Talk:Cot–caught merger

Overview

I moved this page off from "Phonological history of English low back vowels" upon request from one of the readers. This particular section was relatively long in respect to the other sections on the page. As the cot-caught merger can also be further discussed, I feel that the cot-caught merger deserves its own page. Thank you.LakeKayak (talk) 15:18, 12 December 2016 (UTC)

@LakeKayak:The split left an error with the references, as can be seen at the bottom of the page. Can you fix this up? Largoplazo (talk) 15:22, 12 December 2016 (UTC)

Funny enough, I am actually working on that right now. Also, by undoing my edit on "Phonological history of English low back vowels," you may have made it slightly easier for me to fix those edits. Thank you.LakeKayak (talk) 15:26, 12 December 2016 (UTC)

I didn't do anything to Phonological history of English low back vowels. Largoplazo (talk) 15:37, 12 December 2016 (UTC)

I believe that was somebody else. I apologize, sir.LakeKayak (talk) 15:40, 12 December 2016 (UTC)

I have been working on the references for about an hour. They are coming along with still some errors. I will need a day or two to finish the rest. Thank you.LakeKayak (talk) 16:17, 12 December 2016 (UTC)

Somebody help me out and informed me personally. That's done.LakeKayak (talk) 18:18, 12 December 2016 (UTC)

For example, according to PBS, "researchers identified it as a linguistic trend among young Californians some 30 years ago." See California English for details.

At one time, I added this line to the page Phonological history of English low back vowels anonymously. (The line made its way to this page in the split off of that page.) The original intent was to provide a reference explaining the origins of the merger. However, the more I think about it, the more I think that this line really has little context to the bulleted list. So, I feel that the line probably should be removed. Does anybody second this opinion? Thank you.LakeKayak (talk) 15:50, 24 December 2016 (UTC)

As it seems nobody has an opinion, I think I am going to remove it.LakeKayak (talk) 00:04, 28 December 2016 (UTC)

Conventionally represented

@W. P. Uzer: Now I understand what your issue was way back when with calling the notations /ɔ/ and /ɒ/ "conventional". I thought you meant other notations used in the IPA. It may be possible to adjust this by simply changing "conventionally represented by" to "conventionally represented in the IPA by". Over and out.LakeKayak (talk) 01:53, 26 February 2017 (UTC)

Phonetic environment?

I think it goes without saying (but should be mentioned in the article anyway) that /ɔ/ usually remains intact before /r/. But if I hear correctly, I feel many North American speakers who have the merger still use /ɔ/ before /l/ more often than not, which could be efficiently explained as "ɔ → ɑ / _ [-liquid]". But Labov et al. (2006:58–65) mainly talks about pre-velars and nasals, and Wells (1982:473–476) about on, dog, water, etc. Am I crazy to think the merger is resisted before /l/? Nardog (talk) 01:17, 1 August 2017 (UTC)

@Nardog: How many minimal pairs are there before /l/? Mr KEBAB (talk) 16:25, 1 August 2017 (UTC)

@Mr KEBAB: You mean between /ɑl/ and /ɔl/? I can't think of one, at least of native words, but I don't know how that's necessarily germane to my question either. Isn't it enough to just measure the phonetic quality of people's speech , especially if the merger is resisted before /r/? It's just that I feel I've heard /ɑ/ more in words like dog, pot, don, etc. and /ɔ/ more in words like all, tall, doll, Paul, etc. from American English-speaking people.

Or is it the case that there's an allophone of /ɑ/ that occurs before /l/ that I'm not aware of or something?

