Monday, 13 July 2009

uvular r : trill or fricative?

The uvular trill, [ʀ], is very rare among the world’s languages. However the languages that do have it include French, German, and Dutch — though in each case there are other speakers of the language, perhaps the majority, who use a uvular fricative (or something else) instead.
One thing I noticed in Paul Joyce’s German pronunciation pages was that the uvular trill is what he recommends for “consonantal r”.
The German consonantal 'r' is described as a 'roll' or 'trill', by which we mean that the speech organs strike each other several times in quick succession in the articulation of this sound. In northern and central Germany, this sound is made towards the back of the vocal tract, with the back of the tongue raised towards the uvula in order to create a narrow passage. When the airstream moves through this passage, the friction thus created causes the tongue to touch the uvula either once (uvular flap) or several times (uvular roll). The 'r' sound thus created has a rasping throat-clearing quality which can be equated to a less extreme version of the sound produced when gargling.

I think I’d have said that the uvula (the active articulator) touches the back of the tongue (the passive articulator) rather than the other way round: but no matter. The interesting question is whether learners of German should be encouraged to use a uvular trill [ʀ], as Joyce suggests, or rather a uvular fricative [ʁ]. Joyce does add
Although originally used in informal contexts, this [fricative] variant of consonantal 'r' is slowly emerging as the most common pronunciation of the sound in Germany.

My own impression, for what it’s worth, is that a uvular fricative (or approximant) is by now much more widely used than a uvular trill. This agrees with what Mangold says in the current (6th) edition of the Duden Aussprachewörterbuch, where he writes
Bei den ausgebildeten Berufssprechern des Rundfunks und Fernsehens sowie den Berufsschauspielern auf der Bühne und im Film überwiegt deutlich das Reibe-R [ʁ] …
[With educated professional radio and TV announcers, as with professional actors on the stage and in film, the fricative r clearly predominates.]

Furthermore, I think (am I right?) that most learners would find a uvular fricative easier to learn than a trill.
[Picture credit: Sarah Churng.]

33 comments:

  1. Your impression concurs with mine in all points, for what it's worth.

    The trill is still sometimes considered the standard, but it's dying out.

    By the way, the Northern/Southern division one can often read is much too simplified. Stereotypical coastal German has a distinctive alveolar trill, and the Swabian dialect in the South has rhotacised vowels. It's more about marginal vs central, if at all.

    ReplyDelete
  2. Fortunately, I didn't have difficulty learning a uvula trill and can produce it easily in most phonetic environments. But the majority of the Japanese learners of French and German that I have spoke to find a uvula trill more difficult to learn than a uvula fricative.

    I wonder if this has something to do with the length of the uvula. The longer your uvula is, the easier you find it to learn a trill??

    ReplyDelete
  3. I agree with John that learning a uvular fricative is much easier than producing a trill at the same place of articulation. With a trill there has to be a delicate balance between the degree of stricture and the force of the pulmonic airstream to make the velum vibrate. In the case of a fricative all you need is a narrow passage to achieve a friction.

    ReplyDelete
  4. I've never had a problem with the trill, but I know a lot of people who do. My wife, who has several German friends, regularly asks me to produce it, but has yet to imitate it properly (though she has no problem with lateral fricatives, uvular ejectives, or ejective alveolar trills). It could be that, as Shige says, it has something to do with the physiology of one's vocal tract.

    ReplyDelete
  5. If your velum is as stiff as a board for whatever reason, then you will not be able to trill. I myself have problems at times, particularly when my throat is very dry. A little mucus will help!

    ReplyDelete
  6. Can anyone, please, explain the picture by Sarah Churng to me? My mind boggles.

    ReplyDelete
  7. Does the picture mean to show the articulation of a "nasalised" uvula trill?

    ReplyDelete
  8. "Does the picture mean to show the articulation of a "nasalised" uvula trill?"

    I don't suppose it means to show any particular articulation, as opposed to just reperesenting the head of a linguist, but surely if it did it would logically by some sound in the utterance "Want to hear my uvular trill?"

    ReplyDelete
  9. Uvular r in Dutch? (Maybe some posh centenarians...) The sound _is_ changing nowadays, it seems, from a trill to an approximant, but its position is persistently alveolar in common usage, I think

    ReplyDelete
  10. In Dutch? The South, cities like Rotterdam, then parts of Flemish and of Afrikaans? And some posh centenarians.

    ReplyDelete
  11. Yes, I think the uvular fricative is the most common form of r in German. I think that the vast majority of English-speakers would find it easier to learn.

    I'm still not clear on the consonantal/vocalic R point of German. For example, the word "Lehrer" is usually /le:ʁɐ/. That is a vocalic R. What if you make the word feminine in "Lehrerin"? Does it become a consonantal R? When in Germany recently, I said /le:ʁəʁin/ but I don't think that was right. I'm really not sure how you're supposed to say "Lehrerin" in German. (It's important as my mother is a German teacher)

    ReplyDelete
  12. [le:ʁɐʁɪn], at least in real life.

