Singing in Nonsense.

Vittoria Traverso has yet another great Atlas Obscura post:

Before children learn how to speak properly, they go through a period of imitating the sounds they hear, with occasionally hilarious results, at least for their parents. Baby talk evolves into proto-words, so that “octopus” might come out as “appah-duece,” or “strawberry” as “store-belly.” But it’s not just children who ape the sounds of spoken language. There’s a long tradition of songs that “sound” like another language without actually meaning anything. In Italy, for example, beginning in the 1950s, American songs, films, and jingles inspired a diverse range of “American sounding” cultural products.

The most famous is probably “Prisencolinensinainciusol,” a 1972 song composed by legendary Italian entertainer Adriano Celentano and performed by him and his wife, Claudia Mori. The song’s lyrics sound phonetically like American English—or at least what many Italians hear when an American speaks—but are clearly total, utter, delightful nonsense. You really have to hear it to appreciate it. […]

By the time Celentano’s song came out, the sound of American English had been “contaminating” Italian culture for decades. Linguist Giuseppe Antonelli analyzed Italian pop songs produced between 1958 and 2007, and revealed the ways in which Italian singers have incorporated American sounds into their music.

One way was to use intermittent English words, with preference for trendy terms. The most notable example of this is “Tu vuo’ fa l’americano” (“You Want to Be American”), a 1956 song by Renato Carosone about a young Neapolitan who is trying to impress a girl. […]

Similarly, in the 1960s there was a trend of bands in England singing in Italian—with strong English accents. Both phenomena resulted in a similar hybrid sound, one that Italians responded to. According to Francesco Ciabattoni, who teaches Italian culture and literature at Georgetown University, this Anglo-Italian pop genre grew from Italy’s collective interest in America, as well as the British Invasion of the 1960s. “I am not sure how much thinking they put in it, but producers must have realized that imitating English and American sounds would sell more,” he says. Linguistics may have played a role, too. “The phonetic structure of English makes it more suited to rock or pop songs compared with Italian,” he adds.

She goes on to discuss the Divine Comedy (“Nimrod … approaches Dante and Virgil, and says ‘Raphèl maí amècche zabí almi,’ a series of words that has no meaning but, according to some scholars, can sound a little like Old Hebrew”), Charlie Chaplin (“The otherwise silent 1936 film Modern Times sees the comedian performing a song that sounds like a mix of Italian and French, but means absolutely nothing”), and Grammelot, “a system of languages popularized by Commedia dell’arte”:

Grammelot was used by itinerant performers to “sound” like they were performing in a local language by a using macaronic and onomatopoeic elements together with mimicry and mime. Dario Fo, an Italian playwright and actor who won the Nobel Prize for Literature in 1997, featured Grammelot in his 1996 show Mistero Buffo (Comic Mystery Play).

There are, of course, illustrative video clips at the link. (N.b.: “Prisencolinensinainciusol” was discussed at the Log in 2011.) Thanks, Trevor!

Comments

  1. marie-lucie says:

    That Chaplin scene where he is a waiter in some sort of night club and is asked to sing is priceless. Having forgotten the actual words, he ends up using French words just for the sake of sound and the people are quite happy with the performance.

    In The Great Dictator he does not sing but at one point he tells ‘a proverb in his native dialect’ which sounds very convincing to his audience. Crucially, it ends on “Ü!” spoken on a high tone, which carries the day.

  2. Other examples of nonsense lyrics are the work of Cocteau Twins, whose singer Elizabeth Fraser has explained how for much of the band’s early career she drew lyrics from foreign-language texts without knowing or caring what they meant. When so many rock or pop songs are just doggerel, this proved a good strategy to making songs with real staying power were there’s no risk of disappointing people with naff lyrics. (The shoegazing genre, where the words are ordinary English lyrics but the vocals are so low in the mix you can’t actually distinguish them, was another good strategy.)

    The claim that English is better suited to rock music is, within the English-speaking world, so often further refined into the claim that specifically American English is better suited than other accents. Recently I was struck by listening to the work of Glasgow band Camera Obscura because their singer Traceyanne Campbell – while she uses a more general UK English instead of Scots in terms of lexicon – does not hide the phonetic quirks of her native accent, and that feels so shocking for pop music where singers often ape American pronunciation. At the same time, I was listening to Who’s Next, for example, where Roger Daltrey and Pete Townshend sound so very earthy, roots American you’d never guess The Who were an UK act.

  3. Having not listened to “Prisencolinensinainciusol” in a few years, I naturally searched for in on YouTube. I was thrilled to see that there was a new performance available. It has much higher production values than the older videos, and has additions that take the whole thing to an even more outrageous level.

  4. For a time, some American pop punk (most infamously Green Day) adopted a signature Cockney-ish accent.

    To my taste, Spanish is wonderfully suitable for punk, French not at all. However French is an excellent language for rap.

  5. For those fluent in both American and British Englishes, is BrE naff exactly equivalent to AmE tacky?

  6. > Green Day […] adopted a signature Cockney-ish accent.

    I’d love to hear that, what songs should I listen to, and what should I listen for? I just listened to “Basket Case” but didn’t notice much of a Cockney accent. “Whore” and “bore” were both pronounced in a clearly rhotic way.

    Random anecdata: I’ve noticed some Japanese pop singers use over-aspirated /t/s (Japanese /t/s are only slightly aspirated) and diphthongized /o:/s, presumably to sound more like English, whether consciously or not.

  7. Try “Welcome to Paradise”. As I recall it was some of the vowels, if not the r’s. I haven’t listened to it in many years, though.

    Penelope Eckert identified recent California vowel shifts in some other bands’ accents, but the singer of Green Day has actually called himself “an American guy faking an English accent faking an American accent”.

  8. The one thing that puzzles me about “Prisencolinensinainciusol” is the thoroughly un-English lengthened trill that he uses towards the beginning. Maybe just a bit of fun that he couldn’t resist despite the logic of the song?

  9. There are a number of genres of folk music that tend to use a lot of nonsense words (although usually they only form a minority of the lyrics). Traditional cowboy songs are one type (although the use of nonsense words instead of “real” wordplay in cowboy poetry has frequently been maligned). English folk songs use a lot of nonsense syllables too; Bombadill is a particularly English genius loci.

