the difference is between a palatalised consonant and a consonant with a brief j-glide after it and a velarised consonant and a consonant with a brief ɰ-glide after it. I still don't really understand.

[Llawygath]

On the other topic: whatever other errors Ciaran made in his chart, there can't be any such thing as a distinction between broad and plain /g/, /k/ and /x/. No way. What would that even mean? Nobody could pronounce that. If you did have a distinction it would have to consist of some sort of glide after the consonant, which wouldn't really count.

I refer you back to the earlier confusion on my part about what the difference is between a palatalised consonant and a consonant with a brief j-glide after it and a velarised consonant and a consonant with a brief ɰ-glide after it. I still don't really understand.

Sorry. I'm not sure I understand either, but the gist of it all seems to be that a palatalised* consonant has the /j/ pronounced at the same time as it and a consonant with a j-glide after it has the /j/ pronounced immediately after instead, and same for velarized/ɰ-glide consonants.

[off-topic]I'm vacillating on how to spell stuff. I think I like -ise and -yse better than -ize and -yze, but I was brought up with the latter. I also can't decide on theater or theatre &c. The list goes on.[/off-topic]

[Ciarán12]

I refer you back to the earlier confusion on my part about what the difference is between a palatalised consonant and a consonant with a brief j-glide after it and a velarised consonant and a consonant with a brief ɰ-glide after it. I still don't really understand.

Sorry. I'm not sure I understand either, but the gist of it all seems to be that a palatalised* consonant has the /j/ pronounced at the same time as it and a consonant with a j-glide after it has the /j/ pronounced immediately after instead, and same for velarized/ɰ-glide consonants.

But in the case of stops like /g/ and /k/ you can't hear the effect of position of the tongue until after the release anyway, so what's the difference? And /x/ does has a different quality to it if you follow it immediately with a ɰ-glide, or at least it does to me.

[off-topic]I'm vacillating on how to spell stuff. I think I like -ise and -yse better than -ize and -yze, but I was brought up with the latter. I also can't decide on theater or theatre &c. The list goes on.[/off-topic]

I alternate somewhat as well, but I obviously grew up with -ise and -yse so I tend towards those more. I don't think it really matters.

[linguoboy]

But I thought it was phonemically /an/.

As I said, it depends how you treat stress. Perhaps Ciarán can think of some circumstances where an takes primary stress and surfaces as [ˈanˠ], but I can't.

I'm not marking stress because I'm not entirely clear on which dialect does what and I'd probably make a mess of it.

Pretty simply, really: Primary stress on the initial syllable of words. Only Munster has the stress shift to a long vowel in the second or third syllable. And only Ulster has the shortening of long vowels in unstressed syllables, though all dialects reduce unstressed short vowels.

But, like I said, if you apply these rules without finding some way to indicate, for instance, that /aN/ (i.e. ann) receives primary stress but that /an/ does not, you'll get incorrect results.

An tsiopa should be /ən(ʲ)ˈtʲupˠə/. (That is, however you decided to deal with stress and assimilation, there's no phonemic glide and the stressed vowel is /u/.)

Whoops. I wasn't sure if it was a glide or palatalization. Why is it /u/? Is there a rule for that?

The rule (for Munster, at any rate) is that io represents /'i/ before r, s, t, th, and d and is otherwise equivalent to iu. So siota /ˈs'ita/ but sioc /ˈs'uk/ (i.e. [ˈʃʊ̟k]).

I'm not sure where you got the idea that Irish /ʃ/ is alveolo-palatal rather than palato-alveolar.

Everything else I've read says it's alveolo-palatal. Where did you get the idea that it was postalveolar?

ʃ This is a voiceless palato-alveolar fricative and corresponds to ʒ in formation.

ʒ This is a voiced palato-alveolar fricative formed with the tip of the tongue against the lower teeth, the blade being raised toward the back of the teeth-ridge, while the front is raised toward the hard palate.

(Ó Cuív, The Irish of West Muskerry, Co. Cork, p. 41.)

Note that he explicit describes it as coronal ("with the tip of the tongue against the lower teeth"). What are your sources which call it "alveolo-palatal".

[linguoboy]

I refer you back to the earlier confusion on my part about what the difference is between a palatalised consonant and a consonant with a brief j-glide after it and a velarised consonant and a consonant with a brief ɰ-glide after it. I still don't really understand.