Whatever the case, I think there's a room for discussion on the merger's phonetic context in this article. Nardog (talk) 18:01, 1 August 2017 (UTC)

@Nardog: But we don't know whether this [ɔl] (or [ɒl], it's not very easy to distinguish them by ear) is actually still /ɔl/ or is it just an allophonic realization of /ɑl/. That's why I'd like to know whether there are any minimal pairs before /l/ and how widespread is /ɑ/ before /l/ at all (no pun intended). Since LPD also shows the cot-caught merged variants and doesn't allow you to search only for GA variants, it's useless in determining that. The search engine of CEPD is almost just as bad. Mr KEBAB (talk) 11:13, 3 August 2017 (UTC)

@Mr KEBAB: That's true as far as phonology goes, but say in accents where /ɔ/ is merged with /ɑ/ and it is still pronounced differently before /l/, whether the sound is an allophone of /ɑ/ or /ɔ/ is just a matter of analytical choice. But I'm asking phonetically. So I guess another way to phrase my question is: Is the merged phoneme pronounced differently before /l/? I wonder if /l/ is known to round the preceding vowel. Nardog (talk) 22:23, 3 August 2017 (UTC)

I'd speculate it's more backing than rounding, as /l/ in a coda is velarized (or even just a velar approximant). Velarization is essentially the consonantal equivalent of backness, so a vowel becomes backer before a velarized consonant. Whether there is a phoneme /ɔ/ depends on whether there's a contrast in the same phonological environment. I doubt there is; I think for instance doll and all are pronounced with the same vowel. — Eru·tuon 23:01, 3 August 2017 (UTC)

@Nardog: Yes, but if /ɑ/ before /l/ is rare or nonexistent (I'm not sure about the latter), then considering /ɔ/ to be a phoneme may be hard to justify (especially if there are no minimal pairs), because in the case of speakers with cot-caught and north-force mergers the historic /ɔ(ə)r/ is typically conveniently analyzed as /or/ (or /oʊr/ - same thing). Analyses such as /ɑl, tɑl, dɑl, pɑl/ are at this moment a bit more realistic to me than /ɔl, tɔl, dɔl, pɔl/, but I'm open to seeing evidence to the contrary. Have you checked Labov?

Sure, I can't deny that I've heard rounded /ɑ/ before /l/, because I have. This must be a thing for at least a decent amount of speakers, though details may, of course, vary.

You can hear [ɔo̯] for /ʌl/ in Estuary English, so [ɔl] for /ɑl/ in cot-caught merged speakers is perfectly plausible. I remember hearing [ˈkɔo̯tʃɐ] from a Londoner and couldn't figure out what he was talking about (hint - it was 'culture'), so different it was from his usual [ɐ]. I heard it as 'Courtshire' or whatever it was. Confusing as hell if you're not fully tuned into one's accent. Mr KEBAB (talk) 23:16, 3 August 2017 (UTC)

Final stops in "cot" and "caught"

@Wolfdog: Why on earth would we transcribe a glottally masked alveolar stop or especially an unreleased alveolar stop as a glottal stop? Is there a consensus to do that? Sol505000 (talk) 12:52, 23 February 2021 (UTC)

It's quite commonly transcribed that way in the literature: 1, 2, 3. Glottal reinforcement is also commonly transcribed as [ʔt] or [tʔ]. Do you prefer one of those? Wolfdog (talk) 16:59, 23 February 2021 (UTC)

@Wolfdog: By "it" you mean a glottaly masked alveolar stop, right? It's my understanding that [ʔt] (as opposed to [tʔ]) in English involves a released alveolar stop which we can't hear on the recordings. Since we don't know whether the stop is simply unreleased or masked by a glottal stop, it's safer to use a plain ⟨t⟩. Sol505000 (talk) 17:07, 23 February 2021 (UTC)

Yeah, you're the one calling it glottally masked. What I hear is certainly not released. How about [t̚]? Wolfdog (talk) 17:49, 23 February 2021 (UTC)

I see nothing wrong with ⟨t⟩ as it's the phonemic notation and both glottal reinforcement and no audible release are common in this environment and are not the focus in the context. In any case I'd use ⟨ʔ͡t⟩, ⟨ˀt⟩, or ⟨t̚⟩ before ⟨ʔ⟩. Nardog (talk) 11:34, 24 February 2021 (UTC)

