Talk:Sound correspondences between English accents

Table into tables

After spending far too long adding to the proposed redraft of this page (though only really moving information from this page to that, rather than adding undetailed mergers etc.), I've come to the conclusion that it actually is not as desirable as we originally agreed, and that a few changes to the current version would be preferable.

This comes after my recent edits to the current version of this page, in which I split the table up into groups of merged sets (that were already apparent in it) which have made it easier to read (because mergers are confined to one table) and to edit (because smaller tables are easier to work with).

At present, the tables are still laid out horizontally rather than vertically, but I intend on changing this as soon as I can. With that said, most edits that were previously not possible are now, although two of the tables' furthest cells are still too far across to edit at present. Vertical orientation doesn't allow you to compare the values within one set quite as well, but that's a worthy sacrifice for being easy to read, and this page is about correspondences anyway.

Having made that change, I think that a lot of the problems we identified in January can actually be easily overcome, and as such I think that the proposed redraft discussed above is now not worth pursuing. This is mainly for two reasons:

• The article will be easier to edit as the use of the template would be more limited, there will be fewer tables, and correspondences are more intuitively shown on a table with multiple sets.
• The article will be easier to read as it very clearly shows the relationships between multiple sets at a glance, rather than having to explicitly think in terms of splits and mergers (though we can still detail this).

Below, I've compiled the advantages of the proposed redraft and written up ways in which we can make up for the loss of them in the current version of the article.

• The previously agreed-upon redraft...
  • can show more than two mergers for each lexical set: "KIT should reflect its mergers with commA, happY and DRESS but it can only go next to two of them".
    • I propose having more than one KIT table in this situation, organised so that more similar sounds are together (e.g. happY "goes with" KIT better than DRESS in the context of the happY-FLEECE merger). This would create redundancy, so perhaps it's worth keeping the template for such cases, even though it may be slightly harder for uninitiated users to use. Deciding where to draw the lines may be slightly difficult, but I think that as long as we clearly signpost at each table that KIT also has other mergers, this isn't really a problem.
  • allows you to more easily compare the presence of a split or a merger across accents, which is useful for readers who want to compare phoneme boundaries rather than just realizations.
    • this is still relatively easy to do by looking at the lines in the tables
  • includes information on the direction of mergers, which is information that cannot be easily worked into the current version.
    • interesting, but not that important (history of mergers are detailed on their own pages)
  • has recognition that Wells' sets are flawed baked into it; Allophonic splits/Phonemic splits/Into Wells' sets is a useful distinction that is not as clear in the current version.
    • could be implemented by adding a new row/column that details this
  • explicitly names the splits and mergers which take place.
    • could be replaced by linking or notes
  • includes information on the extent to which mergers take place in specific accents.
    • could be replaced by notes
  • lays out the sets in an order which makes them much easier to find (as well as including links to each in the sidebar).
    • quickly finding a certain set may actually be easier in the current layout of the page, since they are naturally grouped by the kind of vowels they are.
  • has a short paragraph explaining each set's system.
    • mostly redundant because that information is contained in each table.
  • is more colo(u)rful, which makes for a more appealing article.
    • I think we might just about live without that.

In my opinion, the few advantages that we lose are not worth the simplicity and intuitive nature of the current style of table. As always, I'd appreciate your thoughts (anyone, but especially @Offa29 given their previous involvement). Thank you! ^{[citation unneeded]} (talk) 17:03, 21 July 2025 (UTC)

I preferred the single large table however I also appreciate your use of clear columns for denoting splits and mergers. I do like having each set listed on its own. May I suggest also adding Parry's FIRE & HOUR sets where possible? I know they were meant for Welsh English yet they have broader significance as the former distinguishes 'hire' (1 syllable) from 'higher' (2 syllables) and the latter distinguishes 'flour' from 'flower' in the same way. 1.126.105.202 (talk) 13:32, 7 August 2025 (UTC)

After having thought about it for a bit longer, I agree that it's better to have each set listed on its own. While the individual 'runs' of sets work for the existing table, this excludes many, many mergers which simply cannot be show like this. I had a go at trying to write out what a complete version of this layout would look like, and it's a horrible mess that's difficult to navigate. I think I was right in the first place when suggesting the current layout for the Draft article.

