Language Log » Spell this (could Irish take the gold?)

Spell this (could Irish take the gold?)

« previous post | next post »

It is a great pity that Irish was not included among the modern European languages considered in the Seymour/Aro/Erskine study of literacy acquisition times that Mark referred to on Saturday. Jim McCloskey once showed me the spelling of the word meaning "will get". It is spelled bhfaighidh. The word is a monosyllable, pronounced roughly like English we (or wee or Wii, or French oui). One craves to know how Irish would fare on Seymour et al.'s shallow/deep and simple/complex dimensions, and whether it might force English to settle for the silver in the European awful spelling system championships.

[Added later: Anyone skeptical of the value of comments to blog posts — and I have certainly been among the skeptics some days — might want to glance at the astonishingly erudite and generally very sensible and relevant comments below, from a variety of people who (unlike me) know something about the Goidelic Celtic languages. They are enough to restore your faith in the whole comments genre.]



26 Comments

  1. Daniel von Brighoff said,

    August 18, 2008 @ 9:56 am

    One does wonder, but for Irish, it could depend quite a bit on which variety the learners speak. Although there has been a standardised spelling for some time now (Caighdeán Oifigiúil "Official Standard"), a normative pronunciation has yet to emerge.

    I'm not sure which dialect McCloskey's ['wi:] represents, but in my own (L2) West Cork pronunciation, this word would be ['vəigʲ], which I think you'll agree represents a considerably closer match of pronunciation to spelling. Of course, Munster dialects have their mismatches as well, although I'm hard pressed to come up with one as gross. The best I can do at the moment is [dˠi:] for duibhe "blacker" (which is either twice or three times the number of letters as phonemes, depending on how you analyse length). Ah, the joys of supradialectal spelling!

  2. Kae Verens said,

    August 18, 2008 @ 10:27 am

    As Daniel noted, it depends on the dialect.

    I grew up in Dublin (not known to be a centre of authority in Irish, maybe), and would pronounce "bhfaighidh" as "vygh'ee", and "duibhe" as "dwiveh', both of which are reasonably close to the original spelling.

  3. Bridget Samuels said,

    August 18, 2008 @ 10:39 am

    Scottish is in some ways even worse than Irish, because they keep some of the silent digraphs that the Caighdeán reform did away with. For instance, the word for night (roughly, [i:xʲə]) is spelled oíche in post-1950's Irish but oidhche in Scottish.

  4. Daithí Mac Síthigh said,

    August 18, 2008 @ 10:50 am

    I recently moved from Ireland to Norfolk, having spoken Irish for years, and I still can't deal with Wymondham, Costessey, Happisburgh and Acle, so I guess the difficulty is all relative! An interesting twist on this, reflected in Daniel's comment, is standardisation: compare for example the Gaeilge (Irish) "Gaeltacht" [Irish-speaking district] with the Gàidhlig (Scots Gaelic) Gàidhealtachd [either Highlands & Islands or Gaelic-speaking district]. All those extra letters…the Scottish will beat the Irish any day, alas.

  5. Keith Gaughan said,

    August 18, 2008 @ 10:58 am

    Also worth keeping in mind is that while Irish spelling isn't all that transparent, it's at least more systematic than English. In this regard, Irish is somewhat like French: the spelling might not reflect the pronunciation directly, it's possible to systematically get from the spelling to the pronunciation in most dialects.

    FWIW, as somebody from Sligo, I'd pronounce 'bhfaighidh' as [βai.hi], and 'duibhe' as [dɪβə]. I can remember an argument I had with a friend of mine when I was in college about the pronunciation of 'spideog' (en. "robin").

  6. John Cowan said,

    August 18, 2008 @ 11:11 am

    There's a fundamental difference between etabnannimous [ræmˈnænɪməs] orthographies and outright maggelitinous ones, a distinction named by the constructed language community. Etabnannimity is rule-governed, though complex: French and Irish spelling typify this quality. Maggelitinousness is just pure orthographic chaos, synchronically and sometimes even diachronically unpredictable, whimsical even.

    About 15% of English words are spelled maggelitinously, another 35% or so require decidedly etabnannimous rules, and the remainder are spelled simply and straightforwardly. An orthographic reform that removed all (and only) the maggelitinous spellings would be an unadulterated Good Thing.