    Of course, there are more recent forms as well, some like scat featuring few identifiable forms at all.

  10. For those fluent in both American and British Englishes, is BrE naff exactly equivalent to AmE tacky?

    I think there are few, if any, “exact equivalents” in the realm of slang.

    As used in the post in question, however, tacky seems far off the mark. For me, “tacky lyrics” doesn’t collocate; I’d say “lame lyrics.” In other contexts (style, behavior) tacky would work.

  11. Are there circumstances where lame would not translate well to naff?

  12. the thoroughly un-English lengthened trill that he uses towards the beginning

    Reminds me of Billy Stewart’s Summertime.

  13. Are there circumstances where lame would not translate well to naff?

    Yes. Really, there is no single-word equivalent in AmE, as naff can be used in broad range of circumstances. I think Wiktionary has it right in defining it as both “In poor taste” (That tie is a bit naff, don’t you think?) and as “Poorly thought out, not workable, or otherwise not very good.” (He thought the idea was a bit naff.)

  14. D. Syrovy says:

    The lyrics to the American songs Celentano recorded in the late 50s were also clearly learned just by listening to the originals: sometimes he gets it right, sometimes it turns out nonsense. They’re worth listening to, not only for their exuberance but for precisely this slightly uncanny effect.

  15. David Marjanović says:

    The Great Dictator’s speech imitates Hitler in nonsense words (some of which are misread German).

    I’ve noticed some Japanese pop singers use over-aspirated /t/s […] and diphthongized /o:/s, presumably to sound more like English, whether consciously or not.

    From what I’ve read, putting on an English accent was popular among German singers in the… 1960s or so.

    There are a number of genres of folk music that tend to use a lot of nonsense words (although usually they only form a minority of the lyrics).

    Oh, that reminds me. There’s that song I’ve sporadically encountered ever since I was old enough to notice (end-80s), most recently last week. Apparently it’s a protest song against the Vietnam war. In the lyrics I can clearly make out with their bombs and with their guns. And then it dissolves into something that isn’t even phonotactically English: repetitions of [iɔhɛ] (…for many years I wondered if it’s in your hands; it’s not) and [d̥z̥ɑɑmɛ]. What is this???

  16. It’s “Zombie” by the Cranberries. Not a protest song about the Vietnam war, but a 1993 protest song against the IRA (written in response to their public murder of two children in a shopping centre in Warrington that year). The phrases you are having trouble with are “In your head” and “Zombie”.

    Another mother’s breaking heart is taking over
    When the violence causes silence
    We must be mistaken

    It’s the same old theme since 1916
    In your head, in your head
    They’re still fighting
    With their tanks and their bombs
    And their bombs and their guns
    In your head, in your head they are dying

    In your head
    In your head
    Zombie, zombie, zombie
    What’s in your head?
    In your head
    Zombie, zombie, zombie

    https://en.wikipedia.org/wiki/Zombie_(song)

  17. David Marjanović says:

    Ah, thanks. I also recognize “they’re still fighting”.

    Still, how can the /d/ of head completely disappear (Trump’s final /g/ in big-league is unreleased, but it’s still there), and how can zombie shift that far?

    1993? I was sure I had encountered it in elementary school, ending 1992. ~:-|

  18. Actually 1994 – my mistake. The bombing was in 1993 but the song came out the year after.

  19. I imagine that the final d of head has been turned into what I think (probably incorrectly) is called a glottal stop; common replacement for t and k in Scots and Irish accents in words like butter, which becomes buh-er, or Mike, who becomes Mi’.

  20. David Marjanović says:

    Yup, this is it. The z is not prestopped (though it is devoiced), and zombie ends in what evidently counts as /ɪ/; but while the d of head seems to be there as a flap at 2 or 3 occasions (2:50, 3:00), it’s still not there most of the time. (Maybe it’s an unreleased first half of a flap!)

    Edit: the d is not a glottal stop. A glottal stop is… holding your breath; I’ve learned to hear those, even the unreleased ones.

  21. There’s that song I’ve sporadically encountered ever since I was old enough to notice (end-80s), most recently last week.

    “Zombie” received a lot of airplay again a couple of weeks ago when Dolores O’Riordan, the lead singer, passed away very unexpectedly, aged 46. Her singing diction was doubtless influenced by her strong County Limerick accent.

  22. Having not listened to “Prisencolinensinainciusol” in a few years, I naturally searched for in on YouTube. I was thrilled to see that there was a new performance available. It has much higher production values than the older videos, and has additions that take the whole thing to an even more outrageous level.

    Thanks, I enjoyed that… but I admit I felt a slight trepidation clicking on the link for fear I was about to be rickrolled.

    “Zombie” received a lot of airplay again a couple of weeks ago when Dolores O’Riordan, the lead singer, passed away very unexpectedly, aged 46.

    Yes, I discovered it (and the band) then, and thought it was a fine, powerful song.

  23. Reminds me of Billy Stewart’s Summertime.

    Hmm – that works, I guess. Objection withdrawn.

  24. I’m not surprised not to hear the /d/ in “head”, and I don’t hear it as peculiarly Irish. You don’t hear the final /d/s in the line “Hey Jude, don’t make it bad” either (on YouTube, listen to the Vevo version introduced by David Frost, and the McCartney White House version).

    As for [d̥z̥ɑɑmɛ], I don’t hear the [d], I do hear the [b], the first vowel is just spread over two notes, and I’d guess the second one is a realisation of final [i] (AAVE-influenced?) that’s pretty common in rock singing; compare Mick Jagger on “Angie”.

    That’s now, of course; when I first heard the song years ago, I didn’t know what the word was either. And I always heard “same old theme” as “same old scene”, but I see now she’s pronouncing /θ/ [tʰ], which is right and fitting, though I’d say [t̪im].

  25. I think there is a general tendency to drop final stops when singing, especially on syllables that are artificially lengthened by the music: “Hey Juuuuu, don make it baaaaa”, even if this violates phonotactic constraints. (The /k/ is preserved because can be transferred to the onset of it.) I blame the singing teachers, who are still trying to make English sound like Italian when sung.

