The "How Do You Pronounce X" Thread

[anteallach]

ask
literally (my daughter gets this one down to two syllables)

In particular I was wondering if any one else had /æs/ for ask in informal speech...

I don't think so, but I don't trust my self-analysis of informal speech.

Is the two-syllable literally something like [ˈlɪtʃli]?

[Vijay]

ask
literally (my daughter gets this one down to two syllables)

In particular I was wondering if any one else had /æs/ for ask in informal speech...

I think I have that before consonants.

[Travis B.]

Is the two-syllable literally something like [ˈlɪtʃli]?

My daughter pronounces it like [ˈɫɘːʁɫi(ː)].

[Salmoneus]

Is the two-syllable literally something like [ˈlɪtʃli]?

I can't speak for Travis' daughter, but I would guess it's /lɪɾli/? Although I can imagine your version too - though it sounds more English than American, I think.

The key, I guess, is whether you begin by deleting the schwa before or after the /r/. If you delete the following schwa first, you get your version, but that requires an irregular shift (tSr@ > tS). If you delete the first schwa first, you get my version through regular sound changes.

[ lIt@r@li > lIt@rli > lɪɾɚli (flapping) > lɪɾli (ɚ and the non-rhotic schwa are both regularly deleted after rhotics by some Americans - or 'Mercans', rather) ]


EDIT: I don't know how to translate from Travis into standard English, but I think that might be the same?

[Travis B.]

How my daughter arrives at [ˈɫɘːʁɫi(ː)] is probably [ˈɫɘtəːʁəːɫi(ː) > ˈɫɘɾəːʁəːɫi(ː) > ˈɫɘɾəːʁɫi(ː) > ˈɫɘəʁɫi(ː) > ˈɫɘːʁɫi(ː)] (if the preceding is confusing, note that while vowels in unstressed syllables can be long, oftentimes they are shorter than short vowels in stressed syllables).

[Nortaneous]

I think in standard notation that's [lɪːɹli].

[Travis B.]

I think in standard notation that's [lɪːɹli].

That is essentially what its equivalent would be in a GA with intervocalic flap elision present.

[quinterbeck]

How my daughter arrives at [ˈɫɘːʁɫi(ː)] is probably [ˈɫɘtəːʁəːɫi(ː) > ˈɫɘɾəːʁəːɫi(ː) > ˈɫɘɾəːʁɫi(ː) > ˈɫɘəʁɫi(ː) > ˈɫɘːʁɫi(ː)] (if the preceding is confusing, note that while vowels in unstressed syllables can be long, oftentimes they are shorter than short vowels in stressed syllables).

What accent of English do you and your family have where /r/ is [ʁ]? I've been meaning to ask for a while... I wasn't aware of [ʁ] being a variant of /r/ in any accent of English. Any other regional accents that have this?

[Kuchigakatai]

What accent of English do you and your family have where /r/ is [ʁ]? I've been meaning to ask for a while... I wasn't aware of [ʁ] being a variant of /r/ in any accent of English. Any other regional accents that have this?

I believe [ʁ] is just his notation for the bunched r... We tried to get him to use [ɹ] ambiguously for it, like most people do, for years. People used to have long discussions with him (maybe I should say "against him") regarding his notation and the reasonableness of what his choice of IPA glyphs represents, but we gave up at some point. For the past few years he's been able to freely post about his pronunciations without flamewars springing up. He does speak in a bit of an unusual way if you hear him (I think it was finlay who once told him he sounds a bit like the cookie monster from Sesame Street), but not as much as you'd think from the notation.

His notation at the moment (and for the past three years or so) has actually gotten easier to read than it used to be. In the past he used to always write his rhotic consonants with the labialization and pharyngealization modifiers, [ʁʷˤ] (in fact, he used to insist doing it in X-SAMPA, so [R_w_?\]; then he started giving Unicode IPA with the X-SAMPA at the bottom as footnotes, and now only uses Unicode IPA). He used to include a lot of length markers [ː], and in some contexts also laminal diacritics for alveolar consonants.

That said, for some years there was a Mexican girl here (usernames Canepari -> Aid’os -> tezcatlip0ca), who as I understand had grown up in the US and was a native English speaker, but was living in northern Mexico again, and she had a worse notation (worse in terms of conversationally irrelevant detail and odd notation choices) than Travis B.--even Travis recognized it as more difficult to read than his. She was a fan of Luciano Canepari's idiosyncratic notation, canIPA.