  7. Breffni said,

    August 18, 2008 @ 11:12 am

    My Irish is poor, but one principle of Irish spelling is that the morphology tends to be represented even if that means the phonology is obscured. So although inflection can change the sound of lexemes in ways that seem quite drastic to people not used to Celtic morphology, the spelling preserves each morpheme.

    In the case of "bhfaighidh", the root morpheme is faigh (which I would pronounce /fai/ – Connacht, maybe also Munster Irish) "get", the -idh suffix is a future-tense marker, and the bh- prefix represents eclipsis – an initial consonant mutation, licensed by a preceding particle like the negator "ní" ("ní bhfaighidh mé", "I won't get") which changes the /f/ into either /w/ or /v/ depending on dialect. If all this was rendered with a shallow orthography, there would be no visible clue to the underlying lexeme. (I would pronounce it /vai/, incidentally, though I might be wrong.)

    English functions kind of similarly in some ways – past-tense "-ed" represents a single morpheme but is used for all three phonological realisations (compare "hated", "chased", "named"). So morphology trumps phonemic transparency there too.

  8. Daniel von Brighoff said,

    August 18, 2008 @ 12:17 pm

    (which I would pronounce /fai/ – Connacht, maybe also Munster Irish)

    Munster Irish has a general rule which transforms final slender gh/dh into slender g, so faigh comes out as ['fˠɑgʲ].

  9. Andrew Carnie said,

    August 18, 2008 @ 12:24 pm

    My favorite is gheobhaidh, roughly [joi:]. Ironically this is really the future of "will get". "Bhfaighidh" is the dependent future found after particles like "an", "go" and "nach".

  10. E W Gilman said,

    August 18, 2008 @ 12:25 pm

    I feel out of place among these people who know what they're talking about, but I thought I'd mention that the Irish write Flann O'Brien under his newspaper name of Myles na Gopaleen once wrote a short sketch in which Irish characters spoke Irish and English ones English spelled as if it were Irish. Years ago a colleague showed me how to read the English parts, but I don't really remember anymore. The skit can be found in the collection "The Best of Myles".

  11. Stephen C. Carlson said,

    August 18, 2008 @ 1:44 pm

    Is there a reason why "etabnannimous" has the pronunciation [ræmˈnænɪməs]?

  12. Mark Liberman said,

    August 18, 2008 @ 2:16 pm

    E.W. Gilman: Flann O'Brien under his newspaper name of Myles na Gopaleen once wrote a short sketch in which Irish characters spoke Irish and English ones English spelled as if it were Irish.

    The book is available from amazon here — well worth your $11.86 — and the sketch in question starts like this:

    Sheán Buidhe: Tabh iú famhnd ánaigh mór seidisius dochúmaints bitheighnd deir Teairli.
    Poiléismeán Bairlí: Bucats obh dem Sur.

  13. Daniel von Brighoff said,

    August 18, 2008 @ 2:21 pm

    That O'Brien piece sounds amusing, I'll have to track it down. In his brilliant parody of turn-of-the-century Irish literature, An béal bocht ("The poor mouth"), he does the opposite, recording an exchange between a native speaker and an Irish learner in which the latter's speech is rendered with the kind of horrible "English phonetics" used in elementary phrase books, e.g.:

    "'Do vool shay an durus amack' – un vwil un aubirch sin ogutt?" arsa an Gaeilgeoir.
    "Nawbocklesh, avic", arsa an Seanduine, agus d'imigh leis ar a bhealach lena cheist ina chloigeann fós gan réiteach uirthi.

    (Standard spelling for these two phrases would be "'Do bhuail sé an doras amach' – an bhfuil an abairt sin agat?" and "Ná bac leis, a mhic!" No more lengthy than the "English" orthography in this case, plus it preserves several important distinctions that the poor learner hasn't yet mastered.)

  14. mollymooly said,

    August 18, 2008 @ 3:47 pm

    Neither party in that extract is speaking Irish

    -Have you found any more seditious documents behind there, Charlie
    -Buckets of them, sir.

    Some words, notably "any" and "sir", have a distinctly Hiberno-English twang

    Manx Gaelic spelling has English influences and looks rather Mylesian (Milesian?)

