The phonetics of a#n vs. an# juncture

« previous post | next post »

Geoff Pullum’s recent post “An aim or a name?” stimulated a surprisingly lively discussion of juncture in English. This morning, I thought I’d encourage this interest in phonetics by posting a random sample of relevant real-world examples.

The distinction between “a name” and “an aim” turned out to hard to find — the 25 million words of conversational (telephone) speech that I searched had plenty of instances of “a name”, but only one instance of “an aim”. So I picked a similar case where both sides of the opposition are represented by dozens of tokens: “a nice” vs. “an ice”, in contexts like “a nice guy” or “a nice one”, vs. “an ice storm” or “an ice cream sundae”.

Here are four random selections from this little collection — see if you can identify them:

1. [audio: http://languagelog.ldc.upenn.edu/myl/Z05543.mp3]
2. [audio: http://languagelog.ldc.upenn.edu/myl/Z00789.mp3]
3. [audio: http://languagelog.ldc.upenn.edu/myl/Z01530.mp3]
4. [audio: http://languagelog.ldc.upenn.edu/myl/Z03368.mp3]


Note that I’m not guaranteeing balance here. Perhaps all four are “an ice”, or all four are “a nice”; or any other combination of categories.

I’ll give you the answer key, with some discussion, tomorrow.

Meanwhile, Bob Ladd’s comment on yesterday’s post (describing findings from Ladd and Schepman, “‘Sagging transitions’ between high pitch accents in English: experimental evidence“, J. Phonetics 31(1) 2003), characterizes the true situation fairly clearly:

The biggest acoustic difference we found was […] that word-final consonants are shorter, on average, than word-initial consonants (in our data, 47 msec vs. 67 msec in the case of /n/). […]There were glottal stops (at the beginning of e.g. Eason) in only 10-15% of cases. In a subsequent perception experiment, we found that people, given a forced choice, could correctly identify the name intended by the speaker only about two-thirds of the time.

The thing that Bob left out is that the within-category variation (e.g. the variation in duration of word-initial [n]) was no doubt of roughly the same magnitude as the average difference between the categories. (The cited paper, alas, doesn’t give variance measures, but only ANOVA F values — a pet peeve of mine…)

Also, although Ladd and Schepman were careful to avoid the obvious effects of facultative disambiguation,  they were measuring “laboratory speech”, i.e. the relatively careful and formal style of speech that you get when you bring people into the lab and have them read (somewhat contrived) passages.

This is necessary in order to get the orthogonal design required for ANOVA and similar statistical analyses, and therefore we phoneticians have been using such techniques for the past century or so; but this approach tends to produce somewhat stereotyped version of the differences under study, human nature being as it is, and to produce results that may be artificial in other ways as well.  This practice is starting to change,  as we learn to approach similar questions using data from collections of thousands of hours of natural speech.



47 Comments

  1. Peter Taylor said,

    May 8, 2010 @ 7:28 am

    I hear “a nice”, unsure but inclined towards “an ice”, “a knife” (although the word boundary sounds more like “an ife”), and “a nice”.

  2. Craig said,

    May 8, 2010 @ 7:34 am

    I live in central maryland, if that makes any difference. I pronounce “a” and “an” with a different “a”. For “an”, my tounge is pulled back to make the sound deeper in the throat (boy, am I a true linguist or not?). My “a” is much closer to the palate but at the back. There is not much difference when casually speaking, but it is certainly discernable.
    I use that difference to guess that the middle two sounds were “an”.

  3. Christian DiCanio said,

    May 8, 2010 @ 7:38 am

    #1, 2, 4 are “an ice” and #3 is “a nice.” I may be totally wrong, but this is my guess. #3 was articulated faster than the others, so this may be biasing my judgement.

  4. Jarek Weckwerth said,

    May 8, 2010 @ 7:38 am

    A “non-native” speaker, I hear a nice in the first two and an ice in the other two. But I’m working on the assumption that the clips come from the kinds of phrases you mentioned; and I would expect an ice storm to be quite different in terms of the rhythm and stress (and, consequently, tonicity) from a nice guy

  5. notrequired said,

    May 8, 2010 @ 7:58 am

    Here’s my two cents:
    1. nice
    2. ice
    3. nice
    4. nice

  6. jih said,

    May 8, 2010 @ 8:17 am

    Any differences in duration would be a by-product of differences in articulation. In “nice”, consonant and diphthong are always planned together, coarticulated. In “an ice” coarticulation may or may not happen. If the /n/ is coarticulated with the following /ai/, then you are not going to hear a difference, I guess.

  7. Frans said,

    May 8, 2010 @ 8:17 am

    I’m going with 1 and 4 “a nice”, and 2 and 3 “an ice”.

  8. Aaron Toivo said,

    May 8, 2010 @ 8:25 am

    Best guess:

    1. nice
    2. ice
    3. ice
    4. nice

  9. Bob said,

    May 8, 2010 @ 8:53 am

    I’m pretty confident #1 is “a nice”. The other three could go either way, I can’t tell.

