6

For example, the word "Emperor" in IPA on Wiktionary for General American is written /ˈɛmpɹɚ/. But that's kind of cheating because ɚ is basically /ɹ/ as far as I can tell. Yet, when you say the word, you say it in syllable form "em-pr-r", it's almost as if there is a silent or subtle glottal stop as in /ɛmpr'r/. Are /ɹ/ and /ɚ/ not really the same sound? To me as an American English speaker they sound the same. How can this word (or words like horror, mirror, etc.) be represented in IPA most accurately to account for the two-syllable "r" sound? Is it about stress somehow?

6
  • 2
    Are you asking about the phonetics or the phonology?
    – Draconis
    Apr 8 at 21:14
  • If it's an 'r' sound followed by an 'r' sound and there's no vowel between, then what is between? I'm an Australian English speaker so can't introspect. Is there any kind of hiatus, is there just a long syllabic 'r'? Apr 9 at 2:02
  • 2
    After thinking about it more I want to say there is an "oo" sound as in "book" in between, because the shape of your mouth changes slightly to that shape, but it is only a tiny bit different than the r shape. Causing a slightly change in intensity of the r. Apr 9 at 3:04
  • 1
    @hippietrail I'm American, and I find that when I say a word like "mirror" or "horror," my tongue, lips and jaw all open slightly in between the first /r/ sound and the second one, producing a little bit of an audible separation between the two sounds. Apr 9 at 14:24
  • 1
    Since American /r/ is rounded, any reduced vowel between /r/'s in mirror or emperor is going to be rounded, too. It's probably a very short centralized [ʊ]. Good observation, Lance. Have a look at Ian Catford's Practical Introduction to Phonetics, which is full of little experiments like you describe.
    – jlawler
    Apr 9 at 15:16
8

Since you tagged this "phonology" rather than "phonetics":

There are a few different ways of representing the second syllable of words like "mirror" in rhotic dialects. Some people treat it as a combination of a vowel /ə/ and a consonant; other people treat it as a syllabic resonant /ɹ̩/. (The same goes for the second syllable of "written", "little", and so on.)

If you prefer the first option, "mirror" would be something like /miɹ.ɹəɹ/, with the schwa acting as the nucleus of the second syllable. If you prefer the second option, it would be more like /miɹ.ɹɹ̩/, with the resonant itself being the nucleus.

Phonetically, if I'm pronouncing the word slowly and carefully, I do tend to have something closer to a schwa in the middle: my tongue lowers slightly, then rises again for the final [ɹ]. If I'm speaking quickly, though, there's no separation between the two [ɹ]s: the realization is more like [miɹː] (contrasting with "mere" [miɹ]). It's possible you actually have a glottal stop in the middle there (though I certainly don't), but the only way to tell would be a recording.

Your example word "emperor" is a bit trickier than "mirror" because there are a few different ways to pronounce it: some people say it with two syllables, others with three. (Personally, I tend to go with three.) I would transcribe these as /ɛm.pɹɹ̩/ versus /ɛm.pɹ̩.ɹɹ̩/—or, if you like schwas, /ɛm.pɹəɹ/, /ɛm.pəɹ.ɹəɹ/. Either way, it has the same /ɹɹ̩/ sequence as "mirror", in my dialect, with the same possible realizations (a slight tongue lowering if pronouncing it slowly and carefully, otherwise just a lengthened sonorant).

4
  • Wow I was not aware American English had a short 'i' in 'mere'. In the non-rhotic Englishes I know better 'i' can never be short before 'r' word-finally, though it can word-internally like in 'squirrel'. Also, isn't there another kind of hiatus than a glottal stop? I find it hard to imagine anyone fully closing their glottis Cockney-style in such words, rather than just a kind of relaxing/easing/centering. Apr 9 at 2:05
  • 1
    @hippietrail Yeah, that sort of relaxing is what I mean by the schwa in careful speech. And I disregarded vowel length here (since the consonant's what's important), but you're right, it should be a long i: before a voiced consonant.
    – Draconis
    Apr 9 at 2:22
  • 1
    @hippietrail For AmE, it’s very common to only describe tense/laxness and not vowel length, so seat is /sit/ while sit is /sɪt/. This is much less common outside AmE. Apr 9 at 6:57
  • @JanusBahsJacquet: Ah yes I usually only get confused when it's all in ASCII but here where Unicode IPA was liberally used it really shouldn't have stood out to me. Thanks for the reminder. Apr 9 at 10:00
4

If you look at official IPA charts (here, here), you won't find the letter /ɚ/ anymore. Esling's chart (the second of those) exemplifies the rhoticity diacritic on regular schwa (ə plus rhotic-hook, i.e. [ ə˞ ], analogous to [a˞] and so on. All that means is, "whatever the vowel is, plus a rhotic quality", which can be any kind of rhotic articulation. However, [r, ɹ, ɻ, ɼ, ʀ] are more-specific rhotic consonants. One essential difference between <ɹ> and <ə˞ > is the the former is a non-syllabic consonant. That can be remedied by writing it as <ɹ̩> (with the syllabic diacritic). However, there is a further claim implicit in writing <ə˞ >, and not as <ɜ˞>, which has to do with the purported tongue position: the latter is a rhoticized open-mid central vowel and the former is a mid central vowel.

Now that we have the "semantics" of the letters out of the way, the next question should be which is most appropriate. One criterion is phonetic. In writing <ɹ̩>, the phonetic claim is that the thing is phonetically the same as <ɹ>, except for syllabicity. That is almost reasonable in my dialect except that <ɹ> is much constricted (the tongue is higher) than supposed <ɹ̩>, so writing the rhotic vowel as <ə˞ > is more phonetically accurate, again for my dialect. I would not be surprised to hear that in rhotic UK dialects, <ɜ˞> is more accurate as a phonetic claims.

But as a phonological claim, I think that /ɹ̩/ is more accurate: it is /r/ which has been made the syllable peak, by deletion of /ə/. It is analogous to syllabic [n̩, m̩, l̩]. On the third or fourth hand, there are many dialects of GAE. In my dialect, "tear" (the verb) and "terror" are a durational minimal pair, likewise "bear" and "bearer", which can lead to some degree of hilarity when context fails (horror and whore). A phonological analysis as [hɔɹ] and [hɔɹɹ̩] is good enough, by my standards, but if I am being very phoneticky, I would write [hɔɹ] and [hɔ:ɹ]. As you should be able to see, that has major consequences for how you write the tense/lax ~ long/short distinction (seat, sit). I do know that some speakers, from Upstate New York, are or were mildly surprised at the [hɔɹ], [hɔ:ɹ] minimal pair: not everybody talks the same as me. In unstressed syllables, I suspect that /əɹəɹ/ just becomes [əɹ] and "tamper" and "tamperer" are actually pronounced the same, although I can certainly over-articulate the syllables to make a distinction.

In other words, accuracy can be judged according to a physical standard, and a cognitive standard, and they don't yield the same results. It depends on what thing you want to "accurately represent".

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.