  26. I can hear all the /d/’s at the end of ‘head’ in Zombie. They are not released but they are there. And they are not simply a cutoff in voicing; there is a definite stop there.

  27. I can hear all the /d/’s at the end of ‘head’ in Zombie. They are not released but they are there. And they are not simply a cutoff in voicing; there is a definite stop there.

    Yes, I think so, too. To me, it’s perfectly clear. Though, since I’ve known the lyrics of the song verbatim for a quarter of a century, I’m a very doubtful witness. All popular music fans are familiar with the “The Girl With Colitis Goes By” phenomenon.

  28. Vonlenska

    Yes. Just so. Love Sigur Rós, especially the early years.

  29. I’m surprised nobody mentioned this instance of a deliberately-created gibberish (not for singing, admittedly), which I very much enjoyed as a child:

    https://en.wikipedia.org/wiki/La_Linea_(TV_series)

    Closer to home (well, from my point of view anyway), a number of the Cirque du Soleil’s songs are in an entirely made-up language (but which seems to have a clearly-defined phonology), which some anglophones call “Cirquish”.

    Breffni: because “theme” in the song is made to rhyme with “sixteen” I too long mistook the word for “scene”.

  30. ə de vivre says:

    It’s only tangentially related (if that), but I spent a wonderful afternoon this past December listening to a Christmas album of quebecois metal bands in a mostly empty coffee shop in the McGill ghetto.

  31. David Marjanović says:

    You don’t hear the final /d/s in the line “Hey Jude, don’t make it bad” either

    I’m not sure if I’ve ever heard a professional sing that one. We had to sing it in school, repeatedly, because the authors of our music “text”book evidently believed it was famous. (Mid-late 90s.)

    there is a definite stop there

    Huh. I’ve never tried to learn Cantonese or the like – and apparently I shouldn’t try!

  32. David Marjanović says:

    You don’t hear the final /d/s in the line “Hey Jude, don’t make it bad” either

    Ah yeah, they’re genuinely missing. I’m not surprised every syllable-final /t/ is missing, too – British glottalization, followed by removal of all glottal stops for singing –, but the extension to /d/ is intriguing.

    (And the /t/ in better is a flap, or an imitation of one. The few remaining instances of /t/ are affricated.)

  33. I think there is a general tendency to drop final stops when singing

    but the extension to /d/ is intriguing

    I don’t agree with this summation. I don’t think McCartney was singing “Hey Juuuuu, don make it baaaaa”. I would argue that he is singing ‘Hey Jude don make it bad’ (dropping the ‘t’ in ‘don’t’ is normal in colloquial English).

    Listening to Hey Jude, the intent to pronounce the /d/ is there, but the actual articulation is extremely weak, and perhaps even comes after the note has been sung. That is a result of attempting to avoid a clearly articulated stop and release during the note, which would ruin the sonority of the song. But failure to clearly enounce the /d/ does not mean it is not there.

    Now, some might say that since the /d/ is virtually inaudible that it’s been elided. But I find that problematic. The intent to add /d/ at both places results in an audible difference. This is not so clear at ‘Jude’ — although when there is word following (as at 1:52 and 3:00), the /d/ is clearly pronounced, attached to the following syllable — but it is clear at ‘bad’. If he were really singing ‘baaa’ the auditory effect would be to sound like a sheep. He does not sound like a sheep because he can be heard to be moving towards a stop. He is actually articulating the stop, even if it isn’t audible during the note.

    Perhaps this is being oversubtle, but as long as there is an audible difference between someone singing ‘Hey Jude don make it bad’ and ‘Hey Jew don make it baaa’, then it’s problematic to say that the /d/ isn’t there.

    In Zombie, the presence of /d/ is even clearer. If Dolores O’Riordan were singing ‘hea’, the absence of a stop would be plainly audible. But the stop can be heard because open vocality cuts out at the /d/. It is not glottal in nature because there is no catch in the throat. This is how /d/ is normally pronounced in many (most?) varieties of English. If this articulation is felt to be unclear in the song it’s probably because phonation when singing is different from phonation during speech.

  34. There’s that song I’ve sporadically encountered ever since I was old enough to notice (end-80s), most recently last week.

    It seems everyone has their own version of it.
    I can’t make out the bolded part “Minun nimeni on Liisa. Olen tyttö, senhän kuulee jo nimestäkin. ” (My name is Liisa. I’m a girl, which is obvious just from the sound of it.)(A translation into Finnish of Astrid Lingren’s Alla vi barn i Bullerbyn.

  35. Actually when I imitate a sheep it’s more like [æʔæʔæʔæʔæʔæʔæʔ] than like [æːːːːː]; there is a glottal pulsing, almost a trill, in the sound.

  36. The phonetic structure of English makes it more suited to rock or pop songs compared with Italian

    Okay, I’m gonna get a little petty about this. I don’t think English or American English or General American Imitating African American Vernacular English (GAIAAVE, the lingua franca of all post-blues pop 😉 ) are inherently better suited for rock or pop than Italian; it’s just that the blues originated with African American speakers, who were then copied by General American English speakers and British English speakers, and that sound has been used so often that it’s basically become a register of pop music. But it’s a sociolinguistic thing, not something about the ‘phonetic structure’ of English. It’s the same reason people fake Southern accents in country music. It’s not that country music needs those particular formants or those phonological restraints to jibe with the music; it’s basically tradition.

    That nitpicking aside, there is a lovely tradition of English-language pop traveling to non-Anglo parts of the world and then being imitated without understanding the meaning. Several years ago I found a clip of some Russian cartoon that seemed to have someone making up fake English to the tune of the Beatles’ “Yesterday.” When traveling to non-English-speaking countries, I’ve heard a lot of cover bands of 1970s-era classic rock with garbled lyrics. A famous example of “so long as it sounds nice and kinda like English” is Swedish producer Max Martin, who has said that it doesn’t matter if a song makes sense as long as the vowel sounds are good and you have a good melody. This sort of disregard for meaning makes sense if you’re able to listen to foreign language songs where all you’re enjoying is melody and vowels + consonants.