  10. jih said,

    May 8, 2010 @ 8:59 am

    (second part)
    It follows that the most appropriate test is not forced choice between the two meanings, but, rather, “Could this be “a nice”?” Positive answers would be expected for all tokens where the speaker intended “a nice” and ALSO for those tokens where the speaker intended “an ice”, but coarticulated across word boundaries.
    It also follows that there should be much more variability in the duration of the word-final /n/ in “an ice” than in the duration of this segment when word initial in “a nice”.

  11. Adrian Bailey (UK) said,

    May 8, 2010 @ 9:35 am

    a nice, a nice, an ice, an ice

  12. Mr Fnortner said,

    May 8, 2010 @ 9:41 am

    How can we tell if our answers differ from chance if we miss one answer (or even if we get all four correct)?

    My answers:
    1. a nice
    2. an ice
    3. an ice
    4. a nice

  13. J.H. said,

    May 8, 2010 @ 9:43 am

    1. “a nice”
    2. “an ice”
    3. “a nice”
    4. “a nice”

  14. Dhananjay said,

    May 8, 2010 @ 10:14 am

    1. a#nice 2. an#ice 3. an#ice 4. a#nice

  15. Chris said,

    May 8, 2010 @ 10:14 am

    nice
    ice
    ice
    nice

  16. Russell said,

    May 8, 2010 @ 10:17 am

    Here are my responses:

    1. First I thought it was “nice,” but then I thought “ice.” Or maybe it was the other way around.
    2-4. See 1.

  17. James said,

    May 8, 2010 @ 10:25 am

    nice
    ice
    ice
    nice

    This answer seems to have a plurality. I couldn’t give any reason for mine whatsoever — just how I perceived them.

  18. D. Sky Onosson said,

    May 8, 2010 @ 11:14 am

    My guesses:

    1.a nice 2.an ice 3.a nice 4.an ice

    However, I doubt that it would be very easy to pick out things like this without hearing more of each individual’s speech. I don’t mean contextual clues as to what the words are, but more examples of each person’s idiolect.

    @ jih: You might be right about that, but we can’t hear articulation, only its acoustic consequences.

  19. Christopher said,

    May 8, 2010 @ 11:16 am

    I’m with J.H. – I heard “nice” on all except 2.

  20. Lee Morgan said,

    May 8, 2010 @ 11:38 am

    I heard:

    1. Nice
    2. Ice
    3. Ice
    4. Ice

  21. unekdoud said,

    May 8, 2010 @ 12:12 pm

    Currently: 2 and 4 nice, 1 and 3 ice.
    I’ll probably go drink a nice cup of coffee with an ice cube in it, then come back and try again.
    (or is it an iced cup of coffee with a nice cube in it?)

  22. Marcus Merritt said,

    May 8, 2010 @ 12:51 pm

    I heard:

    1. a nice
    2. an ice
    3. an ice
    4. a nice

  23. Michael W said,

    May 8, 2010 @ 12:52 pm

    Like Jarek Weckworth I’d expect a big difference between “an ice” on its own and something like “an ice cream [cone]” or “an ice mass” or even “an icicle”. We really ought to get some samples of “anisotropic” in there as well to compare.

    I think most people would have much less trouble at identification if this were accompanied by other speech from the speaker. It’d be nice to see how that affects the ability to decide.

    I’ll give my guesses anyway. I think at least subconsciously I’m judging these as standalone phrases, which is most likely incorrect:
    1. nice (this one I have the toughest time with)
    2. ice
    3. nice
    4. ice

  24. PJ Funnybunny said,

    May 8, 2010 @ 2:27 pm

    nice, ice, ice, ice

  25. Ast A. Moore said,

    May 8, 2010 @ 3:22 pm

    Number three is “an ice”; the others are “a nice.”

  26. Eric said,

    May 8, 2010 @ 3:28 pm

    I hear a nice, an ice, a nice, a knife.

  27. Army1987 said,

    May 8, 2010 @ 3:29 pm

    I’m not a native speaker, but…
    If I disregard the height and frontness of the initial vowel, I hear “an ice, a nice, an ice, a nice”, but the vowel in the last doesn’t quite sound like a typical way of pronouncing “a”, so I guess it’s “an ice”. I agree with D. Sky Onosson about providing other samples of the speakers’ speech.
    It’d be nice: are you sure it wouldn’t bean ice? :-)

    (I think other junctions such as /t#S/ “cat shit” vs. /tS#/ “catch it”, /k#w/ vs. /#kw/ (can’t find a minimal pair right now), /tr/ vs /t#r/ (“nitrate, night rate”), in some accents /t#/ vs /#t/ (“great ape”, “grey tape”) would be easier to tell apart than /n#/ vs /#n/.)

  28. John Cowan said,

    May 8, 2010 @ 4:33 pm

    Boy, this is a blast from the past, although when I was a yoot it was nitrate/night rate/Nye trait, and I see there are still plenty of hits for this. Anyhow, I make it nice/ice/ice/nice, like many others above.

  29. mp said,

    May 8, 2010 @ 4:53 pm

    To me, all four sounded closer to “an ice,” but the second one more so than the others.

