Friday, 24 February 2012

and in French

And so to Adams’s treatment of French. (As before, all transcriptions are his.)

Phoneticians jib at calling a language “phonetic”. What you really mean, we tell our beginner students, is that the writing system is phonemic, faithfully representing the pronunciation of the language. (To us all languages are phonetic in the trivial sense that all languages are spoken and therefore have a phonetics that can be described.)

Adams is unaffected by our preciseness of language.Quite apart from the uncertainties and indeterminacies of its orthography, French is also particularly difficult for speakers of English to pronounce. It sometimes seems as if almost everything that could be different is different: there is no contrastive lexical stress, the voiceless plosives are unaspirated, the vowels — all monophthongs — include front rounded vowels and nasalized vowels, the schwa is (i) rounded and (ii) appears and disappears in complicated ways depending on the surrounding sounds.

For schwa, I think Adams’s wording could have been clearer.
In spoken French, when the letter e with no accent is the only vowel-letter in a syllable and it ends the syllable or word, it is usually silent:
   mouvement [muvmɑ̃]   médecin [medsɛ̃]   ville [vil]
This silent vowel-letter is called mute e. In singing, it is generally pronounced, transcribed as /ə/ (schwa), and sounded as /œ/…
…When a mute e ends a word and the next word begins with a vowel or h, the e is never sounded in speech or singing.
   elle est [ɛlɛ]   comme a [kɔma]
   fatigue amoureuse [fatiɡ amurøːz(ə)]

That might suffice for someone who already knows about the evanescent schwa, but not I think for a beginner who doesn’t know much French. And surely comme a ought to be comme à.

For “nasal” (= nasalized) vowels, Adams rightly warns English speakers not to include an unwanted nasal consonant in words such as onde [õːd(ə)], lamente [lamɑ̃ːt(ə)], impossible [ɛ̃pɔsibl(ə)], embarquer [ɑ̃barke], encore [ɑ̃kɔːr(ə)] (“not [ɑ̃ŋkɔːr(ə)]”).

I am not convinced by Adams’s advice on how to pronounce ɥ. But surely the crucial thing is not the timing of the lip movement, it’s the position of the body of the tongue (front, not back).

Any discussion of French consonants must make a distinction between consonant sounds and consonant-letters, since the letters are often silent.

Hear, hear. Adams devotes over twenty pages to the French consonants, with a comprehensive discussion of the difficult issue of liaison. (I have to confess I wasn’t aware that final c is silent in estomac [ɛstɔma] and tabac [taba]. I did know it isn’t in sec [sɛk] and Poulenc [pulɛ̃ːk].)

For r in French art songs and opera, Adams recommends a tap, ɾ (or rather what he calls “a flip of the tongue”). However, “judicious use of the uvular sound has recently become accepted in some circles for classical singing”.

73 comments:

  1. 'And surely comme a ought to be comma à.'

    commE à? Like: A la guerre comme à la guerre?

    But 'comme a' is not, perhaps, quite impossible, 'comme a-t-il dit/fait etc.' 'j'ai fait comme a-t-il dit', I don't know if this is quite correct or standard, though.

    'comma' in French---existe-t-il? They normally say 'virgule'.

    ReplyDelete
    Replies
    1. Sorry, my typo, now corrected.

      Delete
    2. I don't think Wojciech's sentence works, but you will find plenty of examples of 'comme a dit...' if you do a search.

      Delete
    3. You would say :
      'j'ai fait comme il a dit'

      'comme + a' does not exist in texts such as poems or lyrics.

      The only possibility I see would be in a teacher-student conversation about grammar, where the word 'comme' is discussed for itself.

      "Le mot "comme" a été remplacé par une autre conjonction"

      Delete
    4. Where "comme a" looks wrong, "ell est" just before it definitely is: should be "elle est". It is not Adams's typo — my copy shows this correctly.

      Delete
    5. 'comma' doth seem exist in French, if any trust is to be bestowed upon Larousse. What's the English word for that, the ninth part of a semibreve? http://www.larousse.fr/dictionnaires/francais/comma. But is 'comma a`' [kɔ'ma]?I doubt.

      Delete
    6. A comma is a general name for a small difference in pitch.
      The most common ones (pythagorean and syntonic) are intervals of about 1/9 tone.