[Vijay]

Wow, okay, that explains a lot. Thanks!

[Travis B.]

I thought my /r/ was a bunched /r/ but apparently that involves pharyngealization - and I thought I pharyngealized or epiglottalized my /r/s but now it seems to me that my normal realization is a plain dorsal approximant* (not a fricative) except after a coronal, where then it has postalveolar articulation, or initially or after a rounded vowel, where then it is labialized, and my pharyngealized or epiglottalized realization is not the norm and primarily shows up in codas.

* and it has to be uvular or at least "postvelar" because I have another dorsal approximant for /l/ distinct from it which is further front than it.

[Travis B.]

Note that I still mark length a lot, and this is because in the dialect here vowel length is a very important aspect of word-recognition, because it marks which consonants are fortis and which consonants are lenis (whereas voicing is a less reliable indicator of fortisness versus lenisness). Oftentimes non-native speakers are confusing if they do not get vowel length allophony right. (In more recent times I have spent more time around non-native English-speakers, so I have gotten better at understanding speech without vowel length allophony.)

[Travis B.]

We tried to get him to use [ɹ] ambiguously for it, like most people do, for years.

My reason for resisting this was that my /r/ is nothing like [ɹ], which I actually find quite hard to pronounce in isolation. Only in recent times have I been able to train myself to enunciate [ɹ] or [ɻ] in isolation, and even then it is not natural by any means.

[Nortaneous]

True [ɹ] (i.e. a laminal alveolar approximant) is irrelevant if not fake; may as well appropriate the symbol for bunched r. Abstract phonetic transcription is sort of fake in general.

Sometimes I write ɚ̯ for English coda r, because I analyze START/SQUARE/NORTH/NEAR as parallel to PRICE/FACE/CHOICE/FLEECE, but I usually don't because it's irrelevant.

canIPA seems reasonable.

[Travis B.]

True [ɹ] (i.e. a laminal alveolar approximant) is irrelevant if not fake; may as well appropriate the symbol for bunched r. Abstract phonetic transcription is sort of fake in general.

Sometimes I write ɚ̯ for English coda r, because I analyze START/SQUARE/NORTH/NEAR as parallel to PRICE/FACE/CHOICE/FLEECE, but I usually don't because it's irrelevant.

canIPA seems reasonable.

One can make serious arguments for the diphthonginess of postvocalic /r/ in rhotic English varieties. On the other hand, for the variety here, I would specifically argue against it, because arbitrary "lax" vowels can occur before /r/ in addition to the classic rhotic "diphthongs" thanks to flap elision (for instance I audibly contrast the classic NAE /ɛr/ (which in my dialect I would analyze as /er/) with /ɛ/ (realized as [ɜ]) plus /r/), and in these cases the quality of the /r/ is identical.

[akam chinjir]

True [ɹ] (i.e. a laminal alveolar approximant) is irrelevant if not fake

Why would true [ɹ] have to be specifically laminal?

[Nortaneous]

True [ɹ] (i.e. a laminal alveolar approximant) is irrelevant if not fake

Why would true [ɹ] have to be specifically laminal?

Otherwise why not write it as ɻ? But it probably doesn't matter - do either of them exist?

[akam chinjir]

Otherwise why not write it as ɻ? But it probably doesn't matter - do either of them exist?

Yeah, I'm still wondering. There are some cases where it's given as an allophone of /ɾ/ where it seems plausible, maybe.

[anteallach]

True [ɹ] (i.e. a laminal alveolar approximant) is irrelevant if not fake

Why would true [ɹ] have to be specifically laminal?

Otherwise why not write it as ɻ? But it probably doesn't matter - do either of them exist?

[ɻ] is surely specifically retroflex as opposed to apical alveolar.

But why are you sceptical about their existence? Not every English speaker uses a bunched r.

[akam chinjir]

But why are you sceptical about their existence? Not every English speaker uses a bunched r.

For my part, I don't know enough to justify scepticism, but recently I've come across a bunch of cases where something that's described as an apical coronal approximant ends up having other stuff going on as well---retroflexion, of course, but there's also labialisation, tongue-backing, pharyngealisation; I think it tends to be something that results in accoustic "flatness." And I've been wondering whether the pattern is general. Not that I've done a lot of checking---I certainly haven't gone hunting for articulatory details on a lot of languages that are reported as having /ɹ/.

Do you have any good examples?