  30. Ice (CA) said,

    May 8, 2010 @ 8:06 pm

    1. a nice
    2. an ice
    3. a nice
    4. an ice

    1 and 4 are clearly differentiated for me as a nice / an ice.
    #2 and 3 are hit or miss for me without context, as well as sounding more like disordered speech than a natural utterance.

  31. Varun said,

    May 8, 2010 @ 8:12 pm

    Is this not a classic case of speech recognition??
    I guess it is, and so in order to make some meaningful statement we should also get access to other voice samples.
    Maybe then after finding some interesting features (time between words etc.) we can be in a better position to tell which sample stands for what…

    Anyways my current guess are :) :)
    1) a nice
    2) an ice
    3) a ice
    4) an ice (100% sure)
    — Varun

  32. Scriptor Ignotior said,

    May 8, 2010 @ 9:09 pm

    Almost mere guessing, and involuntarily considering vowel length, intonation, and stereotypes about who would be talking about what:

    1. an ice
    2. a nice
    3. a nice
    4. a nice

    I also involuntarily “corrected” for those stereotypical assumptions. [Bejasus these previews are unruly. No matter what I do, the line beginning with “4” gets spaced apart from the other three.]

  33. Maria said,

    May 8, 2010 @ 9:09 pm

    Non-native speaker here. I heard nice, ice, ice, nice.

  34. Ben said,

    May 8, 2010 @ 11:40 pm

    Native speaker here. At a first listen, I had a lot of trouble deciding based on pure phonetics. But then for each of the four clips I played it back and then immediately after, in my imagination, added a context (“…cream”, “…guy”, etc…) and compared. By contrasting these imaginary contexts, it became pretty clear to me that the answers are:

    1. nice
    2. ice
    3. ice
    4. nice

    And this still seems to the plurality for the answers here, so I’m really curious if it’s correct. And for the other people who got this particular set of answers I wonder if they employed a similar strategy as I (appending imaginary contexts and comparing) or if they found this answer set another way. If this particular set of answers was arrived at in multiple different ways, I’d be more inclined to believe it’s the most likely.

  35. Dave said,

    May 8, 2010 @ 11:41 pm

    I feel as though all four could be wrong, but here’s my best guess:

    #1 a nice
    #2 an ice
    #3 an ice
    #4 a nice

  36. Amanda said,

    May 9, 2010 @ 12:41 am

    1. nice
    2. ice
    3. ice
    4. nice

  37. Scriptor Ignotior said,

    May 9, 2010 @ 1:03 am

    Army1987:

    It’d be nice: are you sure it wouldn’t bean ice?

    O, good! It’d be nice is at least quasi-homophonic with It’d been ice (=It had been ice):

    “Last season the pond didn’t freeze at all; but every other winter it’d been ice for several weeks.”

  38. Ryan Denzer-King said,

    May 9, 2010 @ 1:45 am

    FWIW:
    1. nice
    2. ice
    3. nice
    4. nice

    Impressionistically, I feel like the /n/ in #2 is longer, and that this is what leads me to believe it’s “an ice”. I’ll be interested to see the answers tomorrow.

  39. Nijma said,

    May 9, 2010 @ 1:46 am

    1. a (?)
    2. an
    3. a
    4. an
    …going by the sound of the vowels in the articles. (I’m from the U.S. Great Lakes area).

  40. Nijma said,

    May 9, 2010 @ 2:00 am

    It would be a nice touch to know where the speakers were from.

  41. BW said,

    May 9, 2010 @ 2:31 am

    I heard

    1 nice
    2 ice
    3 ice (at first, but now I’m not sure anymore)
    4 nice

  42. Scriptor Ignotior said,

    May 9, 2010 @ 4:18 am

    Nijma:

    It would be a nice touch to know where the speakers were from.

    O yes, it would! ☺ :)

  43. Sili said,

    May 9, 2010 @ 4:42 am

    1) a nice
    2) an ice
    3) a knife
    4) an ‘ife’

  44. Adouma said,

    May 9, 2010 @ 6:27 am

    I think before we even get the key, the comments have proved that most people can’t reliably hear the difference. GKP was right all along.

  45. jim said,

    May 9, 2010 @ 6:43 pm

    Phrasal homophones aren’t that common in English, I think (now someone will come up with a lengthy list). They’re all over the place in French, though. Listen to Barbara singing “Sid’amour A Mort” for example.

    We distinguish them, like we distinguish single word homophones, by context. “I’ll paint the wall anice cream” vs. “It’s hot; I’ll buy anice cream.”

  46. Peter said,

    May 10, 2010 @ 2:18 am

    /k#w/ vs. /#kw/ (can’t find a minimal pair right now)

    A bit artificial, but “fork whizz” / “four quiz”

  47. Faldone said,

    May 10, 2010 @ 12:11 pm

    I first heard:

    1 a nice

    2 an ice

    3 a nice*

    4 a nice

    *But maybe “an ice”

    After reading a few comments 3 and 4 became “a knife”. Further listening managed to hear 3 as “an ife” as in “she showed be an Eiffel Tower picture she had taken …”

RSS feed for comments on this post