    I think you’re less likely to find English speakers making stuff up to sound foreign in pop music, but that doesn’t mean there aren’t exceptions. The foremost one in my head is Lady Gaga’s “Scheisse,” which starts by saying “I don’t speak German, but I can if you like” and then fake German:

    Ich schleiban austa be clair,
    Es kumpent madre monstère,
    Aus-be aus-can-be flaugen,
    Begun be üske but-bair.

    Ich schleiban austa be clair,
    Es kumpent madre monstère,
    Aus-be aus-can-be flaugen,
    Fräulein uske-be clair

  37. Wenn ist das Nunstück git und Slotermeyer? Ja! Beiherhund das oder die Flipperwald gesput!

  38. Great, you’ve just killed all the German-speakers.

  39. The best, most convincing faux-English I’ve ever heard was by several caged magpies. I actually listened to them very attentively for a minute or so, trying hard to catch at least a single word.

  40. “The Girl With Colitis Goes By”

    Wonderful! And in these well-fed times, it is actually a little plausible, people with pathological malabsorption tend to be thinner and more attractive.

  41. Since the vowels in “bad” and “head” cannot appear in open syllables in English, one could hypothesize that these syllables are more prone to “d-dropping”, since listeners will (subconsciously, of course) know something is wrong and try to reconstruct the /d/. It could be interesting to see if [d] in, e.g. “made”, “seed” is dropped just as often in songs.

    > “The Girl With Colitis Goes By”

    This made my day!

  42. If any Germanophones are left standing after that, I’d like to know about the following claim by Martin Haspelmath. He says that while the verb ‘dig’ conceptually allows four arguments, (a) the digger, (b) the hole, (c) the ground that is dug, and optionally (d) the thing being searched for there (you can dig holes just for the lulz, but it’s not the usual thing), German only allows some of these to be expressed in the same sentence:

    1) Sie gruben ein tiefes Loch in die Erde ‘They dug a deep hole in the.ACC ground’ (a, b, c)

    2) Sie gruben in der Erde nach dem Schatz ‘They dug in the.DAT ground for the treasure’ (a, c, d)

    but not

    3) *Sie gruben ein tiefes Loch nach dem Schatz ‘They dug a deep hole for the treasure’ (a, b, d)

    Is this really true? Why on earth?

  43. ‘They dug a deep hole for the treasure’

    Well even in English for me, this sentence doesn’t parse as a combination of “They dug a deep hole” and “They dug for the treasure”. I can only read it with “a deep hole for the treasure” as one argument – which, if anything, I think would more likely imply that they’re digging a hole in which to bury the treasure.

  44. FWIW, I can report than in Danish, I personally also find the b+d combination slightly strange:

    ??De gravede et dybt hul efter skatten.

    I think it has to do with telicity. “Digging a deep hole” is telic, “digging for the treasure” is atelic. If you’re digging for a treasure, you don’t know how deep of a hole you’ll be digging. You can dig for a treasure, and end up digging a deep hole in the process, though.

    Edit: Come to think of it, we can test it. They dug for the treasure for a month. They dug a big hole in a month. They dug a big hole for the treasure (??) a month.

  45. Unrelated but it’s so LH stuff!

    Scrambled in translation? Norway Olympics team orders 15,000 eggs by mistake

    https://www.theguardian.com/world/2018/feb/08/norway-olympics-team-orders-15000-eggs-by-mistake-south-korea

  46. David Marjanović says:

    you can dig holes just for the lulz, but it’s not the usual thing

    I’m told it’s a very common thing in army training to dig trenches and fill them up again for the instructor’s lulz, though.

    Is this really true?

    Yup.

    Why on earth?

    Lazar has it. (The example feels as if the verb were transitive and intransitive at the same time, which definitely isn’t allowed.) Indeed, sie gruben ein tiefes Loch für den Schatz, a hole to put the treasure in, is A-OK.

  47. Karen,

    Swedish producer Max Martin, who has said that it doesn’t matter if a song makes sense as long as the vowel sounds are good and you have a good melody

    That’s definitely the principle that a lot of artists work on. Oasis are a good example: their songs are designed to be sung full-throatedly after the pub and a feed of chips. “And soooooo Sally can waaaaait, she knows it’s too laaate as we’re walkin’ on byyyy / Her sooooul sliiides awaaay…”. Complete nonsense, but extra points for “walk on by”, which is the “somewhere a dog barked” of pop lyrics.

  48. That’s also the principle Pasternak worked on, at least some of the time. Who cares what it means if it’s got a good beat and you can swoon to it?

  49. Scrambled in translation? Norway Olympics team orders 15,000 eggs by mistake

    Uffda!

  50. Dropping ‘deep’ would get rid of the telicity, I think, but David’s transitive/intransitive problem still stands. I dug a hole (in the ground) for the treasure can still be read Lazar-style, but I think the other reading is now possible, or less impossible.

    Perhaps the success of the Lennon-McCartney writing partnership depended on having one who likes to write nonsense and one who doesn’t care about sense. “Scrambled eggs / Oh my baby how I love your legs / Not as much as I love scrambled eggs.”

  51. People make way too much of lyrics in pop music (don’t get me started on Dylan’s Nobel). They’ve got to be there (barring the occasional instrumental), but they don’t gotta make sense. Oh well a bird bird bird, bird is a word!

  52. one who likes to write nonsense and one who doesn’t care about sense

    Nice distinction, in both senses.

    There aren’t that many Lennon/McCartney songs that were really written together, and ever fewer as time went on, though they did influence each in more intangible ways. But the Lennon/Martin and McCartney/Martin partnerships were at least as important IMHO.

    That’s also the principle Pasternak worked on

    It’s not every day you see Noel Gallagher compared to Pasternak.

  53. People make way too much of lyrics in pop music

    Well, for sufficiently narrow values of pop. Folk and country live and die by the lyrics, though singability is still an issue of course.

  54. Lars (the original one) says:

    I agree with dainichi, in Danish you will normally make the telicity structure explicit as in de gravede et dybt hul da de ledte efter skatten.

    I’m not a native speaker, but I think it sounds more natural in English to use a progressive tense with digging for treasure unless other factors are in play, while they dug a hole is unmarked. Perhaps this is a related phenomenon.

