Help talk:Pronunciation respelling key/Archive 5

Replace "moral" example
I replaced the bad example of "moral" for /ɒr/ with a much better example, "tomorrow", which fits the "orr" /ɒr/ sound in way more dialects than does "moral", which more often has the same /ɔːr/ as "horse, hoarse, pour, forum" (indeed, "moral" with /ɒr/ comes off as rather New Yawker). The current wording is just going to be downright confusing to a much larger subset of editors, especially those not from the northeastern US.

Nardog reverted this with "No, that's exactly the intention (while 'tomorrow' can't be represented by one diaphoneme in our system)" which basically makes no sense in either part. 1) It cannot possibly be our intention to use an example that only pertains to a tiny fraction of English speakers, especially when an everyday word better fits the diaphoneme for a vastly larger population of speakers. 2) Of course we can represent it:. Nothing difficult about that. Regardless, if someone doesn't like "tomorrow" either, then find another example. If no one can find one, then we need to not have "orr" /ɒr/ in there at all; just deem it to be the same kind of minor dialectal variation as dropping the final -r in non-rhotic dialects, and not trying to capture the cot/caught distinction that exists in some dialects and not others. This is supposed to be a very simplistic pronunciation system, so let's keep it that way. This system is not meant to try to capture every possible nuance (like some people saying . Hell, this scheme doesn't even distinguish a full "uh" from a schwa. Not sure how I missed that it does. Regardless, it drops many other phonetic distinctions, so why are we keeping this one, and keeping it in a virtuall intentionally confusing way?  — SMcCandlish ☏ ¢ 😼  18:18, 8 April 2021 (UTC); revised: 05:25, 12 May 2021 (UTC)
 * See English-language vowel changes before historic /r/. Tomorrow is one of the only four (or five if you count morrow separately) words that have /ɒr/ in RP, /ɑr/ in GA, and /ɔr/ in Canada. Since we don't have a diaphoneme for such outliers, it can't be represented by a single transcription in our diaphonemic system; it has to be written e.g. ", ". And since our respellings are mere reiterations of the preceding diaphonemic notations in a different format, there is no respelling standing for "/ɒr/ in RP, /ɑr/ in GA, and /ɔr/ in Canada" either. The diaphoneme and the respelling orr represent what becomes /ɒr/ in RP, /ɔr/ in GA and Canada, and /ɒr/ or /ɑr/ in some parts of the eastern and southern US, i.e. the second row of the table in the aforementioned section.
 * Hell, this scheme doesn't even distinguish a full "uh" from a schwa. It does. Nardog (talk) 23:47, 9 April 2021 (UTC)
 * So pick another example in place of tomorrow, then. The point is that "orr" (/ɔːr/) for moral is an uncommon thing particular to a certain dialect or two, so it is directly confusing to most readers to use it as exemplary.  The vast majority of English speakers pronounce moral with the same "o" sound as in more (/ɔːr/).  I'm not sure why I seem to be having difficulty communicating about this with Nardog in particular. Hopefully someones else will join the conversation.  — SMcCandlish ☏ ¢ 😼  22:06, 11 May 2021 (UTC)
 * Can you think of a better example? 'Moral' fits per the OED and MW, and I'd hardly call RP and GA "a dialect or two", though it does have an alt 'tomorrow' pronunciation in the US (which I don't think I've ever heard) per MW. — kwami (talk) 01:22, 12 May 2021 (UTC)
 * In our diaphonemic system, is not  +  but a single unit, just as  is not  +  but a single unit. Otherwise  for story would mean it is pronounced the same as starry by those with the cot–caught and father–bother mergers. If someone pronounces moral (or any word that falls under the second row in this table) with /ɑ/ or /ɒ/, the example for the respelling orr should be read with /ɑ/ or /ɒ/ in that person's accent.
 * The vast majority of English speakers pronounce moral with the same "o" sound as in more (/ɔːr/). That's not true. It has /ɒr/ in RP, look it up in dictionaries. Nardog (talk) 01:39, 12 May 2021 (UTC)
 * RP isn't even numerically dominant in the UK, and even if it were the UK is only small proportion of English speakers, so you're not contradicting what I said. But please . No amount of nit-picking at my side points is ever going to get around the central issue: using moral as the example here is a poor choice when we can probably find an alternative word selection that uses  in more dialects and so is not going to confuse as many readers. I.e., apply MOS:COMMONALITY, which is a guideline for a good reason.  Instead of arguing for the sake of arguing, help cast about for an alternative word.  Cf. WP:FILIBUSTER.  — SMcCandlish ☏ ¢ 😼  05:22, 12 May 2021 (UTC)
 * Moral has in all dialects. Other words with  include forest, Florida, historic, orange. If you pronounce moral with /ɔr/ the chances are you pronounce these with /ɔr/ as well. That's what both the diaphoneme  and the respelling orr represent; its realization varies across accents, but within each accent the realization is the same across words. You have yet to establish why we need a replacement in the first place.
 * You bring up WP:FILIBUSTER, but could you tell me what viewpoint I'm "pushing ... with which the consensus of the community clearly does not agree"? The diaphonemic nature of seems pretty clear according to Manual of Style/Pronunciation (which strikes me as very much in line with, and motivated by, MOS:COMMONALITY). Nardog (talk) 06:07, 12 May 2021 (UTC)

The very point of having a diaphonemic notation is to make the pronunciation in the most significant varieties of English (GA, Canada, RP, Oz) fully predictable. This sometimes entails making contrasts that do not exist in every variant (such as the /ɔːr/-/ɒr/-contrast). But it's existence in RP should be enough reason to have it in the key Help:IPA/English and consequently also here. Otherwise, diaphonemic notation becomes pointless. It's the only reason we have e.g. /ˈlɛstər/ Leicester instead of /ˈlɛstə/. Not because [ˈlɛstər] is "more common" than [ˈlɛstə], but because /ˈlɛstər/ can be predictably mapped into [ˈlɛstər] in rhotic and as [ˈlɛstə] in non-rhotic varieties. With /ˈlɛstə/, this doesn't work. MOS:COMMONALITY in this case takes the shape of a common denominator–it takes sixths to accomodate halves and thirds. Brits not familiar with the diaphonemic concept might interpret /ˈlɛstər/ as US-centric. –Austronesier (talk) 12:52, 12 May 2021 (UTC)
 * Add What's weird though is that only few words which have the diaphoneme /ɒr/ are actually spelled with -orr- (sorrel is one, I guess), so the respelling is quite unintuitive. That said, I don't have an idea for a better one at the moment. –Austronesier (talk) 13:16, 12 May 2021 (UTC)
 * There are some: abhorrent, correlate, correspond, corridor, corrigendum, corrugate, Gomorrah, horrible, horrid, horrify, horror, incorrigible, Morris, porridge, torrent, torrid. Nardog (talk) 14:50, 12 May 2021 (UTC)
 * Oh, that's a lot. 'The horror! The horror!' that I haven't thought of those. So much for L2-intuition :) –Austronesier (talk) 15:00, 12 May 2021 (UTC)
 * To my knowledge it is a common occurrence for GA-type speakers to either enunciate or adopt the NY/NE type ad hoc in horror in an attempt to differentiate it from whore. Nardog (talk) 15:07, 12 May 2021 (UTC)

