Talk:General American English

Rhotic ɜ instead of ə? ɑ - ä?

In the monophtong diagram, why is there rhotic ɜ where I would expect ə to be? Also, ɑ looks like in the positon for where ä would be. Zbutie3.14 (talk) 04:21, 1 August 2024 (UTC)

⟨ɜ⟩ used to be defined as a "variety of [ə]" by the IPA, without a defined height. This is how it's used in most transcriptions of English. And ⟨ɑ⟩ is just as good a transcription of an open central vowel as ⟨a⟩ or ⟨ɐ⟩, with ⟨ɐ⟩ being arguably the best if you really don't want to use any diacritics. But we use ⟨ɑ⟩ because that's what most sources use (some actually write it ⟨a⟩) and because it varies between back and central for speakers without the cot-caught merger. For the cot-caught-merged majority, the merged phone is typically back, so ⟨ɑ⟩ is the most appropriate symbol for that. Sol505000 (talk) 18:13, 27 November 2024 (UTC)

SQUARE - really /ɛr/?

Shouldn't SQUARE be phonemicized as /ær/? /æ/ before /r/ sounds exactly like pre-nasal TRAP (or very close anyway), except that it's not nasalized (unless idiolectally or regionally, for speakers with a nasal voice quality). I'd write /skwær, ðær, ʃær/ just like we write /mæn, bæn, kæn/. Sol505000 (talk) 16:12, 27 November 2024 (UTC)

Except we go by what credible sources do and there's a long tradition of not employing /æ/ for this vowel in US transcriptions, as with Wells' work, Kortmann's Handbook of Varieties of English, virtually all dictionaries, Labov, etc. Notations like ɛər or ɛr or eər or er are the norm. Wolfdog (talk) 17:16, 27 November 2024 (UTC)

Then this seems to be an issue of sources assuming that the established transcriptional conventions are accurate without investigating the issue themselves, or keeping up to date with others' research (e.g. about /æ/ raising to /eɪ/ in 'sang', 'rang' etc.). An anti-scientific approach, but sadly it's still (somewhat) common in linguistics. If the merger affects all three of /eɪr, ɛr, ær/, surely it makes the least sense to choose ⟨ɛr⟩ (or a non-phonemic symbol such as ⟨eər⟩ or ⟨ɛər⟩) for this sequence. ⟨er⟩ would be fine if we used ⟨e⟩ for FACE, but ⟨ær⟩ is definitely the best from a logical standpoint. /æ/ would then be analyzed as undergoing raising to [ɛə ~ eə] before all sonorants except /l/ (and /ŋ/, where there's no centering glide and it can be safely transcribed /eɪŋ/). In dialects with the TRAP split, ⟨eər⟩ would be even better because you have the /eə/ phoneme.

Aren't there really ANY sources that use /ær/, though? /ɪr, ʊr/ already vary between that and ⟨ir, ur⟩, depending on the source. Sol505000 (talk) 18:18, 27 November 2024 (UTC)

[ɚ] versus [əɹ]

Perhaps Nardog or another editor will know --- I skimmed the Archives but couldn't quickly find such a discussion --- is there any reason why in our vowel chart (and thus other transcriptions on the page) we give NURSE/LETTER as [ɚ] but then for, say, NEAR as [iəɹ] and SQUARE as [ɛəɹ], rather than [iɚ] and [ɛɚ]? Can't remember if we've already talked through this choice. Thanks. Wolfdog (talk) 11:44, 11 March 2025 (UTC)

⟨ɚ⟩ essentially means [ɹ̩] or [ɻ̍] (see e.g. Handbook of the IPA, p. 25), which means [ɚ̯] and [ɹ] or [ɻ] are equivalent. [əɹ] represents a sequence of two sounds. Which is how NEAR and SQUARE manifest in GA in pausa (before a consonant or pause). Nardog (talk) 11:49, 11 March 2025 (UTC)

Ah right, interesting. I was recently reading "Amount of rhoticity in schwar and in vowel+/r/ in American English" (Kuecker et al 2015) that measuring the durations of tokens of NURSE, LETTER, and other vowels+R (START e.g.), concludes (tentatively) that can all be reasonably analyzed as a vowel followed by a separate /r/, though the more precise findings were that "Overall means for all participants show rhoticity of 58% for all vowel+/r/ tokens [START], of 76% for unstressed schwar [LETTER], and 94% for stressed schwar [NURSE]". Intriguing results. Wolfdog (talk) 15:10, 11 March 2025 (UTC)

[ɨ] and [ə]

@Nardog: Flemming & Johnson say it's commonly noted "that some accents of American English have two contrasting reduced vowels" [ɨ] and [ə], but also "it remains unclear whether the distinction between barred-i and schwa is a basic phonemic distinction in the relevant accents of English, or whether it is limited to a restricted environment exemplified by pairs like roses–Rosa’s". I don't believe they end up making any definitive claim about the phonemic status of [ɨ]. Therefore, shouldn't we be cautious about definitively considering [ɨ] an allophone of the phoneme /ə/? (I'm trying to look into more scholarly lit about the weak vowel merger in AmE.) Wolfdog (talk) 11:20, 17 May 2025 (UTC)

They do.

So the minimal contrast between barred-i and schwa in pairs like roses–Rosa's can only arise because of the difference in morphological structure between the two words: in roses the stem boundary precedes the reduced vowel, [[ɹoʊz]əz], while in Rosa's it follows it, [[ɹoʊzə]z]. This contrast does not arise in other contexts, such as within monomorphemic words. (p.94)

Wells (1982: 167f.) argues that RP and other English dialects distinguish unstressed [ɪ] and [ə] in pairs such as Lenin [lɛnɪn] vs. Lennon [lɛnən], and rabbit [ɹæbɪt] vs. abbot [æbət] (Wells's transcriptions). This distinction is not made in the American accents that we are familiar with, so Lenin and Lennon are homophonous, for example. (p. 95)

Your quote is from the introduction section, where the authors are simply laying out the very questions into which they're about to discuss the results of their investigation. Nardog (talk) 13:18, 17 May 2025 (UTC)

Right, and your F&J quote is good evidence for the possibility of it being a single phoneme, but still not something they explicitly assert. I'm just trying to be careful. I feel that the weak vowel merger indeed exists (per Wells) but I wonder if there's also a re-phonemicization involving schwa and KIT. (Is minus taking on the KIT vowel, e.g., rhyming with miss?) Of course reduced-vowel environments inherently make this difficult to easily know, and I don't have any scholarly source. Again, my concern just came from a sense of caution. Wolfdog (talk) 17:00, 17 May 2025 (UTC)

What if, in the Trager note, we simply say the choice to use the symbol ⟨ɨ⟩ dates back to a tradition starting in the 1950s from linguist George L. Trager and others? Wolfdog (talk) 18:14, 17 May 2025 (UTC)

Sure, that at least does not suggest it is a separate phoneme (which I found "a reduced vowel that is..." did). Nardog (talk) 00:51, 18 May 2025 (UTC)