    In the new production of Prisencolinensinainciusol, note the cameo at 4:20.

  55. David Marjanović says:

    I just watched Hey, Jude again. Right at the beginning (0:50), his face is shown up close, and it looks like he doesn’t articulate a [d] in Jude. Then, in bad, he probably does – but by that time he’s been silent for half a second! At 3:40, you can see lack of an articulated [d] in bad.

    but it is clear at ‘bad’. If he were really singing ‘baaa’ the auditory effect would be to sound like a sheep.

    Would that still be the same vowel? [æ] is a wide field, and I’d probably pick a more fronted or epiglottalized spot in it to imitate a sheep than for bad (apart from the glottal constriction mentioned above). IIRC, so did Rowan Atkinson as Blackadder.

    In that sense, the /d/ would still be there, but a [d] is nonetheless completely absent.

    On to Zombie

  56. David Marjanović says:

    BTW, I’m a bit surprised clearly articulated syllable-final stops are supposed to be avoided in English singing. In German, singers are taught to exaggerate them, all as part of a desperate attempt to keep the lyrics acoustically understandable for the audience.

  57. As I say, I blame the Italians, who have no final stops to speak of.

  58. David Marjanović says:

    Zombie: During the “text” part, the [d] in head is probably articulated, if barely. During the “aria” part, I’m probably imagining it.

    The mb in zombie has a strong tendency to be contracted, probably to [b̃] or thereabouts. On the one hand, I tend to hear voiced stops as nasalized and/or prenasalized because they don’t occur in my kinds of German; on the other hand, the beginning of in your is contracted to [ɪ̃j̃]- (as often happens in English, of course), so perhaps this happens to the mb as well.

    All of this could be part of a general effort to avoid closed syllables. In the staccato part, there’s mistaken with a pause after [mɵ]. Sure, the [s] can’t be in the same syllable because then the /t/ would become aspirated, which would be wrong, but why not extend it across both syllables? Here’s a fine specimen of Very West Germanic Consonant Gemination Indeed, |k.j| becoming [kːj] so the /k/ can be onset and coda. I guess O’Riordan would rather not have a coda at all? (…But then, Allen seems to drop the /d/ in inspired and tired near the start of the video.)

    I’m not sure why I used to hear the z as an affricate. Probably because it’s unexpectedly short.

  59. David Marjanović says:

    But why wasn’t German subject to the same Italian tradition? Opera singing in German started only during Mozart’s career (at least according to Austrian mythology, which I haven’t bothered to fact-check).

  60. @ David Marjanović

    I think we’re approaching it from totally different angles.

    You are listening and saying: “I can’t hear it. It’s not there!”

    I am listening and saying: “The sound may be virtually inaudible but he’s not singing ‘Jew’ and ‘baa’. If he were you would hear it.”

    His intention is what makes the difference. Even if his tongue touched the teeth ridge after the note was sung, or just touched it very lightly, the effect would be different from intentionally singing ‘Jew’ and ‘baa’. If he had sung a full-throated ‘Jew’ or ‘baa’, you would know it. I suspect that if a non-native singer were to try that trick (‘I can dispense with the /d/s ‘coz they’re not there’) you would notice the difference.

    In Zombie, the /om/ is turned into a nasalised /o/.

    Sure, the [s] can’t be in the same syllable because then the /t/ would become aspirated, which would be wrong

    I think this is a matter of syllabification. I don’t think it’s normal in English to pronounce ‘mistaken’ as ‘mis-taken’. It’s pronounced ‘mi-staken’ in ordinary speech, which is why /mi/ turns into schwa.

    I’ve never heard Lily Allen. At places she sings with something approaching a Cockney accent, glottal stops and all.

  61. @ David Marjanović

    Do you feel the /d/s are also missing from this Cranberries song?

    Hollywood

    (I will admit, however, that the /k/ in ‘like’ is completely missing).

  62. David Marjanović says:

    I’m not sure, but I think they’re only manifested as the preceding checked vowel, not as [d]. I probably can’t turn the sound on loud enough, though, because then the background music would shake the neighbors… I first thought I could hear one around 4:20, but that tap is glottal, not lingual.

    What is shown as understood is probably understand, and lacks any trace of /d/, like in the South African anthem (“and united we shall stan / […] / in Southafrica, ou lan”).

    (I will admit, however, that the /k/ in ‘like’ is completely missing).

    Contrary to my expectations, it is! Even though every /t/ is there in some form (unreleased in not so, sometimes a pure glottal stop in not Hollywood).

  63. Trond Engen says:

    Bathrobe: His intention is what makes the difference. Even if his tongue touched the teeth ridge after the note was sung, or just touched it very lightly, the effect would be different from intentionally singing ‘Jew’ and ‘baa’. If he had sung a full-throated ‘Jew’ or ‘baa’, you would know it.

    Pronouncing the vowel while thinking the word. This is how Danish works (or at least increasingly from after the onset of the first syllable).

  64. Trond Engen says:

    It’s a joke, of course, but I still believe it.

    The consonants are there as barely audible changes in the pronunciation of the vowel, small adjustments in expectation of a consonant that’s never materialized, probably measurable somewhere in the balance between the formants. I think it works as long as an overt pronunciation is still around to be heard, being a model for people to hear what they don’t hear. If that is lost, you have a complicated vowel system that needs to be resolved, either by exaggerating the variation in the vowel sound into a tonal system or by letting the vowels merge and add morphemes. I think Danish is there, though the process is slowed down by the existence of the written language.

  65. @ David Marjanović

    I can only say we listen to English very differently.

    If I understand you correctly, you agree that there is no /k/ in ‘like’. She is singing ‘lie’, with a long, upwardly inflected (almost yodelled) vocal tail that sounds like /i/. No /k/ there at all.

    The /d/ is definitely there, stopped but not released. If you say that is not a [d], then all I can say is that one of the most common pronunciations of /d/ in syllable-final position in English is not [d].

    There is no glottal stop in ‘not Hollywood’. It’s simply an ordinary /t/, unvoiced, with release delayed until the onset of the following /h/. If it were a straight glottal stop there would only be a catch in the throat, which is not the case. Even if the stop is fairly sudden (that is, accompanied by a sudden occlusion of the vocal chords), there is an alveolar stop there, which makes it a [t].