      Delete
    7. Je t'en remercie, Vincent. My fault. Pitch indeed.

      Delete
  2. The convenient thing about French is that it matches the IPA (suspiciously) well. My impression is that this perfect match is because the IPA is a French invention – might that really be the reason?

    Oh-oh, a supposedly "phonetic" language! ;-) I can so relate, though I more often have to fend off the talk of a supposedly "phonetic" alphabet (every alphabet is phonetic in the trivial sense that it represents sounds; it is orthographies that can more or less directly represent phonemes).

    ReplyDelete
    Replies
    1. Really? I think that [ɑ̃] is a terrible transcription of the "an"/"en" vowel for contemporary standard European French (although it may have been more accurate for the standard European French of the nineteenth century).

      I hear it as closer to [ʌ̃], accompanied by a vertical compression of the lips.

      Delete
  3. Polish is practically phonemic - you have more than one way to write a word, but you never have any trouble reading it; morever, it's compatible backwards, so apart from any vocabulary change, Poles can read things written 500 years ago without trouble. Of course, the newest research suggests there are sounds the natives don't perceive themselves, but still it's functionally a phonemically consistent one.

    Polish alphabet > French alphabet.

    Afaik Serbs use 100% phonemically consistent writing system: what you read--your pronounce, what you say--you write, so it's even better than Polish, though.

    ReplyDelete
    Replies
    1. Thanks to Vuk Stefanović Karadžić!

      In other news, I see that another book with even more languages has been published. It is by Joan Wall, Diction for Singers: A Concise Reference for English, Italian, Latin, German, French and Spanish Pronunciation.

      Delete
    2. Well, not 100% phonemically consistent, no. For example, in spelling we retain the consonant clusters 'ds', 'dš' and 'đs', but we pronounce them with voice assimilation. So:

      odšetati (walk away) /otˈʃeːtati/
      gradski (city adj.) /ˈgratski/ or /ˈgrat͡ski/
      vođstvo (leadership) /ˈvot͡ɕstvo/

      Also, the pitch accent is not shown in the spelling. So 'grad' (city) and 'grad' (hail) are spelled the same, but are pronounced, respectively, /grâːd/ and /grâd/, or in the non-IPA transcription grȃd and grȁd.

      Delete
    3. As a local linguist once quipped, only illiterate people actually write Serbian as they pronounce it.

      Delete
    4. Polish is less than perfectly phonemic, I am afraid. For one, you have words like 'dania/Dania', 'marznąć', 'tysiąchektarowy', 'suita' etc, for which you just have to know the right pronunciaton. For another, the letter-to-sound rules in Polish are a bit complex.

      Delete
    5. You must be joking. It's true that in Polish there are basic reading rules for each letter and digraph. But these are sometimes quite complex. The single letter "ą" can be pronounced in at least 10 ways: ɔm, ɔn, ɔn̠, ɔmʲ, ɔɲ, ɔĩ̯, ɔũ̯, ɔŋ, ɔŋʲ, ɔu̯. And still, there are oddities like "ziścić" (sʔiɕ.t͡ɕit͡ɕ). Or take the stress rules. Always penultimate? Consider "zasoby" vs "byłoby". You can't just depend on orthography. Morphology is very important as well.

      Delete
    6. @ahven: I think you're overphonetizing. Polish grapheme-to-phoneme rules are perfectly cromulent. Your ą example, for example: all the pronunciations are 95% predictable from context and 4% from style. (Yeah, I've just made the numbers up but I have a hunch they aren't far off.) So, by definition, this particular spelling is phonemic. If you know your morphology, things like ziścić are 100% transparent as well. (BTW, it's /z-/, at least in my pronunciation.) Lexical stress is getting regularized by the day. It's those very isolated cases that Wocjciech gave examples of that can pose a small problem from time to time.

      Delete
    7. As I said above, the rules are complex in the sense that there quite a few of them and they have to be learnt by heart by a NNS of Polish. Ziśćić I would say is [zi-], without a glottal stop, that's how I pronounce it, at least. Stress is largely predictable, apart from some substantives like 'matematyka'. But let's face it: Polish is certainly not amongst the least 'phonetic' languages (to use a locution justly condemned by Mr. Wells).