Orange tangent
(One nitpick to the above: orange doesn't fit with the others across dialects. We might need two transcriptions if it were a word we'd bother transcribing.) — kwami (talk) 22:18, 12 May 2021 (UTC)
 * Are you saying orange belongs to the same group as tomorrow, sorry, etc.? Wells (1982: 476) didn't think that. His Longman Pronunciation Dictionary (3rd ed., 2008) reports 80% of AmE respondents preferred /ˈɔːr-/. CEPD and RDPCE don't even include the /ˈɑːr-/ variant, suggesting they don't consider it part of their American models. Kenyon & Knott, Oxford, American Heritage, Random House, OALD, Longman, Cambridge, Collins and Houghton Mifflin, and Macmillan also show the variant with NORTH/FORCE first, if not exclusively. The only dictionaries I could find that put the one with START first were Merriam-Webster, which is admittedly what I usually most trust when it comes to American English, and its learner's dictionary. I checked the past editions of Merriam's Collegiate, and it looks like the last edition that had \ˈȯr-\ first was the 1973 8th edition and they switched to \ˈär-\ first in the 1983 9th, with the completely identical transcriptions and qualifier to the latest web version (last updated 12 May 2021!). Does this mean it's on its way to joing tomorrow? Warrant any change to English-language vowel changes before historic /r/? What say you? Nardog (talk) 23:22, 12 May 2021 (UTC)
 * Not that either. For me at least, orange has the moral vowel, tomorrow and sorry do not. Unless maybe I just got confused by people who have [ɑr] and only noticed it in that one word orange? I concluded that orange was idiosyncratic, but could easily be wrong. — kwami (talk) 00:17, 13 May 2021 (UTC)

/ər/
for intuitivity and familiarity, i think we should use  rather than <ər> for the R-coloured schwa sound. — Preceding unsigned comment added by Omoutuazn (talk • contribs) 12:59, 19 May 2021 (UTC)


 * If we make transcriptions familiar, they become ambiguous and therefore useless. — kwami (talk) 23:04, 19 May 2021 (UTC)

/aʊ/
i propose that in situations where it could be misinterpreted as /oʊ/ (i.e. the "Maui" example), /aʊ/ be respelled . this is seen in textspeak pronunciation respellings like "nao" as well as onomatopoeia like "miaow", so it's clearly intuitive to english speakers, and it's in a lot of pre-existing words like "cacao". — Preceding unsigned comment added by Omoutuazn (talk • contribs) 01:29, 9 May 2021 (UTC)
 * For clarity, it's currently being represented in this system by "ow".  — SMcCandlish ☏ ¢ 😼  22:09, 11 May 2021 (UTC)
 * i know but that leads to confusion like /maʊ/ being respelled — Preceding unsigned comment added by Omoutuazn (talk • contribs) 12:57, 19 May 2021 (UTC)
 * But there will always be such confusions, no matter which conventions we use, unless we abandon any parallel to English orthography. The only practical solution is to not use respelling for confusing words, or to describe them some other way ('rhymes with X' etc.). — kwami (talk) 23:05, 19 May 2021 (UTC)

Representing /aʊ/ as AO
"Particularly, respelling /aʊ/ could prove problematic as there are a variety of monosyllabic words spelled with 'ow'." (§When to use and when not to use). I have a solution: represent /aʊ/ as AO. AO, usually makes an /aʊ/ sound, even though it isn't used in most English words. Electos242 (talk) 18:24, 23 October 2021 (UTC)
 * Sorry, didn't see the previous thread. Electos242 (talk) 18:25, 23 October 2021 (UTC)

ewr
ewr always struck me as suboptimal. It might work for rhotic speakers since it is essentially ew + r for them, but non-rhotic speakers might simply ignore the r instead of diphthongizing it. So how about ure? Compare with Nardog (talk) 19:06, 8 November 2021 (UTC)
 * Done. Nardog (talk) 13:53, 4 December 2021 (UTC)

The GOAT vowel
The digraph here used to represent the /əʊ/ phoneme is rare, appearing only in a few words ('aloha,' 'kohl,' 'mohel,' 'oh,' 'ohm' and 'pharaoh' being, I believe, the only ones in common use). Although used frequently as an exclamation of surprise, it still looks awkward in respelling. A solution I found to be both unambiguous and intuitive is to represent /əʊ/ with 'oa,' except in syllable-final positions, where 'oe' is preferable. This is because 'oa' is rarely syllable-final and might be misread as /əʊə/, while 'oe' scarcely appears intrasyllabically. Also, take note that 'oar' is a much more logical representation of the FORCE vowel that 'ohr,' although I see that this respelling system has the NORTH and FORCE vowels completely merged.

Some examples: kən-TROH-lər, TEL-i-fohn, MOH-tər-boht (current); kən-TROE-lər, TEL-i-foan, MOE-tər-boat (proposed). Maciuf (talk) 19:32, 31 January 2022 (UTC)
 * Awkward is better than variable. oa is usually disyllabic when word-final as you point out, and even so word-internally in Boas, Croat, etc. oe is also not without exceptions like shoe and canoe. Different respellings conveying the same value not only could be confusing to readers but puts an extra burden on editors to remember the rule. Sticking with oh does not have these problems. Rather, oh is an ideal choice as it consistently conveys the same value to everybody precisely because it's rare in actual words and thus has few if any deviations. The same could be said about ah. Nardog (talk) 19:47, 31 January 2022 (UTC)
 * Thank you for your reply. Personally, I can hardly imagine anyone misreading an intrasyllabic 'oa' as /əʊ.æ/ or a syllable-final 'oe' as /uː/; even so, if the same respelling is to be used regardless of position within the syllable, 'oh' is probably the only reasonable choice. I am still of the opinion that the intrasyllabic 'oa' in particular could prove better intelligible to the average reader, but I do agree that the current solution is, however ungraceful it may seem, perspicuous notwithstanding. As for 'ah,' the only other (and most likely better) option is 'aa'; in this case, however, both alternatives are not found commonly in words, unlike 'oa' or 'oe'. Maciuf (talk) 13:17, 1 February 2022 (UTC)

Doubled r's
It has already been stated within the boundries of this talk page that the 'distinction between ar and arr, ir and irr, or and orr, er and err […] is far from intuitive,' and 'given that we have symbols for the short vowels, and these [/ær/, etc.] occur across a syllable boundary, it's hard to see how these ['arr,' etc.] are necessary.' However, no changes in this regard have been introduced since. The idea is simple: instead of writing MARR-ee, MERR-ee, MIRR-ər, MORR-əl, HURR-ee and CUURR-ee-ər, write MA-ree, MEH-ree, MIH-rər, MO-rəl, HUH-ree and CUU-ree-ər (cf STAR-ee, OR-əl, FUR-ee). I hereby recommend that the necessary amendments be introduced. Maciuf (talk) 18:08, 1 February 2022 (UTC)
 * Opposed. EH, IH and UH are worse than double vowels, which are quite intuitive, as they intuitively suggest those vowels are long when they're short. — kwami (talk) 22:48, 1 February 2022 (UTC)
 * MA-ree and MO-rəl would no doubt be read with and, and MEH-ree, MIH-rər, and HUH-ree potentially with  (or [ʌ] when it should be [ɚ] in their accent). Granted, we do include eh, ih, uh, but they are a compromise to be used before aspirated consonants and consonant clusters beginning with /s/. Using them in stressed syllables is counterproductive. Nardog (talk) 00:04, 2 February 2022 (UTC)
 * Thank you for your replies. The usage of 'eh,' etc., however, is not an issue I meant to bring up in this section; MEH-ree, MIH-rər and HUH-ree could just as well be ME-ree, MI-rər and HU-ree if you like. Nevertheless, if you consider the current solution more intuitive, I assume it ought to remain unchanged. Maciuf (talk) 09:26, 2 February 2022 (UTC)
 * My first instinct is to read ME-ree, MI-rər and HU-ree as (or, with a morpheme boundary before ). These won't work either. The diaphonemes we're discussing here occur only in the word-internal position before vowels. In that position, $⟨arr, irr, orr, err⟩$ are unambiguous spellings and all of them occur in native words (marry, mirror, torrent, merry). Not only that, they're unambiguous cross-linguistically, as far as Germanic languages (German, Dutch, Swedish, Norwegian) are concerned anyway, if you can connect the dots (English  =  or  in those languages, whereas both  and  are more or less the same). In all of those languages, a doubled consonant letter denotes a checked vowel (save for compounds, but that's obvious), just as in English. Those respellings are perfect for this guide, AFAICS.
 * Not to mention that $⟨CC⟩$ in orthography denotes a single consonant in non-compounds, just as in German and Dutch. Sol505000 (talk) 00:56, 8 February 2022 (UTC)
 * Thank you for your reply. The negative responses enable me to surmise that the proposed improvement would not be of any benefit. This discussion may therefore be considered closed. Maciuf (talk) 08:47, 16 February 2022 (UTC)

The PRICE vowel
If abidance by the precept of avoiding 'different respellings conveying the same value' (Nardog 2022) does not make possible the elimination of 'oh' from the pronunciation key, how can the co-existence of two possible representations of the /aɪ/ phoneme (including one with a silent 'e') be accordant with the principles of Wikipedia's respelling system? I firmly believe that the use of 'igh' (or, perhaps also, 'uy') instead would be both simpler and similarly unambiguous. Below is a juxtaposition of the current scheme and the two alternatives; I deeply apologise for any errors it might contain.

