Gaelic phonology, especially liquids

[conmaol]

There seems to be a consensus that Gaelic has two distinct (non-lenited) L phonemes - velarised /??/ and palatalised /?/. As I understand it, they are used contrastively in initial position before a back vowel, and the way to represent an /?/ before a back vowel in the orthography involves inserting a silent <e> or <i>, e.g. "lionn", "leac".

However, they don't appear to be used contrastively before a front vowel. At least I can't find any examples. Is this a general feature of velarised and palatalised consonants in Gaelic? That they only contrast (word-initially) before back vowels? Are there any true minimal pairs involving the two L phonemes (words that are pronounced exactly the same apart from the articulation of the L sound)?

Also, is there any particular reason why lenited L is not spelled <lh> in Gaelic? I know that use of <h> to denote lenition has been extended gradually over the centuries - first <ch>, <ph> and <th>, then <sh>, <mh>, <bh>, <dh> etc. Has the spelling system just not caught up with the Ls, Rs and Ns yet? Or is it because these sounds are not felt to be contrastive enough to warrant an <h>?

[Gordon Wells]

Interesting questions. I've got no answer to the first, but no doubt Akerbeltz will be a great source to tap in that respect.

As for the second (lenited L), yes, it is a logical gap that h is not used after l, (or r or n), but since when has any orthographic system been totally "logical"?

Where I am (Uist) there is, still, a strong contrast in the Gaelic of fluent speakers. Ask someone to say "Nochd iad a-nochd" (they appeared tonight) and you'll hear it very clearly.

But if the spelling system is ever going to "catch up" it had better get its skates on. Wholesale Anglicisation of the Gaelic sound system proceeds apace. I have known Celtic Studies graduates with first class degrees who were not aware that liquid lenition actually occurs, let alone reproducing it themselves... And I make no claim to perfection myself, though I do try.

Actually, no, I'm not advocating changing the Gaelic spelling system - again. A "convention" is just that - a "convention".

[GunChleoc]

You read very often that l, n, r don't lenite because the orthography ignores it, but they do lenite. Originall there were four sounds each, but one got dropped and not in a systematic fashion.

http://akerbeltz.org/index.php?title=Li ... _in_Gaelic

http://akerbeltz.org/index.php?title=Th ... _L_N_and_R

[akerbeltz]

Ah I love questions like these. Are you sitting comfortably?

The easy one first. There was a brief period in ... err some centuries back where some writers tried to experiment with lh rh nh (can't remember names) but it didn't catch on. For a simple reason. Gaelic (and Old and Modern Irish) are very specific about disallowing a whole raft of initials in unlenited positions: Fricatives (except s) and weak lnr. Therefore, unless lenition, historical or synchronric, kicks in, all initial lnr are strong (i.e. L L' N N' R i.e. l?? ? n?? ? r?).

With historical I mean those weak initial lnr sounds that are the result of a lost leniting particle (e.g. cha do nigh mi ~ nigh mi where the declarative past has dropped the past tense particle do; rium, riut... (historical frium...)). Synchronic means that the leniting particle is still there e.g. do nead, do leabaidh...

So it's a bit like trying to suggest that vowel intial words in English should be spelled with an ?okina... they ALL have a glottal stop anyway, so no point in writing it. Just a waste of parchment

Remember also that the appearance of the punctum delens postdates lenition by centuries, so the Gaelics were never that bothered about writing stuff that was "known" anyway, especially when lenition was just a boring old phonotactic rule.

The second answer is slightly more... complex. You've got the spelling mixed up with phonology. In terms of lnr (in fact all sounds but the labials) the palatal (or non-palatal quality) is NOT down to the subsequent vowel. It just IS (from a synchronic perspective anyway). So after L N R AND L' N' you can have any vowel permitted for an initial syllable: linn /L'i?N'/ vs leann /L'auN/ etc.

Mininmal pairs are rare though, the only one that springs to mind is lann /LauN/ vs leann /L?auN/. The problem with minimal pairs is that only l and n come in pairs and that the vowel phonology of Gaelic isn't that amenable for minimal pairs. Either way, /La L?a L? L?? Na N?a N? N??/ etc are permissible as initials, even if they don't form minimal pairs.

Or am I misunderstanding your question a bit?

[conmaol]

Thanks for the feedback guys!

So after L N R AND L' N' you can have any vowel permitted for an initial syllable: linn /L'i?N'/ vs leann /L'auN/ etc.

But you can't have words beginning /Le/ or /Li/ right? In other words, there is no possible Gaelic word that begins with a velarised liquid followed immediately by a front vowel. In terms of orthography this would mean finding a word which starts with something like <loe> or <loi> where the <o> is just a silent marker of "broadness" (just as in <leòr> the <e> is a silent marker of slenderness). If this is true, it means that the contrast between velarised/broad L and palatalised/slender L' is not fully contrastive in initial position. If you know that the word starts with a lateral liquid and you know that the next sound is a FRONT vowel, then you can deduce that the liquid is PALATALISED/slender L.

I think this might be the same for all initial consonants - the only words beginning with (phonetic) [ti] or [si] are loanwords like "ticead".