      Delete
    8. I concur there are exceptions, but it's so far from guessing (vide through/though/tough) that I would still argue that Polish can be INDULGENTLY called phonetically consistent It's not Korean, but still. Behaviorists say that altruism is long-term egocentrism--nevertheless, the word exists; I am employing "phonemically consistent" on a similar "licence" if you will, that is without any absolutes.

      Delete
    9. Ad wjarek

      'Lexical stress is getting regularized by the day.'

      I can't help perceiving the all-penultimate stress you are referring to as vulgar and ugly. Besides, there used to be some functional load, if not very heavy: mate'matyka (mathematics), matema'tyka (of a mathematician).

      Ad Master Polish
      in this sense yes, phonetic. There are marginal problems with the formerly voiced 'h', such as forming the dative case from 'Sapieha' (Sapieże, not *Sapiesze, as folks are saying these days) or using the right prefix ('zhabilitować się', to obtain one's 'habilitacja' degree, rather than 'shabilitować się', as people write now, cf. 'zhańbić się', 'to gain ill renown'). In my idiolect, btw, 'h' is largely voiced still. But such errors rather prove your point (than disprove it).

      Delete
    10. Vulgar, hmmm. That's a strong feeling ;) Doesn't it make perfect sense to regularize if a vast majority of word forms are regular? Language is a living thing...

      Incidentally, this was the topic of my first ever linguistic study, in my 3rd year at univ. FWIW (the sample was small), age was by far the most important variable.

      BTW, I was thinking more about past tense verbs, not -yka/-ika borrowings. When I ask my students which Polish words aren't stressed on the penultimate syllable, most can only think of the borrowings. But both groups vary, with the penultimate clearly predominant.

      Some people (e.g. Lech Wałęsa in super-rhythmic preacher mode) will even have penultimate stress in -byśmy etc.

      Delete
    11. I would still argue that Polish can be INDULGENTLY called phonetically consistent It's not Korean, but still.

      To view Korean orthography as the ultimate in phonetical consistency is misleading to say the least, and from the examples so far the situation with Polish seems not that different from Korean. The Korean alphabet rightly gets a lot of credit for its phonetically informed ingenuity, but you have to keep in mind that it was created for 15th-century Middle Korean. Writing is inherently more conservative than speech, and though orthography reforms have managed to keep it reasonably consistent, there are plenty of ambiguities.

      I find Korean quite similar to French orthography in a sense. Because of Korean phonotactics, there are silent letters that surface in certain environments, e.g. 삯 |sags| alone is /sag/ but 삯이 |sags-i| is /sag.ssi/ (Here I'm using a phonemic transcription that is as close as possible to the spelling; the /g/ is /sag/ is actually an unvoiced and unreleased [k], and the /ss/ in /sag.ssi/ would actually be a fortis [ɕ] due to the following vowel.) The |h| is silent in 좋아 |dzoh-a| /dzo.a/ but the combination |h-d| produces /tʰ/ in 좋다 |dzoh-da| /dzo.tʰa/ (again, the /dz/ here is actually lenis unvoiced and is usually thought to be postalveolar). As you can see from the latter example, there are complex assimilation rules, so e.g. 삯만 |sags-man| is pronounced /saŋ.man/.

      The fun comes in when we consider what is not predictable from spelling. 별도 |byʌl-do| 'star also' is straightforward /byʌl.do/, but 별도 |byʌl.do| 'separate' is /byʌl.tto/ because final /l/ in Sino-Korean words trigger the fortis realization of a following coronal obstruent. 단가 |dan.ga| 'individual price' is pronounced /dan.kka/, but 단가 |dan.ga| 'short song' is pronounced /dan.ga/. These are both Sino-Korean words, but the Sino-Korean element |ga| in the first word seems to have a fortis realization in compounds.

      Unwritten 'medial sounds' (사잇소리; insertion of /n/ between elements of compounds or fortis realization of the initial obstruent of the second element) cause the greatest headaches for the learner. 솜이불 |som*i.bul| is /som.ni.bul/ and 안방 |an*baŋ| is /an.ppaŋ/, where I indicate by |*| the 'medial sounds' that are not indicated in the spelling. Sometimes 'medial sounds' are indicated in the spelling by 'ㅅ' |s| but this is not always possible. Native speakers themselves often don't agree on whether a word should be pronounced with a 'medial sound'. 검열 |gʌm.jʌl| can be either /gʌm.njʌl/ or /gʌm.jʌl/; both pronunciations are accepted as standard. 김밥 |gim.bab| is normatively /gim.bab/ but is very frequently pronounced /gim.ppab/. The metro station 학여울 |hag.jʌ.ul| is normatively /haŋ.njʌ.ul/ and hence uses the romanized spelling 'Hangnyeoul' but is often pronounced /ha.gjʌ.ul/ without the 'medial sound'.

