Tuesday, 12 June 2012


As we all know, English words spelt with u that come from Latin, or from Greek via Latin, regularly have a palatal semivowel in English. Thus Latin futur-us gives us BrE ˈfjuːtʃə, AmE ˈfjuːtʃɚ, while Latin fūtĭl-is gives us BrE ˈfjuːtaɪl, AmE ˈfjuːt̬l̩. Greek μουσικ-ή mousik-ē ends up as English music, pronounced ˈmjuːzɪk.

After alveolars the j was lost, or subject to coalescence, depending on the variety of English involved, giving the familiar variability in words such as tube tjuːb ~ tuːb ~ tʃuːb. In East Anglia the yod has been lost more widely, even in words such as music, human (and of course beautiful — here’s the locus classicus). There’s some variability in weak syllables, too, as when ambulance, executive or regular are pronounced without j in certain non-standard accents.

Disregarding these categories of exceptions, though, the rule applies pretty widely, and not only to Greco-Roman borrowings, but also to other long-established ‘international’ words (butane, Cuba, pupa). In more recent loans there may be more variability. So on the one hand Rudolf Nureyev (Russian/Tatar Нуреев, Нуриев nu-) often has nj- in BrE, while the late Sir Peter Ustinov (Russian Устинов uˈstʲinəf) was happy for his name to be anglicized as ˈjuːstɪnɒf; but on the other there is variability in Lithuania, Nicaragua, jaguar and muesli. However Zulu is firmly yodless, and Japanese futon ɸɯ̥toɴ becomes English ˈfuːtɒn, never ˈfjuː-.

When it comes to acronyms (initialisms) we again see variability. When I was a student the University of London Union, ULU, was generally known as ˈjuːluː (and I think still is). We had BUNAC ˈbjuːnæk (the British Universities North America Club) to enable us to buy cheap flights to the USA, while CICCU ˈkɪkjuː (the Cambridge Inter-Collegiate Christian Union) tried to save our souls. I believe CUNY (the City University of New York) is known as ˈkjuːni. But the British medical insurance company BUPA (British United Provident Association) is ˈbuːpə, and the GUI (graphical user interface) by which you operate your computer is a ˈɡuːi.


  1. It's curious that while, in BrE, /j/ in now generally lost after /l/ (lucid, voluminous), it is always retained after a stressed syllable: value, emolument, deluge, etc. Why should that be?

    1. I always said ˈbju:pə and was surprised when I heard BUPA TV commercials pronounce it differently.

      I don't remember hearing anything but ˈʒægju:ə — part from the spelling pronunciation:

      The Commissar for Potters Bar
      Has bought a brand new Jaguar

      [to the tune of The Red Flag]

    2. Max,
      It's because [j] is retained if there's a syllable boundary after the preceding consonant. It can only be dropped, even in AmE, if they are (would be) in the same syllable, which explains why AmE ave[.]nue usually loses it, while ven[.]ue retains it.

    3. I keep /j/ in "voluminous" (also in "lewd", "lure", "lurid", "dilute", and some others, fairly randomly).

      teardrop's principle presumably explains the AmE /j/ in "figure" an its derivatives -- yodless in BrE, although BrE does have /gj/ in "regular", "regulate", "jugular", "coagulate" etc.

      I usually hear AmE "prelude" as /ˈpreɪlud/ rather than my BrE /ˈprɛljuːd/.

    4. vp

      I started a posting then realised it was in contrast to your lju: examples. My point was that there's a whole set of cognates which may have been taken from other languages but would have been obviously Latin derived at the time they were adopted: lunar, lunatic, lunacy etc.

      Personally I wouldn't have u: in lure and lurid. I'm no doubt inconsistent, but my usual choice is a non-rhotic NURSE vowel.

      My pronunciation of voluminous is pulled in opposite directions:
      • The urge to form a 'dark' velar l prompts u: not ju:.
      • The urge to say ju: propel a 'clear' l.

      I'm not sure about lute, but I don't think I've heard ˈlju:tənɪst.

      By contrast, I can't think of any rju: pronunciation — whatever the derivation: cf Rubicon, rubella, Rufus, rule, rune, rue, Rubik. I find it hard to think of words with unstressed ru but discovered serrulated = ˈsɛrjʊleɪtɪd. I think I would pronounce any similar word similarly.