[Gordon Wells]

If you know that the word starts with a lateral liquid and you know that the next sound is a back vowel, then you can deduce that the liquid is velarised/slender L.

I think you mean velarised/broad L?

[conmaol]

Sorry I got my terminology all mixed up. I will re-edit the original post to clarify.

[akerbeltz]

Interesting question. I think it's a matter of perspective. It's clearer with the long vowels which are clearly "vowels" but with short vowels, it's hard to tell if a word like luim is /Luim/ or /Lim/. There's a lot of transitional stuff (not surprisingly really) going on between these very strongly velarised /L N R/ sounds and a following front vowel, so if you're very picky about your IPA, you would probably end up with /l??^uim/ or /l??uⁱm/.

Indeed, you you check the Survey of Scottish Gaelic Dialects, which uses very close transcription, you'll see a plethora of transitional sounds ranging from ^? to ^? for luime.

It's just a matter of convention I guess that on this side of the Irish sea most of these are seen as true dipthongs, rather than a simple vowel with transitional stuff. On the Irish side, I believe the convention is the other way round so words like rith are transcribed as /Rih/.

I must confess I've never considered what this phenomen actually represents from a purely phonological POV, coming at it from the teaching POV mostly.

Flicking through the Survey, transcriptions with dipthongs seem to outnumber those with transitional stuff by far. Sequences of /R/ + front vowel were certainly extant at one point (cf rinn, rith > ruith etc) and given items like lae /Le:/ in Irish, I'd say we're probably looking at a scenario where LNR + front vowel were possible at one stage but in ScG introduced a back vowel or trace vowel to allow for the transition to the front vowel.

[Seonaidh]

When I velarise or palatalise a liquid it is usually coffee...

Dè an diofar eadar:-
1: "a làr" (an làr aice) agus "a làr" (an làr aige)
2: "a nòs" (an nòs aice) agus "a nòs" (an nòs aige)
3: "a rùn" (an rùn aice) agus "a rùn" (an rùn aige)

[conmaol]

From a word list of over 7000 distinct Gaelic word forms I got the following results:

Code: Select all

$ grep '[ei][àòù]' morewords.txt | wc -l
     134
$ grep '[àòù][ei]' morewords.txt | wc -l
     394
$ grep '[aou][èì]' morewords.txt | wc -l
       1
$ grep '[èì][aou]' morewords.txt | wc -l
     183

So there was only one word containing a short back vowel followed by a long front vowel ("doìgh"). Which suggests that the phonemic distinction between broad and slender consonants *is* neutralised before a front vowel. But not before a back vowel or postvocalically in general. I had a hunch about this from the fact that my (very untutored ear) hears lots of /Cj/ onsets in Gaelic but no /Cw/ ones.

[Níall Beag]

I had a hunch about this from the fact that my (very untutored ear) hears lots of /Cj/ onsets in Gaelic but no /Cw/ ones.

Ah, now it makes sense.

The original (theorised) distinction between P-Celtic and Q-Celtic (including Gaelic) was the /Cw/ onset in Q-Celtic, and Gaelic is known to have lost that w-glide. So Welsh "pen" (head) is cognate with the Gaelic "ceann", which in some early stage is likely to have been something along the lines of "kwen".

Yeah, they lost that initial phoneme a long time ago....

[AlasdairBochd]

On the BBC Alba site there is a very good section with old recordings of school students in Uig and Ness, Lewis .
http://www.bbc.co.uk/scotland/alba/radi ... og=foghlam
The different Ls are very clear in their speech and Rs as well, although I can't differentiate them so well. Some good examples of nasaliation as well (if that's the right term). Listen particularly to Sandra Corbett ag aithris and both Fiona Rennie agus Sandra Corbett ag aithris

[akerbeltz]

So there was only one word containing a short back vowel followed by a long front vowel ("doìgh")

And that's a mis-spelling, it's dòigh, which is /d??j/ anyway, with the /j/ coming from the -gh. a/o/u + è/ì are impossible sequences.

Which suggests that the phonemic distinction between broad and slender consonants *is* neutralised before a front vowel.

As I said, it depends to some extent on how you split up the vowel phenomena. I'd also be a bit careful about which dialect you're looking at. There are some where ea still comes out as /a/ so leag /L?ag/ ~ lag /Lag/ (though I don't have a ref) is a distinct possibility. But you could probably argue it's greatly diminished. Which makes sense, especially if the bear the bpmfv back/front phenomenon in mind.

There's also lenition to bear in mind, so even if radical /L N/ and /L' N'/ are short on contrasts, there are loads where you get /N/ vs /n/, /L'/ vs /l/, /N'/ vs /n/ minimal pairs. nigh! /N?i?/ ~ nigh /ni?/, leum! /L?e?m/ ~ leum /le?m/ etc.

lots of /Cj/ onsets in Gaelic but no /Cw/ ones.

Well, Cj you only get with /bj pj mj fj vj/ before a back vowel. Most other consonants aren't a consonant plus a glide but palatal(ised) consonants. Some dialects have C? if I recall rightly but in neither case does that hail back to the P/Q thing but it's a recent development.

It's not that surprising though, that before a front vowel the palatal/non-palatal thing is much less prominent.