      Speaking of metro station names, 선릉 |sʌn.lɯŋ| is normatively /sʌl.lɯŋ/ with regressive assimilation and hence uses the romanized spelling 'Seolleung', but most people seem to say /sʌn.nɯŋ/ with progressive assimilation. But there are also words like 상견례 |saŋ.gjʌn.lje| where the pronunciation with progressive assimilation /saŋ.gjʌn.nje/ is standard and universal.

      The short version is that Korean orthography is full of complex rules and even then there are plenty of ambiguities where it is not possible to predict the pronunciation from the spelling alone. This is not even considering phonemic vowel length distinction, which is not indicated by spelling but has completely disappeared in younger speakers anyway (and which I have ignored completely in the transcriptions above).

      Delete
    12. 별도 should be |bjʌl-do| instead of |byʌl-do|. I started writing the above comment using a more English-like romanization before I changed my mind and made it more IPA-like.

      Delete
    13. ad wjarek

      de sentimentibus non est disputandum. There is no moral obligation, a fortiori, to have nice or neutral feelings about certain linguistic changes. I dislike the changes, not the persons who have them. The amphibrachic rhythm in three-syllable words, I also find it boring. Poles often use it in Italian, too, erroneously, e.g. 'il sinDAco di BerGAmo, where it's il SINdaco di BERGamo. Dactylic.

      About verbal forms you are right, of course. The penultimate stress is somehow coupled with preachery tone of voice or excessive didacticism, which I also dislike.

      If they are clearly winning out I dunno yet, I hope it'll take some time, at least as long as I am alive.

      Delete
    14. I once had the misfortune to have to sit in a waiting room for two hours with nothing to read and only two other people in the room, who were having an animated conversation in Polish (of which I understand not a syllable). I nearly died of linguistic boredom: endless unreduced vowels and penultimate stress!

      By contrast, I can listen to lengthy Russian conversations, where I also understand nothing, with ease.

      Delete
  4. Also in other news, the other day I saw mention of these iOS apps aimed at singers.

    ReplyDelete
  5. To be honest, I've recently been finding French more "phonetic" [sic] than Italian. With a basic understanding of the spelling and liaison rules, I can manage the vast majority of French words on sight; but as a beginner trying to read Italian texts, I'm constantly paralyzed by the question of whether a given 'e' or 'o' ought to be open or closed, or a given 'z' voiced or voiceless.

    ReplyDelete
    Replies
    1. for both, there are historical rules, which presuppose a knowledge of Latin. The 'e' from Latin short 'i', pesce, fish, for instance, is closed. The 'z' from Latin 'di', mezzo, is voiced. This is what they say and teach, at least.

      Delete
    2. /e/ ~ /E/ and /o/ ~ /O/ are not terribly robust oppositions anyway: many accents in the north and the extreme south don't distinguish them phonemically, and even in the centre their lexical incidence in many accents is not the same as in the standard; and even in the standard there aren't that many minimal pairs (this being an allegedly extensive list). So using the wrong one might sound weird but it seldom causes misunderstanding, especially away from central Italy. And I know of exactly *one* minimal pair between /dz/ and /ts/. (At least you didn't mention voiced vs. voiceless /s/; a phonemic distinction between them is as dead as horse/hoarse is in English.)

      Delete
    3. A _romana_ living in Milan once told me that on hearing 'pesca' in her child's school with a narrow 'e' she expected a fishing event, while what they meant was a big peach. Strange that such things should be mistakable for one another in a context of discourse, but in today's world we must not be astounded.

      Delete
  6. (To us all languages are phonetic in the trivial sense that all languages are spoken and therefore have a phonetics that can be described.)

    Except for sign languages. (That is, the are languages that aren't spoken.)