    ReplyDelete
  13. Thanks, Lipman. Both Rs would be consonantal sounds in "Lehrerin".

    ReplyDelete
  14. The question is about the vowel between the ʁs, but in this terminology, the second /r/ in [le:ʁɐʁɪn] would be both consonantal and vocalic. Only "consonantal" would be [le:ʁəʁɪn], and only "vocalic" [le:ʁɐʔɪn], I suppose.

    ReplyDelete
  15. @Kraut:
    A "pick up line" is something you say to a stranger with whom you are hoping to forge a relationship. One traditional pickup line is "Do you come here often?".

    ReplyDelete
  16. "I think I’d have said that the uvula (the active articulator) touches the back of the tongue (the passive articulator) rather than the other way round: but no matter."

    Though the uvula bounces around in the airstream, in my experience, it is the action of the tongue forming the right shape that allows the airstream to make it trill. If you "set up" to trill the uvula, you probably will feel the tongue move UP to the uvula, not the uvula dropping down to the tongue! I don't think it is "no matter" if you're trying to teach someone who can't trill!

    I must say that I find it MUCH harder to teach an apical trill than a uvular one. And the fact that in most cases a fricative is "good enough" means that we have an 'out' if the speaker can't manage to master the uvular trill. Apical trills don't seem to have the benefit of the alternate pronunciation as a fallback position.

    In my experience of speaking French with uvular trills, with less important syllables, the R 'devolves' to the fricative. I wonder if this is true of trilling Germans, too...

    ReplyDelete
  17. @Kraut: I guess that a uvular trill is also somewhat suggestive (think of Eartha Kitt as Cat Woman "RRR!", though she tended to use an apical trill -- see http://tinyurl.com/knzqag )

    ReplyDelete
  18. Question from an interested layman- which would the Edith Piaf Je Ne Regrette RRRRRien be?

    ReplyDelete
  19. 1. Childhood memories from that tape I heard all the time.

    2. Uvular trill (stage standard, as opposed to the regular Parisian fricative).

    ReplyDelete
  20. Edith Piaf is well known for her strong uvular trill. In French, we use both trill and fricative Rs. The uvular trills people make are not as strong as Edith Piaf's. In speech, it is hard to distinguish between the two when someone is talking to you, and as trills aren't very strong, we don't notice them (as trills). But on waveforms and spectrograms, we see that both occur rather frequently.
    Then since what we hear is a fricative one, and since it seems easier to produce, thus uvulars fricative Rs should be the type taught at school (and I think it is)

    ReplyDelete
  21. nɔ̃ ʀjɛ̃ də ʀjɛ̃ nɔ̃ ʒə nə ʀəɡʀɛtə ʀjɛ̃
    ni lə bjɛ̃ kɔ̃ ma fɛ
    ni lə mal tu sa me bjɛ̃n eɡal

    ReplyDelete
  22. I don't get it, John Wells.

    OK, part of the Piaf song is in there somewhere, with half the sounds missing and nasalizations in funny places. Or are my fonts playing up?

    ReplyDelete
  23. Fine here, must be the latter.

    ReplyDelete
  24. @ Lance: must be your fonts. It looks perfect on my screen.
    The orthography is
    Non, rien de rien, non, je ne regrette rien:
    ni le bien qu'on m'a fait
    ni le mal, tout ça m'est bien égal.
    These words are very prominent in my memory this week, because a choir I am in is singing Piaf's song later this week.

    ReplyDelete
  25. @Eric Armstrong: "I must say that I find it MUCH harder to teach an apical trill than a uvular one. And the fact that in most cases a fricative is "good enough" means that we have an 'out' if the speaker can't manage to master the uvular trill. Apical trills don't seem to have the benefit of the alternate pronunciation as a fallback position."

    I have had no luck at all teaching apical trills; or my clients have had no luck learning them. But my own uvular trill, the production of which was hit-or-miss, improved immensely when I began teaching non-regional American pronunciation to an Austrian client; I asked her how she made the sound, did what she said and - hey presto! - it worked. I have since had equally dramatic results teaching uvular trills to my clients.

    I do tend to use a uvular fricative more often than not when speaking French, or English in a French accent, but since I can now produce a genuine trill reasonably consistently, I am no longer limited to doing so faute de mieux.

    ReplyDelete
  26. Does any language distinguish these sounds phonologically, or are they 'always' in free variation?
    For what it's worth, I can't always guarantee to produce a trill - I think it has to do with the amount of mucus (see Kraut's comment above).

    ReplyDelete
  27. But John, shouldn't that last line be:
    ni lə mal tu sa mɛ bjɛ̃n eɡal

    My French is somewhat rusty, but I think I remember at least that much.

    ReplyDelete
  28. David Marjanović15 August 2009 at 19:12

    which would the Edith Piaf Je Ne Regrette RRRRRien be?