The superficial neatness of this layout, while pretty at the moment, does not hold up to expansion.

I agree with you about FIRE and HOUR (@Offa29 said something similar actually). ^{[citation unneeded]} (talk) 20:55, 7 August 2025 (UTC)

Is Not Newfoundland Part of Canada?

Why are all regions of all countries listed together except for Newfoundland? Is Newfoundland a part of Canada or an independent country? If it is a part of Canada, then it should be positioned in the Canadian section. The remaining part labelled as "CaE" (Canadian English) could easily be relabelled as "General Canadian English", or something similar, to make the difference clear. However, promoting secessionist movements in Canada in this way is not more appropriate than listing California separately from the rest of the United States of American for the same reason would be. 1.126.105.202 (talk) 13:24, 7 August 2025 (UTC)

The way regions are laid out is also another problem with this article. I can't believe no-one noticed this one though - maybe they did but it was too difficult to edit before the split into tables. It should be called Standard Canadian English.

I've now drawn up an example table of a new order here. If no-one has any problems with it I'll start reordering the Draft article to fit it, but please let me know if it can be improved! ^{[citation unneeded]} (talk) 22:17, 7 August 2025 (UTC)

Inaccurate Claim

The statement "The lexical sets FACE, DRESS, GOAT, MOUTH, and START do not merge with other sets" regarding the so-called "Unmerged sets" is not entirely factual. For example, in accents where the pin-pen merger exists, there is a partial merger of the DRESS set with the KIT set. It would be more accurate to say they do not completely or fully merge with another any other single set. 1.126.105.202 (talk) 14:09, 7 August 2025 (UTC)

Yeah this is also a symptom of just using the information already in the article. I've renamed it to "Sets not nmerged here", but feel free to change it if you think of something better. Let's not tell lies-to-children. ^{[citation unneeded]} (talk) 22:22, 7 August 2025 (UTC)

All Sets Can Merge

As the tables already show, it is possible for the following lexical sets to merge with other sets. For example:

• DRESS may completely merge with TRAP in Singaporean Standard English.
• FACE may merge with HAPPY in the Scottish English.
• GOAT may merge with THOUGHT in various North England Englishes.
• MOUTH may merge with PRICE in General South African English.
• START may completely merge with TRAP, BATH and PALM in Cameroonian English.

Consequently, the following demonstrably incorrect claims must be removed:

• "The groups are defined such that no mergers of each group's sets take place outside them."
• "The lexical sets FACE, DRESS, GOAT, MOUTH, and START do not merge with other sets."

1.126.107.52 (talk) 19:40, 26 August 2025 (UTC)

I've included these in the article along with a more detailed description of why they are not included in the tables. Basically, it boils down to this article being conceptually flawed because the most mergers that can be shown for each set is two. I'm currently working on a draft version that would fix this problem, but it's quite slow going so it looks like this is the best we can really do at the moment. ^{[citation unneeded]} (talk) 10:52, 7 October 2025 (UTC)

The phoneme /u/

Many dictionaries also contain the phoneme /u/ as in "situation":

• https://www.oxfordlearnersdictionaries.com/definition/english/situation?q=situation

I would be glad if it were added to the "STRUT to GOOSE" section. Many thanks in advance 185.18.68.210 (talk) 10:22, 6 October 2025 (UTC)

The /u/ phoneme (as in situation) is already present in the "STRUT to GOOSE" section as GOOSE. Please let me know if I've misunderstood you! ^{[citation unneeded]} (talk) 21:54, 6 October 2025 (UTC)

