Thursday, 21 January 2010

everyday (?) IPA

Philip Swan, an SIL-trained Australian phonetician, has put up a nice video on YouTube. It consists of three parts: using the IPA to transcribe what a toddler says, using phonetic description for clarity in identifying letters of the alphabet over the phone, and transcribing paralinguistic interjections.
I became aware of the existence of this video from postings on the email list of VASTA, the (American) Voice and Speech Trainers Assn. Inc. Some of the commentators there were sceptical about the accuracy of the transcriptions.
I think that writing the Australian FACE vowel as [æi] is not unreasonable. I have to say, though, that superscripting the [i] symbol to show the less prominent part of a diphthong, as Philip Swan does in his [æi], is not an IPA-approved notation. (If you want to explicitly show “non-syllabic” in IPA, the correct diacritic is meant to be U+032F, a COMBINING INVERTED BREVE BELOW, thus [æi̯].) I am less than enthusiastic, though, about transcribing the Australian PRICE vowel as [ʌɪ], or as Swan has it [ʌɪ]. I would have thought that the first element was usually nearer to fully open cardinal [ɑ] than to open-mid cardinal [ʌ], let alone the Australian STRUT vowel. (Is this the baleful influence of Clive Upton’s unfortunate choice of symbol for the PRICE vowel in the OED/COD?)
I have no idea what the toddler is saying, or trying to say. Has anyone?
To disambiguate letter names in noisy conditions, articulatory descriptions would only work if both parties are well trained in phonetic terminology — which they normally aren’t. That’s why we use keywords: “S as in Sugar, not F as in Freddie”, or “sierra whisky alfa nectar November”.

19 comments:

  1. Making the rounds on Facebook recently:

    India Foxtrot Yankee Oscar Uniform Charlie Alpha November Uniform November Delta Echo Romeo Sierra Tango Alpha November Delta Tango Hotel India Sierra, Charlie Oscar Papa Yankee Alpha November Delta Papa Alpha Sierra Tango Echo Tango Oscar Yankee Oscar Uniform Romeo Sierra Tango Alpha Tango Uniform Sierra

    I "got" most of the Australian IPA choices, though not all. I have the same trouble with [ʌɪ] in PRICE as I hear it in video. Ignorant of Clive Upton, so wondering if ʌ chosen because that symbol is shifted to place vowel quadrilateral where IPA has ɐ. (I've seen this in any number of vowel quads in books and on websites.) Is that the issue, or is something else going on?

    Off to research Clive Upton ...

    ReplyDelete
  2. Forgot to say, I have no idea what baby is trying to say. I think the little one (adorable!) is still in babbling stage.

    ReplyDelete
  3. Re toddler: I have a hunch potatoes are involved...

    ReplyDelete
  4. I have a hunch potatoes are involved...

    Me too. Wasn't cake in there as well? But this kind of thing is quite idiosyncratic, isn't it? (And just serves to show just how strongly speech comprehension is dependent on context. If we knew what the toddler had had for dinner, we would probably be able to make those words out. Excuse my conditionals.)

    With respect to /ʌɪ/: I think Upton's decision is more defendable than its use for Australian E. I imagine the line of thinking is this: (1) PRICE and MOUTH should not be shown with the same starting point. (I agree.) (2) For MOUTH, there's not much choice apart from /a/, which Upton et al. use for TRAP. (Fair enough.) (3) For PRICE, you could go for /ɑ/, as John suggests. But if you look at acoustic vowel charts in Cruttenden (2008) or Hawkins and Midgley (2005), there isn't too much difference between /ʌ/ and /ɑ/ (both should really be shown as retracted ɐ), and if anything is more open, then it's /ʌ/.

    This, I think, doesn't work for Australian English. The sources I have access to all show PRICE as starting far back and START/STRUT as central. (And for me, this is accurate impressionistically; START/STRUT could even be further front.)

    ReplyDelete
  5. WOW!
    What an amazing post. I live in a land where sarcasm does not exist, so I do mean it.

    Working backwards through the blog:

    "That’s why we use keywords: “S as in Sugar, not F as in Freddie”, or “sierra whisky alfa nectar”.”

    NATO agreed in 1956 that a set of words would be the norm for telephonic communication and 'nectar' was not one of them. Alfa is not in my dictionary, and it should be ‘Sierra, whisky, alpha, November”, but if this is a typo I do apologize because they are easy to make as I am sure I will prove. I will not get into the spelling of whiskey!

    I too have no idea what the toddler is saying, but I do remember understanding my own toddlers quite acceptably when they were that age.

    I should hate to be Clive Upton, but I am sure he has a thicker skin than I.

    I like the transcription, or realization of /æi/ with the breve below, to show the Australian accent, but fail to understand why the ligature is used in BrE, when it is clearly the suggested diphthong of AmE. Why do we not use /a/ for
    RP, /æ /for AmE. amd /æi/ for AusE.?

    I would also hate to be Philip Swan, but am really happy to have watched his video. It is funny, amusing, and entertaining.
    I notice he has trouble with letter .
    N as far as I know has four sounds.
    1. The sound of CAPITAL N as in ‘enjoyʲ, and ‘when’.
    2. ‘n’ as in ‘no’, and ‘not’.
    3. ‘nn’ as in ‘bacon’, and London.
    4. ‘ŋ’ as in ‘think.