    I agree that the /d/s in the South African national anthem are missing. It seems that the final /d/ in /nd/ is easily omitted. This is also a tendency in some styles of colloquial English, e.g., the /d/ in ‘and’. Also in expressions like “go roun’ an’ roun'”.

  66. Stu Clayton says:

    1) Sie gruben ein tiefes Loch in die Erde ‘They dug a deep hole in the.ACC ground’
    2) Sie gruben in der Erde nach dem Schatz ‘They dug in the.DAT ground for the treasure’
    3) *Sie gruben ein tiefes Loch nach dem Schatz ‘?’

    The equivalent English sentences 1a) and 2a) are faultless:

    1a) They dug a deep hole in the ground
    2a) They dug in the ground in search of the treasure

    The problem with the German 3) is that it is the result of mindless permutations of 1) and 2). However, mindless permutation of 1a) and 2a) gives the faultless

    3a) They dug a deep hole in search of the treasure

    So what gives with the German ? Above Lazar and David concur in explaining the problem with 3) by “The example feels as if the verb were transitive and intransitive at the same time, which definitely isn’t allowed”. Well, that’s true enough, but I see a more general explanation, having to do with what I’m here calling “mindless permutation”.

    “Ein X graben” means “dig an X”, “nach X graben” means “dig in search of X” or, equivalently, “search for X by digging”. It’s not a good idea to translate by counting and matching grammatical bits, here propositions – the English preposition “for” to match the German preposition “nach”.

    Translation is a kind of reformulation. To reformulate, it rarely suffices to replace parts without thinking. Before replacing old batteries by new ones, you need to check the voltage requirements.

    Semantic batteries have 12 volts, grammatical batteries have 3 volts. Look what happens when you try to replace a semantic battery by a grammatical one, say “in search of” by “for” in 2a)-3a), even staying within English. All the sentences are OK, but the meaning of 3b) is now different from that of 2b):

    2b) They dug in the ground for the treasure
    = “they dug in the ground in search of the treasure” [the purpose is to search]
    3b) They dug a deep hole for the treasure
    = “they dug a deep hole to bury the treasure” [purpose is not to search]

    The sentence 3b) doesn’t have enough power for its circuit logic to function properly.

  67. Maybe it depends on the object:

    “They dug a deep hole for the corpse.”
    “They dug a deep hole for the sequestered carbon.”

  68. >It’s pronounced ‘mi-staken’ in ordinary speech, which is why /mi/ turns into schwa.

    I believe there are weak-vowel non-merged dialects where it’s an [ɪ], which is a checked vowel, so to avoid the aspiration, I think the /s/ would have to be categorized as ambisyllabic.

    > ‘baa’. If he had sung a full-throated ‘Jew’ or ‘baa’, you would know it.

    AFAIK a lot of people actually thought he was singing “jew”. As for “baa”, I think I hear a slight ɛ-like off-glide before the airflow stops.

    > The example feels as if the verb were transitive and intransitive at the same time

    I’m not convinced about the transitivity argument… How does “Sie gruben tiefe Löcher nach dem Schatz” feel to you? At least in Danish, to me, “De gravede dybe huller efter skatten (they dug deep holes in search of the treasure)” sounds completely fine, which proves that (to me, in Danish) “efter skatten (in search of the treasure)” doesn’t necessarily require intransitivity. What it does require is atelicity, and “digging holes” is atelic.

  69. David Marjanović says:

    As for “baa”, I think I hear a slight ɛ-like off-glide before the airflow stops.

    Quite possible. You can watch him slowly close his mouth in anticipation of a [d]. But by the time he might articulate it, he’s already fallen silent for quite a while.

    (I’ll listen to both songs again.)

    At places she sings with something approaching a Cockney accent, glottal stops and all.

    Tha[ʔ̚]’s mainstream southern Bri[ʔ]ish English nowadays. Never le[ʔ̚] su[ʔ̚͡t̚ʃ] peop[o] near a bad mi[ʔ̚͡k̚]rophone.

    How does “Sie gruben tiefe Löcher nach dem Schatz” feel to you?

    The same. Repeated telic activity, it seems, is still telic.

    Sie gruben tiefe Löcher, um den Schatz zu finden would work (same meaning, but only one argument for “dig”). Sie gruben tiefe Löcher auf der Suche nach dem Schatz (ditto) would also work.

  70. David Marjanović says:

    I think we’re approaching it from totally different angles.

    You are listening and saying: “I can’t hear it. It’s not there!”

    I am listening and saying: “The sound may be virtually inaudible but he’s not singing ‘Jew’ and ‘baa’. If he were you would hear it.”

    I think you’re talking about phonemes while I’m talking about sounds. Perhaps I’m saying the Cheshire Cat is gone, including its teeth, but you’re saying its grin is still there; and perhaps we’re even both right.

    All the non-initial oral stops in Hollywood in order. I listened to each several times.