⟨t̚⟩ works because the diacritic is defined as 'no audible release'. ⟨t⟩ is even better because, as Nardog says, the exact phonetic realization of /t/ is not the focus in the context. Plus, we don't use the 'no audible release' diacritic in other contexts (apt, doctor, knock down), do we? Sol505000 (talk) 13:37, 24 February 2021 (UTC)

Well, are we making a decision for just this page or for all American English pages? I'm happy to make the ad hoc change here; otherwise, we can bring this discussion to Talk:American English. Thanks for the input. Wolfdog (talk) 00:07, 25 February 2021 (UTC)

⟨ʔ⟩ isn't a valid transcription of an unreleased alveolar stop without glottalization so yeah, that should be changed. With that being said, I prefer ⟨t⟩ for the unreleased/glottally masked alveolar stop. The unreleased nature of the stops in apt and doctor is never shown in IPA transcriptions of English. It isn't something that's unique to English (it happens in German and Dutch too) and is shared among all major accents anyway. The lack of it (as in Wales and South Africa) is worth mentioning in prose. There's level of reasonable narrowness of transcription before it becomes too hard to read. I'm mentioning the stops in apt and doctor because they're realized with no audible release with much greater consistency than the final stops in cot and caught which heavily depend on the following sound (or the lack of it). For instance, in the phrase not partially the unreleased /t/ can be realized as an unreleased bilabial plosive (producing an Italian-style geminated /p/), which would be interpreted as an allophone of /p/ in other contexts (the same thing can happen in German and Dutch). Though even here ⟨t̚⟩ is an appropriate transcription because the place of articulation varies between bilabial and simultaneous bilabial and alveolar (and it's unreleased in both cases). But when it's flapped, ⟨t̚⟩ is anything but appropriate. Sol505000 (talk) 03:55, 25 February 2021 (UTC)

I think I'd personally prefer ⟨t̚⟩ to represent final /t/ on American English pages, since it reminds readers (or suggests to newcomers) that there is almost always a distinction between this /t/ versus the flap or aspirated variants which we also currently represent on such pages. I don't think this would lead to any kind of frightening slippery slope. Wolfdog (talk) 19:49, 1 March 2021 (UTC)

New Developments since the 1990's?

The article makes the following uncited claim: 'Some speakers in all three regions, particularly younger ones, are beginning to exhibit the merger despite the fact that each region's phonetics should theoretically block it.' This requires clarification: Is it talking about people who have a weak or weakening regional accent in general? Additionally, it seems to me as someone who is native to and lived all my life in the United States in the same area that the contrast is being preserved by younger speakers in other ways: for example, I hear the /ɒ/ sound much more frequently in pronunciation of written o as in 'lot' (and of words spelled with wa- like "water"), and I also hear /ɔ/ more often in words written with the digraph au (though before r it may sound closer to rhoticized /ɒ/), whereas I used to almost always hear /ɑ/ from people both older and my own age for these same words, as if in the last decade or so people in my area who were undergoing the merger have started shifting their pronunciation again to maintain the distinction. If this observation is correct, then the situation may have entered a new stage in which, for areas that were undergoing or had already undergone the merger, speech with the merger and speech without the merger are "averaging out" their differences to reach some kind of new equilibrium, which makes sense as /ɒ/ has previously not existed in many places in the U.S.A. at least. This change in speech patterns may also be due to increased contact between speakers of General American English and those of other accents by means of technology. I also hear the sound of "medium-long" a in "car" or "María", traditionally pronounced /ɑ/, being advanced a bit to almost but not quite /ä/, which I attribute to both resistance to the merger and the influence of Spanish. These evidences are anecdotal on my part as I am not involved in any studies, but in any case I think the claim being made by the article is too strong without a citation, and perhaps also out of date with regard to new developments. Webspidrman (talk) 17:38, 27 December 2021 (UTC)