Maciuf (talk) 22:37, 3 February 2022 (UTC)
 * Our respelling of PRICE is indeed imperfect and I've seen it applied inconsistently in articles (eye after an onset, ye instead of y, etc.). igh works in open syllables but not in closed ones (see e.g. sighz, vighd, g(h)ighst). I like uy, I don't see a situation where it can be misinterpreted (I bet if any complaint were to be raised it would be "why respell e.g. spuy when there's the word spy?", which can be said about many other respelling combinations anyway). Nardog (talk) 14:51, 6 February 2022 (UTC)
 * Yes, uy seems to be pretty good. I'd rather have one respelling for PRICE, so I strongly support the change. Sol505000 (talk) 00:51, 8 February 2022 (UTC)
 * I'll make the switch in a day or two if no one contests. Nardog (talk) 13:49, 11 February 2022 (UTC)
 * It's awkward and unintuitive, but we haven't been able to think of anything better. I'd suggest the mouseover keyword "uy as in buy". I expect a good amount of push-back, though. — kwami (talk) 23:47, 11 February 2022 (UTC)
 * Though I like the uy solution much better than the igh one, I think it comes off as a bit of an overkill in most cases (when only y is very well enough). I think I would support using it to replace eye, but it looks odd and indeed unintuitive to me in the other instances. 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 09:34, 12 February 2022 (UTC)
 * I rather find y to be in more urgent need of an alternative. In fact we used to use vyss to illustrate the limits of our respelling system at the top of Help talk:IPA/English (perma). Adding an e after the consonant somewhat mitigates the obscurity, but the effect is lessened when it conflicts with the doubling of s (I find vuyss superior to all of vys, vyse, vyss, vysse) or when there's a complex coda. Maczkopeti tried to salvage the unhelpful HYNDZ at Ciarán Hinds by substituting HYNDZE but eventually realized it wasn't working and removed the respelling altogether. HUYNDZ is at least better than those. Also the rule is not followed. See Brunei, Dubai, Papyrus, Christiaan Eijkman, Phyllite.
 * And then there's the benefit of consistency. UYN-stuyn is better than EYEN-styne or UYN-styne as the same sound is represented by the same combination of letters. Nardog (talk) 10:25, 12 February 2022 (UTC)
 * Yeah, I understand. Still it seems absolutely redundant to have transcriptions like wuy-OH-ming or SPUY-see. Would a good compromise be using uy in those cases when we currently already have to substitute plain y (i.e. eye AND y-e), and leave it elsewhere? 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 10:55, 12 February 2022 (UTC)
 * An of course fix the cases where the transcriptions were misused. 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 11:12, 12 February 2022 (UTC)
 * The whole idea is to have one respelling for PRICE, so that $⟨uy⟩$ (or any other umabiguous spelling, if we can come up with one) is used like its IPA counterpart . It is simple and consistent. Using two or more respellings is not simple. Sol505000 (talk) 11:14, 12 February 2022 (UTC)
 * I don't think we would have had broo-NYE, pə-PYE-rəs, etc. (which is surprisingly common) if people had thought ny, py more intuitively conveyed . Rather, the inconsistency must stem from the very fact we respell in two ways: they just see eye first and apply it, and they see neye, peye, think it's awkward, and remove the first e, not realizing they could (and should) remove both e's. The fewer symbols represent the same sound the better.
 * I don't see how wy better conveys than wuy, and the spy issue is not unique to . There are always going to be combinations like "why write bayss when there's the word base?", but that's a feature, not a bug, of any respelling system that is consistent and not ad hoc. And ad hoc respellings are prone to ambiguities. Nardog (talk) 11:18, 12 February 2022 (UTC)
 * Do you still object to respelling all PRICE as uy? Nardog (talk) 13:46, 13 February 2022 (UTC)
 * I was never really against, it just seemed to me too much to use it everywhere. But if I’m the only one to think that, of course proceed. 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 13:51, 13 February 2022 (UTC)
 * Ok, doing it now. Nardog (talk) 15:55, 13 February 2022 (UTC)