    I agree with the wisdom of the post, in that he has exceeded the purpose of his role. I also do not know why he says /leɪbiʊl/ instead of /labiʊl/. Is this the latest trend?
    (I use ʊ instead of (ə) and /a/ instead of /æ/ deliberately).
    I enjoyed the post and the video, and hope my typeface is displayed correctly.

    ReplyDelete
  6. Sorry: I should have given references for Upton. I wrote about his transcription at http://www.phon.ucl.ac.uk/home/wells/ipa-english-uni.htm (near the end). Jack Windsor Lewis wrote with greater detail and forcefulness at http://www.yek.me.uk/ipadicts.html.

    ReplyDelete
  7. Aren't PALM/START and STRUT distinguished by length in AuE, there being no reliable qualitative difference between the two?

    ReplyDelete
  8. According to Wikipedia's "NATO phonetic alphabet" article, the spellings "Alfa" and "Juliett" are used by ICAO, ITU, IMO, FAA (all international organisations) and "Alpha", "Juliet" by ANSI (a US organisation).
    "In most versions of the alphabet, the non-English spellings Alfa and Juliett are found. Alfa is spelled with an f as it is in most European languages. The English and French spelling alpha would not be properly pronounced by speakers of some other languages whose native speakers may not know that ph should be pronounced as f. Juliett is spelled with a tt for native French speakers because they may otherwise treat a single final t as silent."

    ReplyDelete
  9. @James Mutter:
    I think /ˈleɪbiəl/ is actually more usual across the dialects (see the OED and M-W). I was more struck by his use of /ælviːˈoʊlɜr/ instead of /ælˈviːəlɜr/.

    As for understanding the toddler, I'm as clueless as the rest of you, but perhaps there is some "butter-peanut" involved?

    ReplyDelete
  10. At one time I thought ælviˈəʊlə was illiterate. The Classical quantities fallacy, you know. I soon observed that trained (where?) linguists and even phoneticians used it. I still don’t like it, but I see Collins now lists it after ælˈvɪələ, and I've found some online dictionaries that do so as well, but in a chaotic way which is not consistent with their treatment of alveolus.

    Even before I saw de cuup's post above, I was fairly sure the toddler is having a succession of stabs at "potatoes". An awe-inspiring reminder of what it takes, I thought.

    ReplyDelete
  11. I've always assumed that Upton's PRICE transcription was based on an [ɐɪ] or similar realisation, with the choice of /ʌɪ/ reflecting the closeness of the first element to the STRUT vowel in some accents.

    I think [ɐɪ] is a reasonable transcription of my own PRICE - the starting point does not feel as open as the [a] of TRAP/BATH - but in my case it's not particularly close to STRUT.

    ReplyDelete
  12. @army1987: Yes, that's one interpretation. You may want to have a look at Felicity Cox's standard description of AusE, p. 345 (warning -- PDF link):

    Illustrations of the IPA: Australian English

    BTW, if you look at the vowel chart on p. 348, you can sort of understand where the /ʌɪ/ represntation may come from: The starting point of PRICE is quite back and less open than START/STRUT. Under a language-independent strict-IPA interpretation, that may be interpreted as [ʌ]. But of course that's mixing apples and oranges, at least a bit. The correspondence between acoustic charts like this and idealised charts as found in e.g. dictionaries or publications with an expressly phonological focus is not perfect.

    Cox herself uses /ɑe/...

    (The Capcha on this site is incredible. Asks to input "fonets".)

    ReplyDelete
  13. @mallamb: I still don’t like it, but I see Collins now lists it after ælˈvɪələ, and I've found some online dictionaries that do so as well

    LPD gives it as the first option for BrE ;)

    ReplyDelete
  14. Australian FACE can even be [ai] I believe.

    ReplyDelete
  15. Nice bit of fun, but my [s] is grooved as well as my [ʃ]... (Of course the IPA chart doesn't distinguish between grooved and slit fricatives anyway)

    ReplyDelete
  16. I am an Australian speaker who uses [æi] for the FACE diphthong, but I want to mention the fact that a more British-esque [ei] is now very common among young Australian women. I hear it often, most recently when I was being served at the local KFC yesterday. (I'm in Adelaide, for what that's worth.)

    ReplyDelete
  17. @John: Belated thanks for the links re Upton. I had, in fact not only read but saved your article to my hard drive, but I had forgotten the details. Both links were very helpful.

    ReplyDelete
  18. Adrian: a more British-esque [ei] is now very common among young Australian women.

    Could this be related to the lowering of TRAP Cox reports in her work? E.g. here? (Warning: PDF link.)

    ReplyDelete
  19. wjarek: I don't know, but bear in mind that Cox's data is mostly from Sydney/NSW, whereas my observations are from Adelaide/SA. The geographical distance is not trivial. In any case, thanks for the PDF link.

    BTW, in case you're interested I've just uploaded a recording of me saying the vowels mentioned in the PDF - heard who'd hoard hard heed hood hod hud had head hid in that order.

    ReplyDelete