Help talk:Pronunciation respelling key

Mattress

Based on syllabication (which I know is itself a continually debated topic), how would we use this key to transcribe "mattress"? After reading Note 1 on the page, about checked vowels, I would assume it's MA-triss. Is that correct? Kbb2 offers something like MATR-iss (or MATR-əss in certain accents), but that looks very odd to me! Wolfdog (talk) 18:27, 7 December 2019 (UTC)

MAT-riss is compliant with Note 1, because [t] is not aspirated in the cluster [tr]. It's not 100% correct syllabication, but it's acceptable because it keeps the vowel checked. MATR-iss is just completely wrong in syllabication - [tr] can't end a syllable in English. Indefatigable (talk) 21:32, 8 December 2019 (UTC)

Kbb2 is adopting Wells's scheme, which is "heretical" in his own admission. Even though /tr/ is affricated, I see nothing wrong with MAT-riss as I don't see how it can mislead readers into pronouncing it incorrectly (whereas MATR-iss could be interpreted as /ˈmætərɪs/ and MA-triss as /ˈmɑːtrɪs/). I think most readers will know it's affricated, consciously or not, upon seeing a notation like MAT-riss just by the virtue of knowing it is one word and the hyphen separates syllables, not words.

This brings me to the thought I've had for some time: Do we really need to respell frustration, historic, etc. as fruh-STRAY-shən, hih-STRORR-ik as opposed to fruss-TRAY-shən, hiss-TORR-ik? I don't really think readers will be misled and produce affrication or aspiration—and even if some of them did, I don't think it outweighs the risk of fruh-, hih- being interpreted as /frɜː-/, /hi(ː)-/. Nardog (talk) 21:55, 8 December 2019 (UTC)

Actually, yes, I very much do prefer your first (rather than second) respellings of frustration and historic (assuming you mean "hih-STORR-ik") -- those look much more natural than the second respellings! With the respelling you've used for Beatrice, I'd mistakenly assume it's pronounced [biˈæʔ.ɹɪs], when in fact the pronunciation is [biˈæ.t(ʃ)ɹɪs]. Honestly, the best-cast scenario seems like it'd be "bee-AT-triss". That's certainly the clearest for a newbie. For mattress, for example, the best-looking to me would be MAT-triss; there's no ambiguity there about aspiration, or affrication. (Can't syllabication, which is has long been debatable anyway, take a backseat here to pronunciation clarity? In fact, isn't that the point of the respelling key?) Likewise, citron would be SIT-trən; hard to imagine SIT-rən is better. Wolfdog (talk) 11:17, 16 December 2019 (UTC)

Then how would you respell fast-track or headdress? Nardog (talk) 11:41, 16 December 2019 (UTC)

FAST-trak and HED-dress. Wolfdog (talk) 14:19, 16 December 2019 (UTC)

And how do you expect the readers to know HED-dress has two /d/ but MAT-triss one /t/? Nardog (talk) 05:35, 17 December 2019 (UTC)

@Nardog: I'd prefer frust-RAY-shən and hist-ORR-ik, which correctly show the plosives as unaspirated. Kbb2 (ex. Mr KEBAB) (talk) 12:04, 16 December 2019 (UTC)

The fact is that the word frustration is pronounced variably as either frus.tration or fru.stration. Depends on the speaker. (That's frus-TRAY or fruh-STRAY... and I'd be willing to bet that frust-RAY is an alternative too, though less common.) Wolfdog (talk) 14:23, 16 December 2019 (UTC)

@Wolfdog: True. /str/ can be affricated (hence the pronunciation "frushchration", in which the postalveolar affricate after /s/ retracts the initial consonant - a natural consequence of the affricated pronunciation of /tr/). How would you differentiate between fru.stration and frust.ration (or frustr.ation - same thing as the second variant)? They look the same to me. I'd denote the difference in affrication as /frʌˈsdreɪ.ʃən/ vs. /frʌˈstreɪ.ʃən/, with /str/ being the only possible onset with the combination /st/ in English. I'm starting to like the maximal onset principle as well - but that's off-topic. Kbb2 (ex. Mr KEBAB) (talk) 17:34, 16 December 2019 (UTC)

Honestly, either pronunciation would end up affricated normally: [frʌsˈdʒɹeɪ.ʃən] (voiced [dʒ]) vs. /frʌˈstʃɹeɪ.ʃən/ (unaspirated [stʃ]) sound equally good to my ears. Is this answering your question? Wolfdog (talk) 22:49, 16 December 2019 (UTC)