  15. felix culpa said,

    August 18, 2008 @ 6:38 pm

    From one who is several generations removed from familiarity with Gaelic speech: who is responsible for the orthography; Scots and Irish to confuse the English, or the English to hobble the Scots and Irish?

  16. Amy de Buitléir said,

    August 18, 2008 @ 7:57 pm

    In my part of Donegal, most people would pronounce bhfaighidh as "wee", or perhaps "wee-ee" if enunciating carefully. I do agree with Keith; Irish spelling may look very strange, but it's very regular.

    For those who don't know Irish: Initial sounds of a word change in certain grammatical situations. I believe that in Welsh they just replace the initial letter, but in both Irish and Gaelic we put the new sound in front. As a general rule, if you see an unpronounceable consonant cluster at the beginning of a word, drop all but the last consonant to figure out what the root word is.

    Finally, the spelling "bhfaighidh" may look strange, but it's doing an admirable job of conveying (to those who know the pattern) how the word would typically be pronounced in each of the three main dialects.

    As an aside, Irish used to represent the softening of a consonant by putting a dot over it. With the advent of typewriters we started placing a "h" after it. With computers it would be easy to use those dots again, but I guess they're gone for good. With dots, we would write that word as "ḃfaiġiḋ", which may look a little more manageable. Or it may not.

  17. Daithí Mac Síthigh said,

    August 19, 2008 @ 6:05 am

    Amy – you're absolutely right about Welsh – take for example how Cymru (Wales) becomes Gymru in Croeso i Gymru (Welcome to Wales), while in Irish the word beginning with c takes a leading G as an urú, such as cuimhne (remembrance) but "cur i gcuimhne" (reminds). I think in Scots Gaelic they're not as triggerhappy with the extra consonant, which if memory serves has something to do with the use of words like 'nan' – for example Comhairle nan Eilean Siar (the local authority in the Western Isles) would be, in Irish – Comhairle na nOileáin rather than nan Eilean (though they do sound very similar).

    The dot for the séimhiú (lenition) is very manageable with new technology, and there is a great story to tell of the history of its rise and fall (there's a series of letters between Government departments, for example, fighting over the cost of typewriters). To my delight, I was able to read the 1960s Dolmen Press republication of An Béal Bocht (honourably mentioned above by Daniel), when doing research on the great Gaelic vs Roman type debate. They used a modified Roman type with the dots in all their glory, and it's quite a sight.

  18. Keith Gaughan said,

    August 19, 2008 @ 9:55 am

    Without a doubt, though, the greatest problem with Irish and Scots Gaelic orthography has always been in its representation of palatalisation. It's so pervasive that the typical mechanism used by other languages (such as repurposing 'j') just don't work all that well, which is the reason behind the rather baroque sequences of vowels you get. Gaelic orthography is screaming out for a better way to represent it along the lines of Cyrillic's soft sign and pairs like а/я, &c., but a little less intrusive, such as repurposing the ogonek to indicate a palatalising vowel.

  19. Amy de Buitléir said,

    August 19, 2008 @ 10:14 am

    Daithí, I would like to know more about the history of the introduction of typewriters into Irish. Can you recommend a book?

    Another interesting thing about Irish spelling is that roughly half of the vowels are just there to indicate how the neighbouring consonants are pronounced! Consonants in Irish can be pronounced either broad or slender. (For example, broad 's' is pronounced as in 'same', while slender 's' is pronounced as in 'shine'.) You know which way to pronounce a consonant by whether the vowels on either side are broad (a, o, u) or slender (i, e). So what happens if a consonant has a broad vowel on one side and a slender consonant on the other? That only happens in a few cases. We have a spelling rule which prevents that: "Caol le caol agus leathan le leathan" (Broad with broad and slender with slender.)

    So if we take a word like "tuirseach" (tired), the sound we're trying to represent is:

    (broad t) + i + (slender rs) + a + (aspirated c)

    If we instead used a system to mark consonants as broad or slender, and brought back the dot for aspiration, we could write this as "tiRSaċ" – only six letters. (In this example I've used uppercase letters for slender consonants, and lowercase for broad.) I'm not suggesting we should write Irish this way. I'm just making the point that what looks like a sequence of nine "sounds" is really just six. (I'm not a linguist, so please excuse my sloppy terminology.)