 * I have to say, I think you folks have lost the forest through the trees here. My understanding of the purpose of respellings is that they are to help the reader understand how to pronounce a word. I don't think this change does that, at all, and in fact would lead persons who haven't read this specific discussion to badly mispronounce these words. Beeblebrox (talk) 19:57, 13 February 2022 (UTC)
 * Just so we understand where you're coming from, how would you pronounce (or how do you think people would pronounce) KEE-nuy? Nardog (talk) 20:34, 13 February 2022 (UTC)
 * I'm not sure actually. It seems to imply there is a "u" sound in there, when there for sure is not. Maybe like "annoy"? It's not a combination you see much in English so it's hard to even understand what it is supposed to mean. Beeblebrox (talk) 22:22, 13 February 2022 (UTC)
 * Buy and guy were not as popular as I thought then. Nardog (talk) 22:44, 13 February 2022 (UTC)
 * To me, "KEE-nuy" looks like the last syllable should be like in buoy or Rapa Nui or chop suey or Louis (but blurred into a single compound vowel rather than two separate ones). Alternatively, it could be a simple u vowel followed by a consonental y, as I suspect some accents might pronounce Cluj-Napoca (although our pronunciation guide there says something else) and as my friend Vida Dujmović pronounces the first syllable of her surname. If you want it to sound like "keen eye", uy is very unintuitive to me. —David Eppstein (talk) 23:26, 13 February 2022 (UTC)
 * The spellings you mention are all different from $⟨uy⟩$, so I don't see how they are relevant. For instance, $⟨ow⟩$ is mostly ambiguous between and  in English, yet $⟨ou⟩$ is almost always  or some other vowel, not . Polish sports commentators notoriously mispronounce Southampton as  (consonantal differences aside) - that's a non-native pronunciation based on a lack of understanding of English orthography - that article says (as expected) that $⟨uy⟩$ for  occurs only before vowels. Obviously, because  starts the following syllable - therefore a monosyllabic $⟨uy⟩$ can't possibly be read . The rendering of Dujmović with  is strongly foreign, by the way.   or   (or possibly  ) is how I'd say it, following English phonotactics. Among the diphthongs,  can start as high as  in some dialects, but normally it starts with  or  (both are considerably lower and more back than your typical English ). English does not have a phonemic  or, nor does it allow a syllable-final  (and an onset  is out of question in English, as it is in Serbo-Croatian).
 * Ah, per Wells, there's an additional phonetic, as in fluid , which is in a free variation with the disyllabic , which is the underlying form . I think it's pretty rare, though (and it's spelled $⟨ui⟩$, not $⟨uy⟩$). This further supports the anglicization of Dujmović as (or , to make it even easier to pronounce).
 * To rephrase it, this $⟨uy⟩$ is used the same as the most common usage of $⟨uy⟩$ in English, as in "buy" or "guy". As all of our respellings are! We're not using $⟨i⟩$ to represent the FLEECE vowel, which is how is spelled in Spanish, French, Italian, etc. Sol505000 (talk) 13:49, 14 February 2022 (UTC)
 * Your first mistake is in "The spellings". I was not talking about spellings. I was talking about pronunciations. The pronunciations that I, trying to understand "uy" as an intuitive pronounciation guide, would use to interpret it. Based on my intuition, at least, "uy" totally fails at being an intuitive way of representing the vowel sound in "buy" or "guy" because, separated from those words, it just doesn't look like it should be pronounced that way. —David Eppstein (talk) 18:54, 17 February 2022 (UTC)
 * I'm of the same opinion as Beeblebrox. I was alerted to this matter by Nardog's alteration today of the respelling of "Lyceus" to "luy-SEE-əs" in the Apollo article, and I don't think there's any way that a naive reader would interpret "luy" as representing the syllable /laɪ/. Moreover, the former system of representing the /aɪ/ phoneme used on this page—"eye" in syllable-initial positions and "y" elsewhere—is exactly the system used in The World Book Encyclopedia, a reference work that many American readers, at least, are familiar with from their schooldays. (I think, however, that the statement "When y is followed by a consonant within the same syllable, place an e after the consonant as necessary" is totally uncalled for. "TYT" is a better respelling of tight than "TYTE".) I wish that Nardog would have waited for further input here before altering this page and proceeding to alter pronunciation respellings in actual articles. Deor (talk) 21:49, 13 February 2022 (UTC)
 * I don't think there's any way that a naive reader would interpret "luy" as representing the syllable /laɪ/ Really? What would they think it is instead? "TYT" is a better respelling of tight than "TYTE" I'm sure you'll have a tough time finding company on that. As I noted above, vyss was held as an example to illustrate the limits of respellings, and look at what we have at Rachel Weisz now. I'll undo the change for now and maybe start an RfC but I wish I was doing so because I was persuaded by more compelling arguments. Nardog (talk) 22:44, 13 February 2022 (UTC)
 * I know that it sounds odd in the context of this discussion, but to my mind one virtue of a good respelling system is that one sound corresponds as nearly as possible to one respelling. The addition of extraneous "silent e's", as in "TYTE", serves no purpose that I can see. Deor (talk) 19:53, 14 February 2022 (UTC)
 * I wholeheartedly agree that "UY" is a very non-intuitive respelling of the /aɪ/ sound. I saw some of Nardog's changes today and initially thought it was a misguided attempt by a British editor to represent a West Country dialect, standing for /ɔɪ/ or /uːɪ/. If I saw it in an unfamiliar word without the IPA present, there is no way that I would guess that it is intended to represent /aɪ/. CodeTalker (talk) 23:38, 13 February 2022 (UTC)
 * Same here. It gives no help whatever, because BUY is so unusual. To me, it is too reminiscent of OY as in BOY. I had no idea whatever what was intended, BUY simply didn't occur to me. I wonder if we need an additional notation for these awkward vowel sounds such that 'tight' is respelled t[eye]t? (where the word in square brackets is the most common three letter word that has the appropriate pronunciation. We are in danger of reinventing IPA here! --John Maynard Friedman (talk) 00:02, 14 February 2022 (UTC) updated 00:04, 14 February 2022 (UTC)
 * Remember we have uu for . We should probably just discourage respellings altogether when there are no good options (as we do for ow ). Nardog (talk) 00:11, 14 February 2022 (UTC)
 * It has been mentioned that some cannot but see a 'u' sound in 'uy'; however, 'u' is most commonly pronounced /ʌ/, which in combination with 'y' yields /ʌɪ/ — another rendition of /aɪ/. There is no /uɪ/ diphthong in English. I would not expect the naive reader to anticipate quirky dialect emulations within an encyclopaedia. Maciuf (talk) 11:18, 14 February 2022 (UTC)
 * Exactly. Even if you read $⟨uy⟩$ as +  (which is not a possible combination in English, except in very rare circumstances where  occurs before a stressed syllable that begins with a ), you'd come up with something very close to the intended . In some dialects it may even be identical. Sol505000 (talk) 13:31, 14 February 2022 (UTC)
 * The only place UY is pronounced /aɪ/ is at the end of a syllable. Adding a consonant after UY (for me) breaks the analogy with BUY and GUY. Can you honestly say that if you saw an unfamiliar place name "Fuyt", you would immediately assume that the only reasonable pronunciation was a homonym for "fight"? That would be pretty far down the list for me. CodeTalker (talk) 17:40, 14 February 2022 (UTC)
 * Although I have no feelings for or against any particular transcription, I agree with some that  is certainly unintuitive, even despite the rarities "guy" and "buy". Critics here don't need to answer the question "Other than, how else would readers interpret ?" It's enough merely for critics to assert that readers will react to  with "Huh? What in the world is this?" However, as others have said, any uniform transcription of this phoneme will lead to reader problems. The topic of this debate, then, ought to be which one will yield the least problems. (Sidenote: I'm personally apt to interpret  as , as in screwy or phooey. Changing shy to for example is more likely to be read as some English approximation of an East Asian word. Meanwhile, might seem like more than die.) Wolfdog (talk) 22:31, 14 February 2022 (UTC)
 * Again, why would any reader interpret it as, a disyllabic sequence? We already divide all words into syllables. would be , just as greater is  and palace is . Why would someone assume that  is treated any different? Sol505000 (talk) 23:11, 14 February 2022 (UTC)
 * Just told you that I did. I also just explained why. I was talking about what it intuitively looks like or not. Wolfdog (talk) 03:18, 15 February 2022 (UTC)

While a final decision for uy is pending, can I suggest replacing current eye cases in a close syllable with y…e and limiting eye to syllables where only the diphthong occurs (and to the sequence)? i.e., "ice" would become YSSE instead of EYESS, "Einstein" would be YNE-styne rather than EYEN-styne and so on. What do you think? 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 16:21, 15 February 2022 (UTC)
 * That would make item respelled Y-təm, which looks like . And my brain read your example as Weinstein. Nardog (talk) 19:08, 15 February 2022 (UTC)
 * No, I explicitly stated “limiting eye to syllables where only the diphthong occurs”, such as in EYE-təm. Asking any native speakers, how would you tentatively read yne? 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 19:11, 15 February 2022 (UTC)
 * Oh, my bad. But since it would not reduce the total number of respellings, I see little motivation for this change. Nardog (talk) 19:18, 15 February 2022 (UTC)
 * It would not reduce the number of respellings but it would limit eye only to when strictly necessary. At first I had thought a general change to ye for all instances of syllable-final (YE-təm, ə-LYE etc.), but then I realized it would be interpreted as  in YE-təm-like transcriptions. 〜 イヴァンスクルージ九十八 ［IvanScrooge98］  （ 会話 ） 19:31, 15 February 2022 (UTC)
 * I realize that my suggestions are unlikely to find favor here, but I think it would be a good idea to restrict eye to syllables consisting of nothing but the vowel in question ("EYE-təm" for item, for instance) and use just y for all other cases ("YS" for ice, "YN-styn" for Einstein, etc.). Deor (talk) 22:05, 15 February 2022 (UTC)
 * Actually, I like that. I might interpret "YSSE" as "hissy" less the 'h'. I'd have a hard time seeing it as "ice". "YS" is odd enough that I'd know not to try to read it as English orthography, and so remember the value 'Y' is supposed to have. So here are the two least-bad options IMO:
 * {| class="wikitable mw-collapsible mw-collapsed" role="presentation"