@Wolfdog: Not quite. I meant /frʌˈsdreɪ.ʃən/ vs. /frʌˈstreɪ.ʃən/, so that the initial syllable has no coda. I'm just applying the theory that those plosives that are written with ⟨p, t, k⟩ in English orthography but are categorically unaspirated (meaning: are preceded by a voiceless consonant) actually belong to the /b, d, ɡ/ set (as in "spy" /sbaɪ/, "lisp" /lɪsb/, "stir" /sdɜːr/, "next" /nɛksd/ [not /nɛɡsd/, which blocks pre-glottalization and elongates the vowel], "sky" /sɡaɪ/ and "task" /tæsɡ/ or /tɑːsɡ/). /str/ would then be the only onset that would be the exception to that rule. My question was "how would you differentiate between fru.stration and frust.ration".

EDIT: [fɹʌˈstʃɹeɪ.ʃən] does contain an aspirated /t/. In this case it manifests as an affricate, formed with the following postalveolar approximant. /str/ is the only syllable onset that can contain an aspirated /t/ (it's /str/, rather than /tr/ with /s/ belonging to the coda of the preceding syllable because the same can happen in "street" [stɹiːt ~ stʃɹiːt ~ ʃtʃɹiːt] - notice that the affrication is variable even here). I wonder if there are e.g. Scottish speakers who would differentiate between "fru.stration"/"frust.ration" and "frus.tration" as [fɾʌstɾeʃən] (with two taps) vs. [fɾʌstʃɹeʃən] (with a tap and a postalveolar affricate). Kbb2 (ex. Mr KEBAB) (talk) 08:20, 17 December 2019 (UTC)

The difference between pure tenuis [t] and [d] to me is so minimal as to be almost inaudible. Therefore, I barely, if at all, would differentiate fru.stration and frust.ration. I guess I'm missing the point of this experiment/question. I do, however, like the sound of your "Scottish experiment"; that findings of such a study would indeed be interesting. This conversation is reminding me think of the typical-American pronunciation of sixteen as [sɪks.tʰiːn] vs. the typical-British [sɪk.stiːn] (or even [sɪk.sdiːn]). Wolfdog (talk) 20:25, 17 December 2019 (UTC)

I just don't think it's practicable to try to replicate the exact allophonic realization. Someone seeing MAT-riss for the first time may indeed pronounce it [ˈmæʔɹəs], but once they know it's one word and once it's in their personal lexicon they are much more likely to pronounce it with an affricate—unless they are aware of the allophonic difference, but such a person wouldn't need a respelling in the first place. The goal should be to convey the phonemic values, and MA-triss, MATR-iss, and MAT-triss are all more susceptible to conveying a wrong sequence of phonemes than MAT-riss. Same with fruh-, hih- vs. fruss-, hiss-. (An alternative cure to all this, by the way, is to proscribe respelling these words, as we already do for other cases.) Nardog (talk) 05:35, 17 December 2019 (UTC)

I'm not seeing how my option (MAT-triss) has the same "susceptibility" as the others. Of course, the point of the respelling key is to simplify for nonexperts. That being said, maybe you're right that we should just proscribe respelling these types of words. Wolfdog (talk) 20:25, 17 December 2019 (UTC)

If someone really is confused about how to pronounce a respelling, couldn't they go to an English dictionary website that has audio, or get a dictionary app, and look up the word and listen to it? I think in most cases they would be able to relate what they hear to how the word has been respelled. (Unless the dictionary butchers or otherwise does a poor job of rendering the pronunciation but I guess that's a separate issue.) Senjoro Nie (talk) 14:42, 16 January 2020 (UTC)

The words we transcribe using IPA and respelling are most typically the kinds general dictionaries do not include (see also WP:LEADPRON). Nardog (talk) 12:12, 17 January 2020 (UTC)

kh

Consonants

Rspl.	Example(s)	IPA
kh	loch, Chanukah	/x/

Incredibly helpful. 2001:9E8:26:3D00:2EFD:A1FF:FE73:84E6 (talk) 11:18, 5 November 2020 (UTC)

What is your point? /x/ is a marginal sound only found in loanwords so there are hardly any less obscure words that have it than those two. Nardog (talk) 16:31, 5 November 2020 (UTC)

'ur' sound

This seems to be used for words with 'er' such as 'herd', and that doesn't work for Scottish English where it's a separate sound. Additionally, I don't believe the two examples in the table have the same vowel sound as each other in some Scottish dialects. 115.70.7.33 (talk) 09:16, 29 November 2020 (UTC)