  20. Amy de Buitléir said,

    August 19, 2008 @ 10:18 am

    Ah, Keith, you snuck your post in ahead of mine. Sorry for the duplication. But I guess my post is still useful a sort of "translation" of your post for non-linguists.

  21. Daniel von Brighoff said,

    August 19, 2008 @ 10:49 am

    Just a short technical note to Amy's post: "slender" (caol) corresponds to "palatalised" in phonetic jargon and "broad" (leathan) to "velarised". The corresponding popular terms for languages like Russian would be "soft" (мягкий) and "hard" (твёрдый). A broad IPA transcription of tuirseach would therefore be [ˈt̪ˠirʲʃəx]. ([t̪ˠɪrʲˈʃax] for Munster.).

    Oh, and I thought I'd come up with a lovely example of crazy spelling for Munster when I stumbled upon neamhthaibhseach [ˈnʲafˠɪʃəx] "trivial" (from taibhse "[great] size"). But then a couple of more fluent speakers asked me, "Why don't you just write neafaiseach like the rest of us?" As a bonus, the full pre-reform spelling had dh just before the bh. That's 18 letters for 7 phonemes! (Though, of course, only 14 if using dots above.)

  22. Keith Gaughan said,

    August 19, 2008 @ 2:05 pm

    @Amy, ah well. :-)

    Just to demonstrate what I mean, using Amy's example of 'tuirseach', it'd be spelt 'tirsąch', and Daniel's example of 'neafaiseach', would be 'nąfisąch'. You could reintroduce the punctus delens for lenition, which would give 'tirsąċ' and 'nąfisąċ'. Or for a more dramatic example, we can spell it 'nąṁṫiḃsąċ'. Not pretty, but at least clearer.

  23. Colin Watson said,

    August 20, 2008 @ 3:29 am

    My favourite example was gained from encountering a road sign to the town whose Anglicised form is "Ardee". The Irish form is "Baile Atha Fhirdhia", with a pretty similar pronunciation if you discount the leading [bˠalʲə] 'town' and the slender dh [j]; 12 letters to 5.

  24. David Marjanović said,

    August 22, 2008 @ 8:16 am

    pronounced roughly like English we (or wee or Wii, or French oui).

    Nitpicking alert: French oui has a distinct [u], the beginning of which is just consonantic enough that the word doesn't need (and never gets) a glottal stop in front of it when it's utterance-initial. The [u] is probably at least as long as the [i] for most speakers. And then, of course, some people put [ç] behind every utterance-final [e] and [i].

    From one who is several generations removed from familiarity with Gaelic speech: who is responsible for the orthography; Scots and Irish to confuse the English, or the English to hobble the Scots and Irish?

    A couple of sound changes are responsible*, as well as the abovementioned desire to write not just the pronunciation but also the grammatical makeup of each word (not done in Welsh and Breton).

    * This is similar to Mongolian written in the Mongolian script. The orthography used for this e. g. writes a consonant that has long disappeared, producing previously inexistent long vowels that are not written as such — similar to English gh. In other words, it is written the way Genghis Khan pronounced it, not the way it's pronounced nowadays. Mongolian in the Cyrillic script, incidentally, is beautifully phonemic for the modern language — the one good side of Stalinism.

  25. marie-lucie said,

    August 23, 2008 @ 9:06 pm

    @ David Marjanovic: (sorry, I don't know how to do the final consonant)

    French oui has a distinct [u], the beginning of which is just consonantic enough that the word doesn't need (and never gets) a glottal stop in front of it when it's utterance-initial. The [u] is probably at least as long as the [i] for most speakers.

    Who have you been talking with, or listening to?

    I think that perhaps what you are perceiving as initial [u] is an anticipatory lip-rounding before the [w] in a very deliberate pronunciation. And French speakers normally do not add a glottal stop in front of word-initial vowels unless they are making a deliberate effort at pronouncing a word, in which case they might also produce an [h] (listen to Edith Piaf's recording of her song "Milord", in which the word Allez sounds like [hale] since it occurs just after the singer has taken a deep breath).

  26. John Cowan said,

    April 7, 2009 @ 9:04 pm

    Stephen C. Carlson: Certainly, but I don't remember enough Etabnanni to say exactly why.

    Of course the Goedelic languages should be written in Cyrillic to start with. So much more sensible.

RSS feed for comments on this post