    got: [ɾ].
    picture: unreleased [k̚]; I’m not sure if I hear [t͡ʃ] or just [ʃ], though it’s probably the former.
    head: absent, replaced by a sharp intake of breath.
    head again: unreleased [d̚].
    it’s: [t͡s].
    bed: voiceless but released [d̥]. That’s unexpected…
    bed again: same thing again! Perhaps connected to the immediately following inhalation (in both cases).
    most: [d̥], released, and then of starts with its own voice onset.
    not: [t̚ʼ] – there’s an alveolar and a glottal closure, and neither is audibly released. Interestingly, the /h/ of the following Hollywood is reduced to a short [ɦ].
    Hollywood: a short flap that fades out during its “release”.
    like: absent.
    understand (I’m pretty sure that’s it and not “understood”): [ɾ̃] like in American.
    understand: [d̥]. BTW, that’s also what this starts with.
    understand: nasalization of the preceding vowel, no stop (oral or nasal).
    not: [t̚ʼ] again, and Hollywood starts with [ɦ].
    Hollywood: …might be an unreleased voiceless flap, or absent.
    like: absent all 3 times.
    not: [t̚ʼ], Hollywood with [ɦ].
    Hollywood: I think there’s genuinely no stop there this time. Perhaps an approximant aimed at a flap but not going that far.
    like: absent.
    understand: just [n] this time.
    understand: [d̥].
    understand: no stop, little if any nasalization of the vowel; it’s just the vowel quality that indicates /n/ anymore.
    not: probably [t̚ʼ] again; Hollywood with [h] this time.
    Hollywood: no closure, no constriction that I can hear, and I tried probably 10 times with -wood alone. BTW, have you noticed the [v]? We might be having ourselves a boukólos rule here.
    like: absent all 3 times.
    anybody: [b̥]; probably counts as initial.
    anybody: [ɾ]; there begins with [ɾ], too.
    anybody: voiced [b] this time.
    anybody: [ɾ].
    get: [ɾ] all 6 times.
    picture: as above, though with clearer [t].
    should: drives me crazy, but is probably absent after all, leaving a checked vowel behind.
    greatest: [t]. Released unaspirated fortis, no glottalization.
    greatest: [d̥] or [t] (or somewhere between, it’s a spectrum); released, irony even begins with its own voice onset.
    shoot: [ɾ̥].
    it’s: [d̥͡s], possibly glottalized a bit. (Lenis-to-fortis affricates are common in word-final position in my dialect and even in Austrian Standard German. I don’t have a [t] in deutsch.)
    not: [t̚], apparently not glottalized; not forming an affricate with the following [s].
    at: [ɾ].
    not: [t̚ʼ] again; Hollywood with [ɦ].
    Hollywood: approximant aimed at a flap?
    like: absent.
    understand: like last time.
    not: [t̚ʼ] again; Hollywood with [h].
    Hollywood: unreleased approximant!
    like: absent all 3 times.
    not (3:22): this time I’m sure it’s just [ʔ̚]. Followed by [h].
    Hollywood: sacrificed to inhalation.
    like: absent.
    not: [t̚ʼ] again; Hollywood with [h].
    like: absent all 3 times.
    anybody: [b] both times.
    anybody: [ɾ] both times.
    get: [ɾ] all 6 times.
    not: probably [t̚]; Hollywood with [h].
    Hollywood: [ɾ], released!
    not: probably [t̚ʼ]; Hollywood with [h].
    Hollywood: [ɾ] again.
    like: absent to the end, followed by [hi] at 4:38.

    That was exhausting. I’ll try the other two songs this evening at the earliest, together with an attempt to figure out what’s going on in mistaken.

    So, not counting understand where -/nd/ is a special environment, final /d/ is sometimes replaced by premature inhalation, but otherwise usually present as a consonant in a dazzling variety of forms, some of which are very reduced indeed. Complete absence as a sound still leaves a “checked”/”lax” vowel, which phonotactically requires a following consonant phoneme, as the grin of the Cheshire cat.

    Edit: putting three paragraphs into a single <i> tag puts the first and the last but not the middle paragraph in italics. ~:-|

  71. I have no /t/ (not merely no [t]) in picture. Bizarrely, only the OED and the Oxford American recognize this pronunciation for AmE; the native AmE dictionaries insist on /t/. I have to wonder if they just aren’t listening.

    Maddeningly, there is no dictionary of Hiberno-English, and probably never will be.

  72. Perhaps we are talking about phonemes and sounds, but when you identify sounds so minutely, right down to a [d̥], [ɾ] or [ɾ̥], I think you are pushing the limits of phonetic analysis. When you claim “head: absent, replaced by a sharp intake of breath”, I’m at a loss what to say. I can hear the /d/ quite clearly.

    And I’m completely mystified that you can’t hear something as clear as ‘understood’. No, it doesn’t rhyme with ‘Hollywood’, but it’s not ‘understand’, either.

    Listen to the second example of ‘understood’ at understood (a woman speaking) . It sounds a bit like that.

  73. David Marjanović says:

    And I’m completely mystified that you can’t hear something as clear as ‘understood’.

    I’m not familiar with any Irish English accent. All I can say is that the vowel is very far from both [ʊ] and [æ]; given that it doesn’t rhyme with the /ʊ/ in Hollywood, I figured it probably isn’t /ʊ/…

    the second example of ‘understood’

    How do I get there? The label “1 out of 915” seems to be a label for the whole page; there’s no link there or at the end of the long video.

  74. Yeah, it’s not a model of good website navigation.

    There’s a green circle below the video, with an arrow inside pointing to the right. It’s a link to the next video.

    The speaker is Thandie Newton.

  75. > I’m not familiar with any Irish English accent. All I can say is that the vowel is very far from both [ʊ] and [æ]

    It’s clearly “understood”, but it’s also very clear that the two allophones of /ʊ/ in “Hollywood” and “understood” are very far from each other. The former shows assimilative rounding from the /w/ and thus sounds almost like a /u/, “Holly wooed”. The latter is quite unrounded/front/low and almost enters STRUT territory, “under stud”. Her allophones don’t strike me as particularly Irish to me, though. I think they would work in RP and GA too, but I’ll let native speakers correct me if I’m wrong.

  76. Since nobody has mentioned it, I’ll mention that the British comedian Mike Reid recorded an English version as “Freezin’ cold in 89 twoso.” An English cover of an Italian imitation of English! What won’t they think of next?

    You might enjoy Cliff Nazarro: https://www.youtube.com/watch?v=aIdmfYBvQMw

  77. David Marjanović says:

    Ah, thanks. Yes, that’s a similarly unrounded & lowered allophone. I have encountered such pronunciations, but hadn’t figured out that it can depend on the absence of preceding /w/, so the lack of rhyme with Hollywood threw me off. Also, Newton’s understood happens to end in a clearly audible consonant (released even), while in the song the only hint of any final consonant is the checked vowel in which the syllable phonetically ends (every time). I agree I probably just imagined the nasalization; most likely I misinterpreted the centralization.

    I think you are pushing the limits of phonetic analysis

    I absolutely am. I’m at the limit of what I can do without looking at spectrograms (having, after all, never learned how to interpret those). After all, I’m trying to find out what it is you hear as /d/. English does a lot of things to plosives that I’m not used to from other languages; and indeed I found out that I had overlooked several of those when I listened to the Hollywood song the first time.