    ReplyDelete
  7. The constructed-language community divides difficult spelling systems into the etabnannimous and the maggelitous (or maggelitinous) types. Etabnannimous systems have complex but predictable spelling-to-sound rules, as for example the name of the Etabnanni language, quite regularly pronounced [ræmˈnanni]. Maggelitous systems, per contra, are beyond prediction. I confess I do not even remember how Maggel is pronounced, but it's nothing like [maggel].

    French orthography is almost entirely etabnannimous; one of the exceptions is oignon, traditionally pronounced ognon, though a spelling pronunciation is beginning to catch on. English words are, if I remember correctly, about 50% straightforward, 35% etabnannimous, and 15% maggelitous.

    ReplyDelete
    Replies
    1. If a spelling pronunciation is beginning to catch on, it would be among NNSs, and in spite of the Académie recognizing the spelling "ognon".

      Delete
    2. French orthography is almost entirely etabnannimous

      ennui but ennemi
      abdomen but examen
      patient but tient
      agenda but hacienda

      Would a native French speaker know from the spelling how the name of the conglomerate Vivendi is pronounced?

      Delete
  8. The Adams text is excellent. Sure, linguists and phoneticians are bound to quibble with some of its terminology, but the information an English-speaking singer needs is thoroughly presented. I have used it for Italian and German for three years now in a university course I teach on an overview of lyric diction.

    When it comes to French, I mainly agree with Adams's descriptions, but differ enough from his approach to have made my own handouts. Last year, the university printed them in a perfect-bound softcover book. vp kindly mentioned it in a comment on this blog on February 2; here it is, for anyone else who is interested.

    ReplyDelete
  9. >I have to confess I wasn’t aware that final c is silent in estomac [ɛstɔma] and tabac [taba].<

    That's very big of you. But then I believe you've never done tobacco, or even a lot of eating. I've always been a martyr to my 'nerfs', and cannot imagine ever having pronounced the f, but that did not save me from having to admit embarrassingly late in life that I wasn’t aware that final f is silent in 'cerf'.

    ReplyDelete
    Replies
    1. 'to admit embarrassingly late in life that I wasn’t aware that final f is silent in 'cerf'.'

      because you never did much hunting? Fortunate beasts!

      It would interest me what, if anything, Adams says on 'aou^t'. Oo or oot?

      Delete
    2. Or ah-oo or ah-oot?Thanks to you I have been doing a horrible amount of hunting, but this is a very elusive beast. I was appalled by oot, let alone and ah-oot, when I first heard them, and had never thought to investigate such outrages on the internet. Not having a copy of Adams, I thought a squinny at the present state of play might enable me to contribute a succinct statement of their inappropriateness in any form of phonation, but I can only report that there seems no possibility of creating order out of the chaos that still persists. You may be interested in this: http://atilf.atilf.fr/tlf.htm.

      Delete
    3. This dictionary says 'oo' or 'oot', without giving any guide-line. For aught I can see, at least. I am used to thinking that 'oot' be a hyper-correct form. But if so, then a-oo(t) would be super-hyper-correct. Have you actually ever heard them from anyone but a French beginner? (in fact the word may look as though it were to be pronounced that way).

      All of this 'wild hunt' helps you little with French harts and their mute letters, however.

      Delete
    4. I too used to think that 'oot' was a hyper-correct form, and a-oo(t) super-hyper-correct, but if you read the article in TLFi I linked to you will see it's always been a lot more complicated than that. And yes, I have heard some of them from NSs for close on 60 years, especially, and it seems to me increasingly, 'oot'.

      Delete
    5. Fascinating: I had no idea that any pronunciation other than /ut/ was possible.

      PS I find http://www.cnrtl.fr/definition/aout easier to search.

      Delete
    6. ad mallamb
      NSs for close on 60 years, especially, and it seems to me increasingly, 'oot'.

      Yes, to me too(t). Spelling-pronunciation taking over?

      Delete
    7. Wouldn't the t-less realization be the spelling pronunciation? (to match, say, ragoût).

      Delete
    8. most final -t in French are mute... even in 'but' /byt/ nowadays, the 't' seems to have been restored under the influence of spelling, if we are to trust Monsieur Littre':

      http://francois.gannaz.free.fr/Littre/xmlittre.php?requete=but

      Delete
    9. I have never heard /aut/; and /a-u/ only in the song ("A la mi-Août").