The glide is a secondary formation. It doesn't occur after all consonants. And the tongue position varies. As long as I've got Ó Cuív handy:

ʲ[*], a glide having the spread-lip position and the tongue-raising of a palatal consonant. The tongue-raising is less before more open vowels than before more closed vowels. Thus before uː, u, the tongue is raised towards the iː position, e.g. fʹʲuː fiú, tʹʲuv tiugh, while before oː, o, etc. it is raised to approximately the e position, e.g. bʹʲoːrʹ beoir, bʹʲog beag. It should be remembered that the tongue-raising for the ʲ glide is less than for the fricative consonant j before the corresponding vowels.

[*]In Ó Cuív's notation, ʹ is used to represent palatalisation and ʲ represents a palatal glide.

[Ciarán12]

This insecurity about my pronunciation of Irish has been bothering my for a while now, and I'd like to get to the bottom of it. I'm fairly sure that the majority of what seems odd on paper about my pronunciation can be attributed to my lack of understanding of IPA. As such, I'd like to by-pass it an just go for making recordings of how I pronounce words. This page contains a Lexical set for Irish, which it explains is used to distinguish the various features of pronunciation amongst the dialects. If I make a recording of me saying all of these words (the underlined ones in question, not the whole sentences), would I (or rather you, linguoboy) be able tell me how they match up to the dialects (or help me out with a transcription maybe)? As I've said before, it's not the end of the world if I don't have the pronunciation of any particular dialect, but I'd like a reference point for any further discussions on Irish phonology, and this is the only way I can think of doing that.

[mōdgethanc]

On the other topic: whatever other errors Ciaran made in his chart, there can't be any such thing as a distinction between broad and plain /g/, /k/ and /x/. No way. What would that even mean? Nobody could pronounce that. If you did have a distinction it would have to consist of some sort of glide after the consonant, which wouldn't really count.

What are you talking about? It's completely possible to have palatalized velar consonants. Polish, Greek, Russian, Icelandic and Scottish Gaelic all have them. They aren't any harder than any other palatalized consonants are.

In Irish I'm told they're actually palatals, but I doubt this, at least for the stops. True palatal stops are rather uncommon; they tend to be affricated like in Hungarian.

[linguoboy]

On the other topic: whatever other errors Ciaran made in his chart, there can't be any such thing as a distinction between broad and plain /g/, /k/ and /x/. No way. What would that even mean? Nobody could pronounce that. If you did have a distinction it would have to consist of some sort of glide after the consonant, which wouldn't really count.

What are you talking about?

"broad" = "velarised". (The usual term for "palatalised" in Irish phonetics is "slender" [caol].)

[mōdgethanc]

Well, shit.

[Ciarán12]

From this thread:

Irish doesn't really have a "standard" accent, only a standard grammar and orthography. <á> as either [ɑː] or [ɔː] both look okay to me. Word-final /j/ as [ɟ] is a Munster feature, yes. He sounds pretty native to me.

Doesn't it? https://en.wikipedia.org/wiki/Help:IPA_for_Irish seems to be an attempt at a pan-dialectal accent for transcribing words, and this is the accent I see prescribed in the books on Irish I've looked at (admittedly not many). Of course native speakers are going to speak with whatever accent they grew up with, and non-natives might pick one to imitate, but this looks about as standard as it gets to me.

As I'm sure you well know, there's a difference between phonemic and phonetic transcription, so it's entirely possible to transcribe the phonemes of Irish in a pan-dialectal way, but there realisations are going to be different. Irish phonemics have a standard version, but not their phonetic realisation. I'm not sure what dialect specifically it is that they've taken as "standard" for that chart, but there are definitely some things I find strange about it. Firstly, /ɣʲ/ is not [j] for me at least, and I don't recall hearing it as [j] often, so that's kind of weird. /n̠ʲ/ would be [ɲ], /ŋʲ/ would be [ŋʲ] and /ɾʲ/ would be [ʐ] (and definitely not [ɾʲ] like the chart suggests!).

As for the vowels <a> and <á> just being short and long versions of the same sound hasn't been the case in most of Ireland since Old Irish, or possibly Middle Irish. Where that chart has /ɔ/ and /ʊ/ I have [ʊ] in both cases. That chart looks anything but standard to me. And all of the Irish learning materials I've come across have taught different dialects, with different pronunciations (one of the things that make learning Irish so difficult is the lack of a standard pronunciation that is generally followed by textbooks).