eye-DEN-tə-tee EYE-təm PRY-mə-ree PRY-vət bi-HYNDE LY-brə-ree TY-təl SY-əns sə-SY-ə-tee və-RY-ə-tee TRY-əl ə-RYVE di-SYDE di-FYNE di-SKRYBE EK-sər-syze prə-VYDE REE-ə-lyze sə-PRYZE ə-LYNE ə-SYNE di-ZYNE ə-PLY di-NY juu-LY neer-BY eye-DEN-tə-fy OK-yoo-py hy-POTH-ə-sis sy-KOL-ə-jee BYTE NYFE pri-SYSE PRYDE WYPE WYZE BY-nə-ree EYE-və-ree NY-jər TY-gər BY-bəl GRYNDE BRY-ən dy-əg-NOHS EYE-ə-wə oh-HY-oh BRY-tən dy-NAM-ik hy-EE-nə PY-thən RYME SPY ty-KOON wy-OH-ming POL-tər-gyste ROT-vy-lər shy-AN EYE-ki-doh TY-lənd PARR-ə-gwy eye-DEN-tə-tee EYE-təm PRY-mə-ree PRY-vət bi-HYND LY-brə-ree TY-təl SY-əns sə-SY-ə-tee və-RY-ə-tee TRY-əl ə-RYV di-SYD di-FYN di-SKRYB EK-sər-syz prə-VYD REE-ə-lyz sə-PRYZ ə-LYN ə-SYN di-ZYN ə-PLY di-NY juu-LY neer-BY eye-DEN-tə-fy OK-yoo-py hy-POTH-ə-sis sy-KOL-ə-jee BYT NYF pri-SYS PRYD WYP WYZ BY-nə-ree EYE-və-ree NY-jər TY-gər BY-bəl GRYND BRY-ən dy-əg-NOHS EYE-ə-wə oh-HY-oh BRY-tən dy-NAM-ik hy-EE-nə PY-thən RYM SPY ty-KOON wy-OH-ming POL-tər-gyst ROT-vy-lər shy-AN EYE-ki-doh TY-lənd PARR-ə-gwy uy-DEN-tə-tee UY-təm PRUY-mə-ree PRUY-vət bi-HUYND LUY-brə-ree TUY-təl SUY-əns sə-SUY-ə-tee və-RUY-ə-tee TRUY-əl ə-RUYV di-SUYD di-FUYN di-SKRUYB EK-sər-suyz prə-VUYD REE-ə-luyz sə-PRUYZ ə-LUYN ə-SUYN di-ZUYN ə-PLUY di-NUY juu-LUY neer-BUY uy-DEN-tə-fuy OK-yoo-puy huy-POTH-ə-sis suy-KOL-ə-jee BUYT NUYF pri-SUYS PRUYD WUYP WUYZ BUY-nə-ree UY-və-ree NUY-jər TUY-gər BUY-bəl GRUYND BRUY-ən duy-əg-NOHS UY-ə-wə oh-HUY-oh BRUY-tən duy-NAM-ik huy-EE-nə PUY-thən RUYM SPUY tuy-KOON wuy-OH-ming POL-tər-guyst ROT-vuy-lər shuy-AN UY-ki-doh TUY-lənd PARR-ə-gwuy
 * colspan="3" | Y vs UY respellings of /aɪ/
 * Current
 * 'Y'
 * 'Uy'
 * 'Uy'
 * }
 * Either way, it would be good to get rid of the silent E's. But it's a lot easier for me to recognize "RYM" than "RUYM", "di-SKRYB" than "di-SKRUYB", so I support Deor's proposal despite it having context rules, with the UY proposal as my second choice. — kwami (talk) 02:49, 16 February 2022 (UTC)
 * It ought to be remembered that using 'igh' is also a possibility. Although it only naturally occurs in open syllables or when directly followed by 't,' it does not strike one as anomalous in respellings such as ə-RIGHV and EK-sə-sighz; however, it is admitedly awkward in some contexts — especially before consonant clusters — as in GRIGHND, pri-SIGHSS and di-SKRIGHB. Maciuf (talk) 09:34, 17 February 2022 (UTC)
 * For weird that it might look, I would still prefer it to y for cases like RYM or ə-RYV. We should bear in mind that this help is thought to be an intuitive replacement for IPA, and plain y in closed syllables would mostly make me think of an sound. 〜 イヴァンスクルージ九十八 ［IvanScrooge98］  （ 会話 ） 10:09, 17 February 2022 (UTC)

And that is the thing, respellings should be intuitive. If the average reader is instead puzzled by them, they are not doing their job. The point is not to show how clever you are, it is to help the reader know how to say uncommon words. Beeblebrox (talk) 20:59, 17 February 2022 (UTC)
 * But no system is going to be completely intuitive. One pronunciation of "Balt" and "malt" is rendered "BOLT" and "MOLT", which is certainly not intuitive. Best we can do is minimize the unintuitive cases, at the same time we try to minimize the number of contextual rules. — kwami (talk) 21:26, 18 February 2022 (UTC)
 * But then again, di-ZYN, BYT, NYF, pri-SYS, PRYD, … without e seem to expand the number of unintuitive cases instead of minimizing them. How many readers are going to look at WYP and reasonably not end up saying “whip”? 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 22:10, 18 February 2022 (UTC)
 * Because $⟨y⟩$ usually stands for before a single consonant. WYP is likely to be read . Sol505000 (talk) 01:32, 19 February 2022 (UTC)
 * I’m still waiting for enough natives to confirm this is intuitive to them. 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 07:08, 19 February 2022 (UTC)
 * One pronunciation of "Balt" and "malt" is rendered "BOLT" and "MOLT", which is certainly not intuitive. This issue is solved by not including the alternative pronunciation, as recommended by the help page; those who shorten the vowel are likely to read BAWLT as /bɒlt/ anyway. Contrariwise, you cannot avoid using the PRICE vowel. How many readers are going to look at WYP and reasonably not end up saying “whip”? Actually, quite many — meaning, of course, those who distinguish between /w/ and /hw/. Because ⟨y⟩ usually stands for /aɪ/ before a single consonant. WYP is likely to be read /waɪp/. You are refering to an article that does not consider pre-consonantal ⟨y⟩ when in in a word-final syllable, in which case it is pronounced /ɪ/, as in 'wyn' and 'polyp.' I’m still waiting for enough natives to confirm this is intuitive to them. Yes, YNE-styne is preferable to EYEN-styne. Maciuf (talk) 15:48, 19 February 2022 (UTC)
 * Thank you. (And of course I was simplifying not counting the /hw/–/w/ distinction, for the sake of my point. :D) 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 17:48, 19 February 2022 (UTC)
 * Part of the problem is of course that English contains tons of loanwords and place names in particular often adhere to no coherent rule at all (Bexar County being my personal favorite example,the x is silent, it's pronounced just like "bear"). The example that brought this to my attention had to do with Kenai, the name of a town, a borough, a peninsula,a mountain range, and an enormous lake. Having lived on said peninsula for many years and worked in tourism-related business for a lot of that time I have heard it mispronounced many times, but always the "ken" part, which people tend to pronounce like the name "Ken". Can't really fault them for that, there is no clue that there is a long e there. I don't think I've heard a native English speaker mispronounce the "ai" part, I think that is fairly intuitive. That's the other problem, this is all a bit subjective. Beeblebrox (talk) 20:44, 19 February 2022 (UTC)
 * Yeah, and that's also part of the point: since English can be so unpredictable, we should try to use the most predictable sound-to-spelling correspondence as possible, while also trying to be straightforward and avoiding using more than one (or two) spelling(s) for the same sound. 〜 イヴァンスクルージ九十八 ［IvanScrooge98］ （ 会話 ） 22:16, 19 February 2022 (UTC)