      In de La Fontaine's La Cigale et la fourmi the pronunciation is clearly /u/:

      "Je vous paierai, lui dit-elle,
      Avant l'Oût, foi d'animal,
      Intérêt et principal. "

      Delete
    10. how can you tell that that be without a -t? For euphonic reasons? Anyway I feel reinforced in my anti-t prejudice (as far as aou^t goes).

      Delete
  10. interesting, most French singers I know tend to go with a trill for /r/ in more classical/traditional singing.

    ReplyDelete
  11. I am impressed by the poetics of these 'prove you are not a robot' passwords we are requested to enter here. The last one in my case was 'ionswar ropyrig', which has a vaguely Nordic, maybe,too, vaguely Celtic, ring to't. How ever are they generated? Wise humans or robots who take it out on humans. In the latter case there must be some software and I'd be interested in what it is like.

    ReplyDelete
  12. Thanks, wjarek. Fascinating. And the new CAPTCHA software on here seems to be keeping the spam out. I wonder how long for.

    Perhaps this time I'll remember what the acronym stands for, but it's too clever by half, and I always forget. With a brain disintegrating like this and dodgy eyes, this version of it will soon be making bona fide posting impossible for the postcocious. Don't either of you think it's getting damnably difficult?

    ReplyDelete
    Replies
    1. It's not the CAPTCHA, it's the fact that you need to have an account to post. Either Google, OpenID, WordJournal...

      Delete
    2. We're also getting better at keeping the spam out because I have realized that I need to check every couple of days or so, and have discovered how to do that efficiently using the tools available in Blogger. (Also how to reinstate genuine postings that Blogger thought was spam.)

      I've cleared all the spam from the last 12 months' postings, but there is still some old spam left there in the earlier years.

      Delete
  13. Don't either of you think it's getting damnably difficult?


    I do. The world belongs to the young. Yet, 'up with the postcocious', e. g. us---sez I.

    ReplyDelete
  14. I think I've seen captchas that are much more difficult. The ones on here are usually doable on first attempt. And you can always try the audio version...

    And wrt to their poeticness (poeticity?), I actually have the impression that they used to be more poetic. Or maybe at least more perversely pertinent to the topic. I've even commented on this at some point.

    ReplyDelete
    Replies
    1. Well, poetry is not the right word. I mean, these captions are always or nearly always suggestive of some language, sometimes Romance, sometimes something English-like, sometimes something else---like Nordic. But never Bantu or Tibeto-Burmese, or Burushaski, for that matter. Does the Wikipedia entry explain how they generate them? Maybe they chop and shuffle real words from an assortment of languages. In which case they'd be well-advised to extend that assortment.

      Delete
    2. I think this depends on the specific capcha implementation. I once had an installation of a content management system where the captcha was strictly CVCV.

      Delete
  15. I've now tried the audio version and I think it's more difficult. Numbers embedded in babble and noise. In English, of course ;)

    ReplyDelete
  16. Some questions need to be answered. For example, differential phonology. How do the ʃ, tʃ, dʒ differ from language to language, Spanish, Italian, German, English, French, in terms of place of articulation, manner etc. Apicality, laminality, absence or presence of palatalization?

    Anyone?

    We also need to solve the mystery of the German and English o, since if you compare the diagrams in Cambridge dictionary and Duden, they are exactly the same.

    English should've kept o for ɔ and ɒ untouched.

    Wojciechs needs to tell us the pronunciation of those unintuitive words from Polish quoted by ahven. And the word for mathematics.

    ReplyDelete
  17. tysiąchektarowy: /tɨɕɔntsxɛktäˈrɔvɨ/ 1000 hectars large, 'ch' is normally pronounced /x/. t is dental, /ts/ stands here for one sound, like 'z' in German, I can't get in the linking 'moustache'.

    marznąć: /märz'nɔntɕ/, freeze, 'rz' is normally /ʒ/.

    dania: däɲä dishes, courses.

    Dania, däɲjä Denmark. Before 1936 they spelt the country 'Danja', which was truer to pronunciation. d, like t, is dental in Polish.