[mōdgethanc]

As I'm sure you well know, there's a difference between phonemic and phonetic transcription, so it's entirely possible to transcribe the phonemes of Irish in a pan-dialectal way, but there realisations are going to be different. Irish phonemics have a standard version, but not their phonetic realisation. I'm not sure what dialect specifically it is that they've taken as "standard" for that chart, but there are definitely some things I find strange about it. Firstly, /ɣʲ/ is not [j] for me at least, and I don't recall hearing it as [j] often, so that's kind of weird. /n̠ʲ/ would be [ɲ], /ŋʲ/ would be [ŋʲ] and /ɾʲ/ would be [ʐ] (and definitely not [ɾʲ] like the chart suggests!).

As for the vowels <a> and <á> just being short and long versions of the same sound hasn't been the case in most of Ireland since Old Irish, or possibly Middle Irish. Where that chart has /ɔ/ and /ʊ/ I have [ʊ] in both cases. That chart looks anything but standard to me. And all of the Irish learning materials I've come across have taught different dialects, with different pronunciations (one of the things that make learning Irish so difficult is the lack of a standard pronunciation that is generally followed by textbooks).

I suppose the point is it's not based on one particular dialect but is a (somewhat artificial) abstraction of all of them.

I don't know what you think the slender counterpart of /ɣ/ is; it makes sense to me that it would be [j] or maybe [ʝ]. I understand dialects vary on how many nasal and lateral sounds they have, since there were four kinds of /n/ and four /l/ sounds in Old Irish and they merged with each other in different ways, so all have at least /nˠ/ and /nʲ, and some might have /ɲ/. I haven't heard any have /ŋʲ/, but Scottish Gaelic does, so maybe. I always thought /ɾʲ/ was exactly what it looks like, but if I understand righly you say it's [ʐ] in every dialect, not just Munster. As for the vowels, I think the long /a/ is usually backed to [ɑː], but they both have a lot of allophones. The short vowels /ɪ, ʊ/ and /ɛ, ɔ/ seem to alternate a lot too.

[Ciarán12]

I suppose the point is it's not based on one particular dialect but is a (somewhat artificial) abstraction of all of them.

I'm just not familiar with what particular abstraction this is. It's not a standard I've come across before.

I don't know what you think the slender counterpart of /ɣ/ is; it makes sense to me that it would be [j] or maybe [ʝ].

It's [ʝ] for me.

I understand dialects vary on how many nasal and lateral sounds they have, since there were four kinds of /n/ and four /l/ sounds in Old Irish and they merged with each other in different ways, so all have at least /nˠ/ and /nʲ, and some might have /ɲ/.

Well, the way I speak (and also the only way I'm aware of in any dialect) has [ɲ] for /nʲ/. Some may also have [ɲ] for /ŋʲ/ (I have [ŋʲ] for /ŋʲ/). As for differing /n/ phonemes, I think some dialects have a separate /n/ for <nn> (I think I remember linguoboy saying in Munster it was [ŋ]), but for me <nn> only changes the pronunciation of a preceding <a> or <o>. For example, to me <sin> ("this") and <sinn> ("we") are both [ʃɪɲ] (but maybe some dialects differentiate between them) .

I haven't heard any have /ŋʲ/, but Scottish Gaelic does, so maybe. I always thought /ɾʲ/ was exactly what it looks like, but if I understand righly you say it's [ʐ] in every dialect, not just Munster.

I vaguely remember being told that it is realised differently in some dialects, but I've never actually heard a different realisation myself. I dug up three videos of native speakers speaking the three main dialects for comparison:

Munster
http://www.youtube.com/watch?v=NQPtqDBS4MM

For specific examples, listen to díriú at 0:16 and athuair at 0:21

Connacht
http://www.youtube.com/watch?v=LhqC3Vd0Qzo

Specific examples: scéal na haimsire at 0:02 and TG4 go gairid at 0:05

Ulster
http://www.youtube.com/watch?v=Ml_FRgT08OU
Specific example: Is mise Áine Ní Bhreisleáin at 0:04

These all sound more or less like [ʐ] to me.

As for the vowels, I think the long /a/ is usually backed to [ɑː], but they both have a lot of allophones.

True, but the only dialect where <á> even comes close to having the same quality as <a> is in Ulster Irish (and I don't think they actually are the same vowel, just close), and Ulster Irish is definitely not the standard, if there could be said to be one.

The short vowels /ɪ, ʊ/ and /ɛ, ɔ/ seem to alternate a lot too.

Possibly, I think I'd need examples to be sure though.