Not enough natives -- I'm native English-speaking, and prefer Y. Yes, people are going to initially read WYP as "wip", but that's why we link to the key. If you don't check the key, you're going to get a lot of pronunciations wrong, just as with any dictionary. (Even with IPA in the OED, you need to check the key to understand which IPA conventions they're using.) The question for me is, once a reader checks the key, are they going to remember that Y is the /ai/ vowel? I expect that most of the time it should work pretty well. — kwami (talk) 22:39, 19 February 2022 (UTC)
 * I concur with the comment above. (Also a native speaker of English, for whatever that's worth.) Deor (talk) 22:54, 19 February 2022 (UTC)
 * The whole point of this respelling scheme is to satisfy people who refuse to refer to a key. If we could count on readers checking the key to understand a phonetic notation, we could scrap this respelling scheme entirely and just use IPA. Indefatigable (talk) 23:01, 19 February 2022 (UTC)
 * We can't idiot-proof WP. I think of this as similar to metric vs imperial. It would of course be preferably to use only metric on WP, but there are a lot of Usonians (the majority of native English-speakers) who are unfamiliar with metric and find it alienating. The same with IPA. But just because we stick to the basic Latin alphabet as an alt convention doesn't mean we can get away without having any rules. "UU" for the PUT vowel is unintuitive, and you can't be sure what it means without checking the key. Whether we use Y or UY or IGH, readers are going to have to check the key to have any confidence in the pronunciation. — kwami (talk) 12:26, 20 February 2022 (UTC)
 * Again, the way forward will be picking the transcription that yields the fewest confusions. (All transcriptions will yield some.) Wolfdog (talk) 13:42, 20 February 2022 (UTC)
 * Intuitively representing /ʊ/ is difficult; employing 'u' and doubling the following consonant (the best solution I could think of) is far from satisfactory, hence the use of 'uu,' which, though obscure, is not likely to be misread (and, unlike 'ʊ,' does not require a special character); in the case of /aɪ/, there seem to be better choices than the analogous 'y.' Maciuf (talk) 19:09, 20 February 2022 (UTC)
 * I think UU is from the BBC. Don't remember what their /ai/ was; elsewhere I've seen AHY. UY is similar to UU in that, while it has no obvious meaning to the reader, at least it can't mean anything else. IMO 'Y' is more user-friendly than UY or AHY, but at the expense of having other readings. — kwami (talk) 19:47, 22 February 2022 (UTC)

Literally no normal person understands these LITERAL greek characters or how they sre supposed to be pronounced normally,
Let alone in the ways they seem to be used here. We live in an age where everyone has microphones in their pockets, why not ditch this and use audio pronunciation instead? 2603:900B:405:EF1E:1CEE:EE20:3DE2:C169 (talk) 09:54, 14 May 2022 (UTC)


 * Many people (especially in the UK, where most dictionaries use IPA) understand those characters. Those who don't may hover over the individual symbols for prompts; besides, there already is a fairly intuitive respelling system, and you may add recorded pronunciations. All those formats should be retained as both respelling and IPA have the advantage of being more accurate, not requiring the use of an audio device and being easier for editors to enter; audio pronunciation could also be confusing for people of various accents. The reason IPA may be preferred to respelling is that it is a widespread standard among phoneticians and is familiar to many, specifically Britons. Maciuf (talk) 14:30, 20 July 2022 (UTC)

PRYSE or PRYCE
While reading the rules for the EYE sound, I noticed that the example given, "price" PRYSE, could be seen as having a /z/ sound, pronounced the same as prize. Within the rules of the respelling key, this is unambiguous but for someone who doesn’t know the rules, they could see YSE pronounced like eyes. We already have respellings such as gh, ss and tch to help the reader. I think replacing the s in respellings like PRYSE with a c would be more helpful for the reader. SlimyGecko7 (talk) 19:09, 24 August 2022 (UTC)


 * Oh, it should be written with ss. I’ll change that in the key. SlimyGecko7 (talk) 19:00, 7 September 2022 (UTC)
 * No, that makes it "priss". I don't know if there is any way to write "price" in our system. There used to be, but currently, like "Balt" and "Maui", that word should probably just be avoided. — kwami (talk) 08:39, 9 September 2022 (UTC)
 * (Also, should we have an example of "YEYE"? — kwami (talk))

Semi-protected edit request on 21 March 2023
Change the word 'dew' under the 'ew' sound example to 'few'; as 'dew' is pronounced /uː/, and not /ju:/. Thequantaleaper (talk) 23:13, 21 March 2023 (UTC)
 * Red information icon with gradient background.svg Not done: That's not true (in our diaphonemic system). Nardog (talk) 00:31, 22 March 2023 (UTC)
 * The purpose of providing examples is to avoid ambiguity or dialectical/regional variance. American English speakers do not pronounce "dew" as /dju:/, so using it as an example for all English speakers is inappropriate. "Few", on the other hand, is commonly pronounced /fjuː/ across all regions, making it more appropriate for an example here. Thequantaleaper (talk) 01:34, 22 March 2023 (UTC)
 * No, that's rather the strength of the respelling system and dew better illustrates it than few. Those with yod-dropping will read dew as /duː/, those with yod-coalescence as /dʒuː/, and those with neither as /djuː/. And we want them to. See MOS:DIAPHONEMIC for the idea behind this. Nardog (talk) 01:41, 22 March 2023 (UTC)
 * The purpose of these specific examples shouldn't be to illustrate anything but the intended phoneme. If the majority of readers are sounding out 'dew' as /du:/ or /dʒuː/, instead of /dju:/, then they are not benefiting from the example. Rather, they're just getting unnecessarily confused by the discrepancy with the other well-suited examples.
 * Yes, the word 'dew' may very well illustrate the strengths of the respelling system; however, 'few' better illustrates the intended phoneme. The readers who actually need the examples for elucidation (and don't pronounce 'dew' as /dju:/) simply do not benefit from it here. Thequantaleaper (talk) 02:13, 22 March 2023 (UTC)
 * It's not a discrepancy. It illustrates that the diaphoneme /dj/ stands for /dj/, /d/ or /dʒ/ in specific accents. Nardog (talk) 04:09, 22 March 2023 (UTC)

Pronunciation of cure/lure/pure
As there is some dispute about this, we should discuss. IMO, "lure" is a poor [sic] example to use because I'm some dialects it is pronounced l'your and in others as l'oor (like poor). So it seems to me too that "pure" is a better choice. --𝕁𝕄𝔽 (talk) 19:01, 1 July 2023 (UTC)