    'suita' (suite, in music) is /'switä/. But the word looks as if it could just as well be pronounced /'sujtä/ and I am aware of at least one case of such mispronunciation.

    matematyka, mätɛmätɨkä, t is dental. With stress on the antepenultimate, this means 'mathematics', penultimate -- 'a mathematician's'.

    No 100 p.c. guarantee of accuracy of the above, be warned. I am an 'ignorant amateur', to use the _epitheton ornans_ coined by a justly irate Mr. Wells. But it does seem to me that I've got the thing right.

    ReplyDelete
  18. sorry, it's /'märznɔntɕ/, freeze, marznąć.

    ReplyDelete
  19. @Wojciech: If you want to be very accurate and insist on /ä/, then you would probably want /ɲ/ for the second nasal in marznąć.

    For suita, I have what you transcribe, but I have very definitely heard people saying /suˈita/.

    ReplyDelete
    Replies
    1. ɲ for /'märznɔntɕ/,thus /'märznɔɲtɕ/?

      Well, I myself seem to have 'n' rather than 'ɲ' here. The tongue is in the 'dental' position, as if in anticipation of the following non-palatal /t/ of /tɕ/. I can't discern any difference between that and /lɔnt/ for 'ląd', firm land. I would not be tempted to spell this word as 'marznońć'.

      /su'ita/, 3 syllables, yes, this too. The word is tricky.

      Delete
    2. have 'n' rather than 'ɲ' here

      Really? Do you really feel the same nasal in marznąć and marznący? That's sort of against the rules of Polish phonology ;)

      The dental contact doesn't matter. But what is the front of the tongue doing?

      Delete
    3. Well, I ain't sure no more, it thinketh me so only. I only know that I have to put some conscious effort into saying 'zmarznońć' (with a ɲ). Maybe you're right, I'd like to see some X-rays... . The front of the tongue doesn't seem to rise much. Sometimes I have the impression the vowel itself is nasal, rather than decomposed into an o and an n or an ɲ. Curious.

      Delete
    4. On reflection and more experimentation I must finally admit that I have a different nasal consonant in 'marznąć' and 'marznący'. Vicisti! as Z.K. would have said.

      Delete
    5. I was really just asking a question.

      I wouldn't be surprised if someone found that there is in fact "another palatal nasal" in Polish in this kind of context (i.e. different from your regular ń in e.g. pranie).

      For example, the accepted orthodoxy about ą before fricatives is that the second element is a nasalised [w] or [j]. But there was an old paper by Jassem in which he argued for more (and different) options on the basis of the position of the tip of the tongue. Can't remember what they were exactly. So maybe here you have something that is not 100% the same as a "regular ń"... ?

      Delete
    6. yes, but your question sent me thinking...

      and yes, I suspect that what I have in 'marznąć' is different both from 'n' and 'ń'. Such things are largely inscrutable, to me at least.

      Delete
  20. The Finnish writing system is virtually phonemic. I say "virtually" because even though your average Finn will say that Finnish is completely "phonetic", there are some instances where the spelling doesn't exactly follow the pronunciation. An example of this would be /k/ initial phonemes following /e/ as in hernekeitto (pea soup), where a glottal stop precedes the /k/ in keitto (soup). This results in a pronunciation of "hernekkeitto". This phenomenon is in fact well documented in more in-depth Finnish grammar books as "loppukahdennus" (lit. end doubling). Interestingly, some regional dialects of Finnish lack this feature (western coastal dialects). Elsewhere in Finnish, double consonants are crucial to understanding, as the glottalised phoneme boundary dictates the word meaning, for example, kuka (who), kukka (flower).

    ReplyDelete
    Replies
    1. Finnish is in fact quite famous for its 'phoneticity' and from you I have learnt why it is not perfect in this respect. Nothing is perfect in this world... Do all composita ..VC... have this type of doubling?

      Delete
    2. Finnish often isn't very phonemic: "minä olen" is pronounced moːn.

      Finnish would be phonemic if it were pronounced as it is written. But then, any language would be phonemic if it were pronounced as it is written.

      Delete
    3. But is 'mo:n' for 'minae olen' standard literary Finnish? I doubt. For aught I know, it is a 'parallel language', spoken Finnish, like, say, Swiss German (spoken) as different from Standard German (written). Is that wrong?

      Delete
    4. It is a similar situation, yes.

      Delete