    An extremely long (like, 10 contacts) uvular trill.

    Uvular trill (stage standard, as opposed to the regular Parisian fricative).

    I keep reading that. But I live in Paris most of the year now, and nobody there ever uses [ʁ]. There are some French people who do, but not in Paris.

    What happens in Paris (and not in Québec) is that the [ʀ] becomes partly or entirely voiceless in all positions except between vowels, and usually this voiceless trill turns partly or entirely into a voiceless fricative. (This sound, BTW, is closer to [x] than to [χ] – unlike the dorso-uvular [q], [ʀ] does not entail a retraction of the tongue.)

    In German, /r/ isn't pronounced as a consonant in most of these positions. In other words, [ʀ] is always followed by a vowel, and therefore isn't devoiced.

    The uvular trills people make are not as strong as Edith Piaf's.

    That is, much shorter: usually a single contact, like in contemporary German.

    The trill is still sometimes considered the standard, but it's dying out.

    Yes, I think the uvular fricative is the most common form of r in German.

    I'm told by reliable sources (such as this professor of linguistics who lives and teaches in Bremen) that this is the case in northern Germany. But until recently, I (from Austria) didn't know anyone used that sound in German at all. I use a short uvular trill (at most two contacts).

    By the way, the Northern/Southern division one can often read is much too simplified. Stereotypical coastal German has a distinctive alveolar trill, and the Swabian dialect in the South has rhotacised vowels. It's more about marginal vs central, if at all.

    Probably it's even more complicated. It does seem to me that everyone who speaks Bavarian dialect uses the laminal-alveolar trill, but in Austria, where closely related dialects are spoken, at most half of the country still uses it, while the rest switched to the uvular trill generations ago.

    I'm still not clear on the consonantal/vocalic R point of German. For example, the word "Lehrer" is usually /le:ʁɐ/. That is a vocalic R. What if you make the word feminine in "Lehrerin"? Does it become a consonantal R? When in Germany recently, I said /le:ʁəʁin/ but I don't think that was right. I'm really not sure how you're supposed to say "Lehrerin" in German.

    Here's what I figured out when I was little:

    1) Starting from the spelling, every r changes the preceding vowel (if there is one) into a diphthong that begins with the written vowel and ends in [ɐ̯].
    2) Unstressed [ɛɐ̯] becomes [ɐ].
    3) [aɐ̯] and [aːɐ̯] consist of too similar sounds to form a diphthong and thus both become [aː].
    4) Now that the presence of r is marked (never mind rule 3...), it can disappear, and so it does, except when a vowel follows.

    For northern Germany, you have to add an additional rule:

    5) Rule 1 does not apply if the preceding vowel is long.

    Then it gets complicated (parts of western Germany are partially rhotic, and other parts reportedly lack rule 3), but I'll ignore that.

    So, ignoring the subtleties of /ɪ/ realizations, Lehrerin is [ˈleːɐ̯ʀɐʀɪn] for me, and [ˈleːʀɐʀɪn] or rather [ˈleːʁɐʁɪn] in northern Germany; Lehrer is the same without the -[ʀɪn] part.

    Where I come from, people have a lot of trouble imagining that the French word libération, which its unstressed short [eʀ], could at all exist; it contradicts non-northern Standard German phonology in no less than four points.

    But John, shouldn't that last line be:
    ni lə mal tu sa mɛ bjɛ̃n eɡal


    Most Parisians, though not all, pronounce est with /e/ rather than /ɛ/. But I don't know if Piaf did.

    ReplyDelete
  29. Hi..
    The French language is perhaps the most well-known example of uvular R language.This is great article about "uvular r". I liked this topic. I will surely bookmark it for future use. Good Work! Please keep going and continue.I will keep visiting this blog very often.
    Thank you very much for sharing this.
    pc-spiele

    ReplyDelete
  30. To much information for in a very little amount of time, I'll try to get a little further information on this to clear this up for me. I am still new with this so I might be jumping ahead a little to fast.

    ReplyDelete
  31. Interesting enough, a new book on French phonetics and diction adopted as a text at UCF
    ignores the uvular trill /R/ completely and uses only the aveolar trill /r/. I was wondering whether the uvular trill /R/ had passed from current educated usage in France.

    ReplyDelete
  32. Graham: "Does any language distinguish these sounds phonologically, or are they 'always' in free variation?"

    For apical vs uvular R, yes: Portuguese and Arabic, at least; possibly also Pharsi. (Also Spanish, for apical trill vs apical flap.)

    ReplyDelete
  33. Uvular trills are VERY common in Dutch, anyone who says otherwise has clearly never been here. And it's not just old posh people and the South either, you hear it in Utrecht (a lot) and Amsterdam too, from young and old alike. It's a curious thing in that it seems to appear irregularly during speech, most especially when stressing words.

    ReplyDelete