 * Again, read the text preceding the tables and MOS:DIAPHONEMIC. The diaphonemes //ljʊər// represents "/ljʊər/ or /lʊər/ depending on accent". The whole point of including the word in the table is to illustrate is essentially a shorthand for "equivalent to  for those with yod-dropping and to  for those without it". Nardog (talk) 19:53, 1 July 2023 (UTC)
 * Could you say exactly that at footnote-3? --𝕁𝕄𝔽 (talk) 20:24, 1 July 2023 (UTC)
 * Why? Virtually all other symbols listed have similar variation, but we don't have notes explaining about them. The whole purpose of this respelling system is to provide a simpler alternative to the IPA key that English speakers would find more familiar. So they're going to pronounce the same as however they pronounce the word "lure", which is exactly our aim. There's no need to clutter the tables with notes about variation; we should just direct those interested in deeper discussions about sounds to the IPA key, as we already do. The only notes in this key are all about telling editors what to use in which situations. Nardog (talk) 20:47, 1 July 2023 (UTC)
 * Because as it stands, it has already confused two editors. I for one have never heard the word "lure" pronounced as anything but loor nor the word "cure" pronounced as anything other than c'your (but then again I had never heard "beautiful" pronounced as bootiful until I heard a rural Norfolk accent). It looks like a mistake. But I can appreciate that we could easily drift into trying to reinvent IPA, so maybe there really is no easy answer. --𝕁𝕄𝔽 (talk) 21:09, 1 July 2023 (UTC)
 * Would neural be better? It's another word affected by yod-dropping, which seems more advanced in /lj/ than in /nj/. Nardog (talk) 21:14, 1 July 2023 (UTC)
 * Or the famous Bushism, "Nuclear" pronounced noo-q-lar.
 * I'm wondering if the problem is the use of a pair of words that, although spelled the same, are typically pronounced differently. Are we trying to get one line to do two different tasks? Assuming that the most common use of RESPELL is for a locality name, an average editor will look for sample words that are closest in pronunciation of the name, as they would say it. If there are two words on the same line that are sounded locally very differently, the editor is left confused about what to do or which to choose. [Is there really anywhere that pronounces "cute" as coot.] --𝕁𝕄𝔽 (talk) 23:20, 1 July 2023 (UTC)
 * Nuclear (or nucular—yes, WP:WHAAOE) doesn't contain /jʊər/.
 * That's not a job for respelling (nor IPA in slashes). The strength of respelling is that it automatically maps to each speaker's accent. English orthography simply does not have the wherewithal to convey differences between accents.
 * (Yod-dropping after non-coronals is found in East Anglia, as you've mentioned.) Nardog (talk) 23:32, 1 July 2023 (UTC)
 * I agree with JMF; remove "dew" and "lure" and add footnote. Keep table to clear examples.
 * Might help to have a footnote about how some dialects add or remove a "y" sound in some words. The point of the table is to show the reader "symbol A = sound B" and "symbol C = sound D" and if there is regular diaphonemic variation (e.g. caught) then the reader will adjust how they pronounce sound B or sound D automatically. But that isn't the case for words like "dew" and "lure" they can represent either sound B or sound D.
 * I would be more convinced by Nardog if there were precedence in reputable sources (i.e. dictionaries). Merriam Webster mentions *in the detailed explanatory notes* that some southern English dialects pronounce the "y" sound after l (lewd), s (suit), or z (presume), but they don't put those words in the pronunciation key. Fredlesaltique (talk) 06:31, 2 July 2023 (UTC)
 * I'm going to remove neural *for now* until we agree to include a word that can represent two different sounds in the table. Fredlesaltique (talk) 06:42, 2 July 2023 (UTC)
 * Please read yod-dropping. "Dew" and "lure" are precisely words where "they can represent either sound B or sound D". Respellings like dew, lure are automatically going to be read as [duː, lʊə(r)] by those with yod-dropping and as [djuː, ljʊə(r)] by those without it (and as [dʒuː] by those with yod-coalescence).
 * they don't put those words in the pronunciation key That's because Merriam-Webster specializes in American English. It's trivially easy to find dictionaries that have /Cj/ for BrE and /C/ for AmE in those words:.
 * And the words have been in the table for years. You're disregarding BRD by removing them. Nardog (talk) 07:34, 2 July 2023 (UTC)
 * If the table had just one word, well known to be pronounced differently by locality, there would be no issue. "Neural" would be fine, as its two pronunciations are well known. "Neutral" is perhaps an even better-known example. "Cure" and "lure" are particularly poor examples. --𝕁𝕄𝔽 (talk) 07:50, 2 July 2023 (UTC)
 * No, that wouldn't illustrate the fact yod-dropping does not occur in cute, cure etc. in the vast majority of varieties of English. Nardog (talk) 07:53, 2 July 2023 (UTC)
 * Absolutely. But this is not an article about IPA or linguistics. The object is to advise editors of a word to use as a respelling key. Giving two almost invariant (-ly different) words on the same entry is confusing. Two entries delivers the desired result. --𝕁𝕄𝔽 (talk) 08:16, 2 July 2023 (UTC)
 * I don't follow. The desired result is to have the reader understand that the respellings ew and ure represent a situation that may be affected by yod-dropping/coalescence after coronals but not in other environments. Nardog (talk) 20:37, 2 July 2023 (UTC)

Example
I wonder if a real example would help? Compare Newport, Wales and Newport, Rhode Island. The former retains the yod, the latter drops it. A US visitor to Wales might need to be forewarned that the city is locally pronounced like "pure"; a UK visitor to RI say it like "lure". This is what we need the table to convey. ==𝕁𝕄𝔽 (talk) 09:02, 2 July 2023 (UTC)
 * Maybe a better example: there is a suburb of Dublin called Terenure. The yod is retained, its third syllable is pronounced like "cure", not "lure" (LOOR). The table needs to advise editors how to distinguish. --𝕁𝕄𝔽 (talk) 09:12, 2 July 2023 (UTC)
 * What phonotactic constraints like yod-dropping mean is that the affected speakers are often incapable of undoing the merger, at least without some conscious effort. And the whole point of the key is to illustrate the values of the symbols with words everybody is familiar with. Any word that requires a transcription on Wikipedia would not achieve the goal (WP:LEADPRON). Nardog (talk) 20:36, 2 July 2023 (UTC)

Elephant in the room not mentioned in the preamble
This, the yod dropping phenomenon, is probably one of the best known differences between en.US and en.UK pronunciation ("dew" pronounced doo or d'you). But it is not mentioned in the preamble. The "father" and "farther", "wine" and "whine", and "cot" and "caught" differences are far more subtle. , as this is your area of expertise, do you agree and if so, could you add some explanatory text? --𝕁𝕄𝔽 (talk) 09:45, 2 July 2023 (UTC)