    5. I don't remember hearing anything but ˈʒægju:ə — part from the spelling pronunciation:

      I'll assume that initial consonant is a typo unless you say otherwise - I've only ever heard it with the affricate.

    6. Although I vacillate between ju: and u: in voluminous where it is stressed, I constantly have j before an unstressed vowel in voluble, valuable.

      Similarly with səˈlu:t (and just possibly səlju:t) but consistently ˈsælju ̩teɪʃn̩.

    7. David,
      It's interesting what you write about serrulated, because pronouncing it with [rj] shows that the common rule that non-rhotic speakers only pronounce [r] before a vowel is inaccurate. It even makes the observation "non-rhotic speakers have no coda [r]" questionable, since the syllable boundary is between [r] and [j] here.

  2. You're correct on CUNY. But the State University of New York (SUNY) is a yodless ˈsuːni for Americans.

    1. That's because /suni/ would be the normal representative of *suny in AmE, which drops yods after /s/.

  3. Stressed /rju/ went the way of the dodo, and at about the same time, IIRC. Of course, this did not affect /ɪw/ in accents with no through-threw merger.

  4. Are there any common words with stressed /gj/?

    The only such words I could find in LPD were all uncommon: "gules" (heraldic term for red, which I've read but never uttered), "gular" (related to the throat) and "gewgaw" (a trifle).

    Why is stressed /kj/ so much more common than stressed /gj/? I'm guessing it goes back to French, and possibly beyond that to Latin.

    1. vp: I agree that [gj] is less frequent than [kj], but not nearly as exotic as, say, [gw] compared to [kw].
      By "stressed /gj/" I assume you're after "tautosyllabic [gj]"? If so, I think that:
      1) [gj] must be tautosyllabic in singular [ˈsɪŋgjələɹ], ungulate [ˈʌŋgjəlǝt], strangulate [ˈstɹæŋgjəˌleɪt], angular [ˈæŋgjələɹ], etc. because the initial syllables of these words cannot end in [ŋg].
      2) Similarly, [gj] must be tautosyllabic in fulgurite [ˈfʌlgjǝˌɹʌjt] because the initial syllable cannot end in [lg].
      3) [gj] must be tautosyllabic in ague [ˈeɪˌgjuː], argue [ˈɑɹˌgjuː], orgulous [ˈoɹgjələs], virgule [ˈvəɹˌgjuːl], etc. because the initial syllables would be otherwise be "superheavy" syllables which are unexpected word-medially. (Compare unacceptable hypothetical *[ˈeɪgˌnuː], *[ˈɑɹgˌnuː], *[ˈoɹgnələs], *[ˈvəɹgˌnuːl].)
      4) The rhyme [ɑɹg] is shunned more generally, also arguing that [gj] is tautosyllabic in argue, arguable, argumentative, argumentation, etc.
      5) Less convincingly, for me at least, [gj] must be tautosyllabic in contiguity [ˌkɑntəˈgjuːəti], exiguity [ˌɛksəˈgjuːɪti], ambiguity [ˌæmbəˈgjuːəti], Montague [ˈmɑntəˌgjuː], etc., else I (like most North Americans) would pronounce the unstressed vowel before [g] as [ɪ] rather than schwa.
      6) Least convincingly, I simply "feel" that [gj] is also tautosyllabic in legume [lɪˈgjuːm] ~ [ˈlɛˌgjuːm], ligule [ˈlɪˌgjuːl], tegument [ˈtɛgjəmǝnt], augury [ˈɑgjəɹi], Caligula [kǝˈlɪgjəlǝ], jugular [ˈʤʌgjələɹ], regular [ˈɹɛgjələɹ], etc.

    2. I actually meant that /gj/ must precede the stressed vowel. But you have provided some good examples in "contiguity" and "ambiguity" -- thanks!

  5. What about puma? When I arrived in NZ in 2003 (from the UK) I was surprised to hear [ˈpuːmə]. Is this common in the States too?

  6. plutoman: Yes, [ˈpuːmə] among people who actually use the word. I'm a California native, and I've never heard any other pronunciation. (Although we don't really use the word here. I prefer "mountain lion," and the newspeople usually say "cougar.")

    1. I've heard puːmə fairly often in California, but usually referring to the sportswear company. I was pretty surprised by the yodlessness at first.

  7. Thanks Julie and vp. I now see there was a whole post in late Feb this year devoted to puma, tuna and similar words.