    When you claim “head: absent, replaced by a sharp intake of breath”, I’m at a loss what to say. I can hear the /d/ quite clearly.

    In this particular instance, she’s most likely aiming at it, but before she actually gets to it, she inhales (audibly), leaving a truncated [hɛ] behind. In most others, she doesn’t.

    Of course, head is the only possible word here, so I suppose you do what fluent speakers do and extrapolate without even noticing. But I’m not used to plosives falling off, so I notice. And while /hɛ/ wasn’t even a phonotactically possible word(-final syllable) in English before Lisa Simpson said /mɛ/, it’s a word (“hey”) in my dialect, and a possible word-final syllable in Standard German (though because non-initial /h/ is so rare, the only example I can come up with would be Sinuhe the Egyptian…).

  78. The latter is quite unrounded/front/low and almost enters STRUT territory, “under stud”. Her allophones don’t strike me as particularly Irish to me, though. I think they would work in RP and GA too, but I’ll let native speakers correct me if I’m wrong.

    Agreed, it’s not Irish, it’s pop-singerese. They take all sorts of liberties with vowels, partly for reasons of sonority: ‘(under)stood’ is held for longer than ‘(Holly)wood’, and so she probably aims for a more open vowel. Then there are various conventional affectations and allusions to other dialects, like the (lowered and retracted?) /i/ in Lisa Stansfield’s pronunciation of ‘baby’ (first line and passim), which strikes me as a contemporary-R&B thing. Be it noted I don’t endorse the song in any way.

  79. Re the Stansfield vowel, I misremembered: the “baby” in question is in the first line of the chorus, after the spoken intro; and it’s the first syllable that suffers from what I think is retracted tongue root. So [be̙bɛ] or thereabouts. The point being that there’s little enough you can safely deduce about English vowel allophony from singing.

  80. David Marjanović says:

    [be̙bɛ]

    Oh yes! That’s very common in pop singing, has been for decades.

    Edit: now that I’ve actually listened to the first minute, it’s striking how Stansfield uses different accents to speak and to sing; the singing accent could be a jazz-derived AAVE imitation.

  81. Lisa Stansfield is from Manchester, and to my mind there’s quite a Mancunian sound in her voice, especially in the spoken intro. Some of that seems to carry over into the singing. I find it hard to say how much of her sound is her natural accent and how much is due to her adopting an American soul/blues singing style.

  82. David Marjanović says:

    Soul and blues are probably better candidates than jazz.

  83. > ‘(under)stood’ is held for longer than ‘(Holly)wood’, and so she probably aims for a more open vowel.

    My take was she uses a rounded/back allophone for “wood” because it is a “high note”, sung with a falsetto-like voice, either by convention or because it helps her sing it that way.

    In several places, Lisa Stansfield sings “back” with a weird /k/ that sounds like a [x] or [χ] to me, which is a phenomenon I hadn’t noticed before.

  84. My take was she uses a rounded/back allophone for “wood” because it is a “high note”, sung with a falsetto-like voice, either by convention or because it helps her sing it that way.

    Maybe. But I was a little surprised at Thandie Newton’s pronunciation of ‘understood’. She was born in London and presumably has a Southern English accent. I would never use that pronunciation of ‘stood’, which for me always rhymes with ‘could’, ‘should’, ‘hood’, or ‘good’.

    So I wonder whether the kind or pronunciation used by Thandie Newton might not be common in southern England, and might also be current in the part of Ireland that Dolores O’Riordan came from. Somehow it didn’t strike me as just an artifice of the singing.

  85. It’s not “falsetto-like”; she’s just singing falsetto. Women can, you know.

  86. What’s the difference between a countertenor and a pair of dentures?

    One is a falsetto voice, and the other is a falsetto teeth.

  87. @Bathrobe: The vowel that Newton uses there seems consistent with other innovative values of /ʊ/ that I’ve heard from southern English folk; she uses it as well in “babyhood” at 1:08, and “should” at 1:44.

  88. As a child, my wife heard a strange word in Bing Crosby’s rendition of I’m Dreaming of a White Christmas, and asked her adults, “What’s a /kɑrdəraɪ/”? The lyric is of course “(With every Christmas) card I write”, but sure enough, Bing omits the /t/ altogether, not even a glottal stop.

  89. Trond Engen says:

    Heh. I’ve always heard that as “like every Christmas card I know”.

  90. I think you have conflated verse 1 line 2, “Just like the ones I used to know” with verse 2 line 2, “With every Christmas card I write”. It’s the most forgettable of Berlin’s lyrics, not that it’s actually possible for an American to forget it. Perhaps because of this, it has been translated into dozens of languages: the first verse of the official Norwegian translation is “Jeg drømmer om en hvit glede / Drømmer om jul med hvite trær / Om små barn som synger / og spurv i klynger / snedrev og i vintervær.” Which I hope is thoroughly boring and vapid, because if not it’s a mistranslation.

  91. Trond Engen says:

    I think you have conflated verse 1 line 2

    You’re right.

    the official Norwegian translation

    I’ve never heard it. And it’s not because we don’t do that. Other Holiday Classics such as Snømannen Kalle (Frosty the Snowman), Rudolf er rød på nesen (Rudolf the Red-Nosed Reindeer) and Jeg så mamma kysse nissen (I Saw Mommy Kissing Santa Claus) are in every shopping venue in December.

    Which I hope is thoroughly boring and vapid, because if not it’s a mistranslation.

    Not the case here, but good translation of songs (and other poetry) is good mistranslation. You reimagine it in another language and do whatever it takes to make all poetic elements work together to create the original sense in the listener.

  92. Lars (the original one) says:

    February 10, 2018 at 8:22 am — I somehow until now missed Trond’s earnest paean of admiration for the feats of signal recovery that Danes continually perform, but I must temper it with the admission that /s/ and many coda stops, though unreleased given the chance, will in fact interrupt voicing (or rather murmuring) and present an opportunity for resynchronization.

Trackbacks

  1. […] Hat looks at the surprisingly profound roots of singing in nonsense, in different cultures and over the age […]

Speak Your Mind

*