 * Imma be dropping out of this discussion cuz I'm getting pissed off and the lack of reliance on reliable sources is fucking unbelievable. Hopefully someone who has the time and patience can stop this. This table was perfectly fine for years until some editors decided to complicate things a couple years ago because they could. Find a word with consistent pronunciation in actual dictionaries, then include it and change things. Otherwise stop gatekeeping and fucking around with the table. Fredlesaltique (talk) 14:23, 2 July 2023 (UTC)
 * Which dictionaries? Take the simple word "dew". US dictionaries will tell you that it is pronounced like "doo"; UK dictionaries will tell you that it is pronounced like "d'you". Citations for both are trivially easy to find. So we must to recognise the equal validity of both. Let me give you an example. There is a town in England, about 30 miles north of London, called Luton, with London's fourth international airport. A foreign visitor needs to know should they say "LOOTon" or "L'YOUton". Allegedly it has actually happened. --𝕁𝕄𝔽 (talk) 15:46, 2 July 2023 (UTC)
 * I don't see harm in adding it to the three mentioned, though I fail to see how it's any less subtle than the existing ones. Any more than that, I'd be hesitant because it's already covered in more detail by Help:IPA/English, which the text already points to. What do you mean by "some explanatory text"? Nardog (talk) 20:39, 2 July 2023 (UTC)
 * I mean that this question seems to come up repeatedly so it really needs some kind of FAQ to pre-empt this kind of fruitless debate where [I suspect] you to consider it just obvious, read the IPA – but to most editors the IPA may as well be in Greek. As Fredlesaltique has just demonstrated, they get frustrated and walk away. The explanation needs starts from the premise that this is HELP:RESPELL, not HELP:IPA. On that basis, I'm even struggling to see the value of the tables as they stand: they do an excellent job of providing an IPA mapping but so what? For what purpose? Surely what they should be doing is providing a respell syllable?  --𝕁𝕄𝔽 (talk) 23:35, 2 July 2023 (UTC)
 * I'm afraid I'm still not clear on what exactly you mean by "this question" or "this kind of fruitless debate". Do you mean yod-dropping in particular or the diaphonemic nature of the system more broadly? Do you have suggested wording in mind?
 * The respelling system exists as a compromise in response to the complaint that IPA is unintelligible. That's why MOS:PRON says a respelling must always be preceded by the equivalent IPA. (There's a good chance it'll be deprecated once instant IPA-to-audio becomes available here, which I'd welcome.) So readers shouldn't generally need to look at the key (especially if the respelling doesn't contain less intuitive symbols like uu and ow), and editors shouldn't edit respellings unless they're already familiar with our IPA for English. Nardog (talk) 02:03, 4 July 2023 (UTC)
 * Let's return to first principles. The purpose of RESPELL is to indicate how a word is pronounced. Two main uses arise: place-names and unintuitively pronounced 'ordinary' words. The most frequent usage is for place-names, but let's take the easy one first.
 * Ordinary word pronunciations are not, and must not be made to be, dialect sensitive. MOS:PRON absolutely applies, it is unnecessary to distinguish pure/lure, father/farther etc and editors who try to do so are in policy violation.
 * If it is a place-name, there is a "correct" local pronunciation: if you mistakenly put a yod into "Luton", you will not be understood. That is not really true of the two Newports, at most it will cause local amusement. Nevertheless, editors feel that they have to get it right and expect this help article to tell them how to do it. So a means to distinguish between pure and lure is needed. The diphonic policy cannot apply and the preamble to the article is thus misleading. And the line in the table that says pure and lure are the same is just wrong in this context, editors keep challenging it, get told MOS:PRON applies: that is the "fruitless debate". It does not and should not apply here.
 * So the article as it stands is not only not responsive to the primary use, it is arguably dismissive of it. That needs to change IMO. But first let's see if we are any closer to a common understanding of the problem. --𝕁𝕄𝔽 (talk) 11:43, 4 July 2023 (UTC)
 * Luton can only begin with a bare (see Longman Pronunciation Dictionary / Cambridge English Pronouncing Dictionary). It's not an example of . Sol505000 (talk) 13:13, 4 July 2023 (UTC)
 * My point exactly. As I already remarked above, a US visitor needs to know that they should say "LOOTon" (not "L'YOUton"), lest they believe mistakenly that (when in England) they must "correct" their en.us for their native yod-dropping custom. Allegedly it has actually happened. (I don't expect anyone to add a RESPELL to Luton, it is just an illustrative example.) See also the River Thames, where ESL speakers regularly and mistakenly lisp the leading Th. --𝕁𝕄𝔽 (talk) 15:28, 4 July 2023 (UTC)
 * Is it though? is respelled LOO, not LEW. The correct respelling of "Luton" is LOOT-ən (or LOO-tən, dictionaries disagree on the syllabification). This can only be interpreted as lacking . Sol505000 (talk) 18:15, 4 July 2023 (UTC)
 * Again, my point exactly. The table declares (see cure / lure) that LOO and LEW are diphonic and no distinction should be made. Putting words in your mouth I know but your "don't be silly, see any good English dictionary" is exactly what said, except that he started from an American dictionary and, entirely correctly, got a different answer. It is for precisely this reason we have MOS:PRON and rightly so for 'ordinary' words. But not, I argue, for place names. --𝕁𝕄𝔽 (talk) 18:51, 4 July 2023 (UTC)

/r/ and /hw/
The transcriptions of /r/ and /hw/ aren’t all that accurate, and I think they ought to be replaced with a /ɹ/ and /ʍ/ respectively. Luna Netto (talk) 03:59, 22 September 2023 (UTC)


 * this isn't phonetics.  and  are standard. — kwami (talk) 04:39, 22 September 2023 (UTC)
 * And the whole point of this respelling key is the use of familiar symbols. If we start switching to complex ones the average person doesn't understand without looking them up, we'd just switch entirely to IPA and not have these "old-school dictionary"-style respelling keys.  — SMcCandlish ☏ ¢ 😼  05:53, 22 September 2023 (UTC)

Is there a more accessible example for kh?
I thought that loch was pronounced just like "lock" and that in Chanukah, the C was silent and it was pronounced just like Hanukah. I freely admit that I'm probably completely wrong ... but I think the example needs to be something everyone knows how to pronounce, not two words that not everyone who speaks English would know. --B (talk) 23:09, 19 December 2023 (UTC)


 * AFAIK, there is no such word. It seems that any English word with /x/ may be pronounced with /k/ or /h/ instead, and that includes names like 'Bach'. We could say 'like German achtung,' if that helps any. — kwami (talk) 08:35, 20 December 2023 (UTC)
 * That's the point. If a given speaker lacks /x/ in their inventory and utters [k] or [h], that is exactly what is meant by . Nardog (talk) 12:06, 20 December 2023 (UTC)
 * Also, if you pronounce the holiday with an [x], it's not going to matter whether you spell it "Chanukah" or "Hanukkah" (reg. the 'C' being silent). — kwami (talk) 12:15, 20 December 2023 (UTC)
 * As any Scot will tell you, loch is certainly not pronounced just like lock. RESPELL only works for English (in England) and Anglo-American. You can pretend that Loch Lomond is pronounced Lock Lomond, that Lough Neagh is Lock Nay, that Llangollen is Lan-gol-len or even that Cagliari is Cal-e-are-e or even Cag-lee-are-e, but you certainly shouldn't put any of those fake pronunciations into articles. And no, Chanukah is not pronounced Hanukah or even Chan-oo-kah. English does not have gutterals, that's just how it is. --𝕁𝕄𝔽 (talk) 17:50, 20 December 2023 (UTC)
 * So you're arguing that loch and Chanukah are not English words? — kwami (talk) 03:17, 21 December 2023 (UTC)
 * No, I'm arguing that Wikipedia should ⁁not state falsely that lock and hanukah is their correct pronunciation. Maybe somewhere in the body we might report the conventional (mis)pronunciation by English speakers but we must make that qualification clear. The case is significantly different to, for example, Worcester, England. 𝕁𝕄𝔽 (talk) 09:18, 21 December 2023 (UTC)
 * Providing false information is rather at odds with being an encyclopedia. — kwami (talk) 09:24, 21 December 2023 (UTC)
 * Whoops! That'll teach me to be extra careful when editing on mobile. I intended to write "should not state". --𝕁𝕄𝔽 (talk) 10:43, 21 December 2023 (UTC)
 * I was wondering!
 * We'd need to mention something if people weren't consistent. E.g. if Gaelic, Yiddish and German-derived words had /x/ for different people, then the premise of the respelling would collapse. I don't know if the people who have /x/ in loch are the same ones who have it in Hanukkah, and if those are the same ones who have it in Bach. If not, then we don't have the regularity that we count on for respelling. But also the /k/ substitution for loch and Bach is at odds with the /h/ substitution for Hanukkah, so the last at least doesn't fit. But loch and Bach might still work. — kwami (talk) 11:21, 21 December 2023 (UTC)
 * Wikipedia isn't about what's "correct" (to whom?), it's about what's observed in the world. Nardog (talk) 11:27, 21 December 2023 (UTC)


 * I do think adding German achtung would be a good choice if the intent is to explain what /x/ is for speakers who contrast it with /k/ and /h/. Double sharp (talk) 04:41, 21 December 2023 (UTC)
 * That is not the intent, see WP:DIAPHONEMIC. For speakers who don't have /x/, simply means whatever they substitute for it. Nardog (talk) 04:47, 21 December 2023 (UTC)
 * Agreed. I shouldn't have made that suggestion.
 * If you pronounce thy the same as dye, then for you //ð// is pronounced [d]. If you pronounce it the same as vie, then it's [v] -- both common pronunciations. Other words with //ð// will presumably be the same. //x// is no different.
 * As for why comparing it to German is a bad idea, consider the variation of English vowels between dialects. If we were to compare them to other languages, the comparison would only work for speakers of a particular dialect and be misleading for everyone else. Better to say "the KIT vowel" etc., which is essentially what this key does. Same for consonants. — kwami (talk) 04:58, 21 December 2023 (UTC)
 * Well, if that is not the intent, then I agree with . Double sharp (talk) 06:01, 21 December 2023 (UTC)