[linguoboy]

I suppose the point is it's not based on one particular dialect but is a (somewhat artificial) abstraction of all of them.

I thought we'd covered this before, but perhaps not. Even though that article doesn't cite the authoritative source (i.e. Lárchanúint don Ghaeilge by Dónall P. Ó Baoill), I've always assume the pronunciations given are based on the Lárchanúint, an attempt to create a compromise standard of pronunciation in the 80s. No one I know really uses it; serious learners always model their speech on a particular dialect and those who have been through the school system (like Ciarán) tend to mix and match depending on what teachers they have. Thus they also end up with a compromise, but a different one than the formally-constructed version.

Well, the way I speak (and also the only way I'm aware of in any dialect) has [ɲ] for /nʲ/. Some may also have [ɲ] for /ŋʲ/ (I have [ŋʲ] for /ŋʲ/). As for differing /n/ phonemes, I think some dialects have a separate /n/ for <nn> (I think I remember linguoboy saying in Munster it was [ŋ]), but for me <nn> only changes the pronunciation of a preceding <a> or <o>. For example, to me <sin> ("this") and <sinn> ("we") are both [ʃɪɲ] (but maybe some dialects differentiate between them).

According to authoritative sources, most dialects distinguish [ɲ] from [nʲ]. IME, most native English-speakers find this very difficult to perceive, because they can't even tell [ɲ] from [nj]. I think this chart misquotes Ó Cuív, however, because I own his book and he quite clearly states that fortis slender /n/ is [ŋʲ], not [ɲ] (although when broad it is dental and velarised rather than velar--another tricky distinction for Anglophones), and this is what I hear from the native speakers on the Pimsleur recordings. Thus I distinguish sin ([ʃɪnʲ) from sinn ([ʃiːŋʲ]).

[mōdgethanc]

I'm just not familiar with what particular abstraction this is. It's not a standard I've come across before.

It's probably because it's intended for foreign learners who might not have access to native speakers, I guess.

It's [ʝ] for me.

Maybe it varies between dialects, or maybe they're in complementary distribution. Hell, Spanish has both [ʝ] and [j], why not Irish?

Well, the way I speak (and also the only way I'm aware of in any dialect) has [ɲ] for /nʲ/. Some may also have [ɲ] for /ŋʲ/ (I have [ŋʲ] for /ŋʲ/). As for differing /n/ phonemes, I think some dialects have a separate /n/ for <nn> (I think I remember linguoboy saying in Munster it was [ŋ]), but for me <nn> only changes the pronunciation of a preceding <a> or <o>. For example, to me <sin> ("this") and <sinn> ("we") are both [ʃɪɲ] (but maybe some dialects differentiate between them) .

Is this [ɲ] a true palatal, or is it alveolo-palatal?

I find it curious that Wikipedia doesn't mention this [ŋʲ]. But then, it doesn't mention [ʐ] either.

[Ciarán12]

I'm just not familiar with what particular abstraction this is. It's not a standard I've come across before.

It's probably because it's intended for foreign learners who might not have access to native speakers, I guess.

Maybe. I don't usually pay all that much attention to phonetic transcriptions in learning material because, from the beginning, I've based my pronunciations on how I had already learned to pronounce it in school.

It's [ʝ] for me.

Maybe it varies between dialects, or maybe they're in complementary distribution. Hell, Spanish has both [ʝ] and [j], why not Irish?

I... am confused. I just realised I do have [j], but I'm trying to work out what the rules for it's distribution are. Giúdach - [gʲu:d̺ˠəx], but lenited it's Ghiúdach - [ʝu:d̺ˠəx], but dhá - [ɣɒ:], but Cad a dhéanann tú? - [kˠɑd̺ˠ ə 'je:n̪ˠən̪ˠ t̪u:]. I don't know how to analyse this.

Well, the way I speak (and also the only way I'm aware of in any dialect) has [ɲ] for /nʲ/. Some may also have [ɲ] for /ŋʲ/ (I have [ŋʲ] for /ŋʲ/). As for differing /n/ phonemes, I think some dialects have a separate /n/ for <nn> (I think I remember linguoboy saying in Munster it was [ŋ]), but for me <nn> only changes the pronunciation of a preceding <a> or <o>. For example, to me <sin> ("this") and <sinn> ("we") are both [ʃɪɲ] (but maybe some dialects differentiate between them) .

Is this [ɲ] a true palatal, or is it alveolo-palatal?

I don't know, I can't tell the difference. It might be alveolo-palatal.