Hi! Thanks for your reply. However, I can't find the /u/ phoneme (as in situation) in the "STRUT to GOOSE" section. GOOSE shows /uː/ not /u/ 185.18.68.210 (talk) 21:53, 14 October 2025 (UTC)

It's not a phoneme. See . Nardog (talk) 01:30, 7 October 2025 (UTC)

Hello! Thanks for your response. You could argue that it's not a phoneme, but you could say exactly the same about /i/, which is included in the table as "happY". See: https://www.oxfordlearnersdictionaries.com/about/english/pronunciation_english

So, I think that both /i/ and /u/ should be treated in the same way. If /i/ exists in the table, it's reasonable to ask why /u/ doesn't 185.18.68.210 (talk) 21:56, 14 October 2025 (UTC)

/i/ is not a phoneme either. One could say it's an archiphoneme, but in reality it's just a conventional shorthand for "/iː/ or /ɪ/ depending on accent". See happy tensing. Nardog (talk) 13:35, 15 October 2025 (UTC)

But /i/ is in the table in this article. So either it should be removed or /u/ should be added as well. This is called consistency 185.18.68.210 (talk) 15:19, 15 October 2025 (UTC)

I've had a deeper look and I think that the main reason this article has /i/ and not /u/ is that Wells defined happY as a lexical set but didn't do the same for sitUation/actUal, and this article uses his sets as its basis. Neither /i/ nor /u/ are actually phonemes in their own right because, though they occasionally have independent realizations (see CaE, Cockney, Cumbrian, Pitmatic, Scouse, all forms of Ulster English, Newfoundland English etc. for happY in the table), their definition is purely based on environment, so they aren't phonemes but positional allophones.

When we talk about the 'happY lexical set', we really mean 'the words containing a positional allophone of the KIT lexical set' (in the positions outlined here and here), and when we might have talked about the actUal lexical set, we really mean 'the words containing a positional allophone of the GOOSE lexical set'. happY sometimes gains the realization of FLEECE, just as actUal sometimes gains the realization of FOOT, but they don't represent phonemes any more than think (a positional allophone of KIT which gains the realisation of TRAP in some accents) represents a phoneme.

However, if I rewrote the previous paragraph using BATH for happy, TRAP for KIT and PALM for FLEECE, I would be describing the process that led to the BATH lexical set, which does describe a phoneme in accents with the TRAP-BATH split. While this might seem to be a contradiction, the difference here is that the version of the BATH set that became static can't be easily described using a rule (gas [ɡæs] vs glass [ɡlɑːs], mass [mæs] vs pass [pɑːs], is it [plæstɪk] or [plɑːstɪk]?).

I'd appreciate some feedback here, because I'm not entirely certain, though I think that's along the right lines. The next question to ask seems to be 'should we get rid of happY?' but I feel that there's little point because it does convey useful information and is true to Wells' system. Perhaps a note explaining that happY doesn't actually describe a phoneme would be a good idea? If you can find some good sources describing whether /u/ uses /uː/ or /ʊ/ in lots of different accents, then it might be worth adding, but there's also some appeal to sticking to Wells' original sets because that's what people are familiar with. ^{[citation unneeded]} (talk) 15:17, 15 October 2025 (UTC)

Okay, I get your perspective now. Still, I wonder how /u/ is realised in modern RP. As /uː/ (as in GOOSE) or /ʊ/ (as in FOOT)? Do you know the answer perhaps? 185.18.68.210 (talk) 15:28, 15 October 2025 (UTC)

As an SSBE speaker, I can confirm that it's pronounced using the GOOSE vowel [ʉw] rather than the FOOT vowel [ɵ], at least in my ideolect. ^{[citation unneeded]} (talk) 17:26, 15 October 2025 (UTC)

Table design

While I think the current layout of information is pretty good, I think that the "Lexical set" and "Examples" columns should not disappear when scrolling horizontally in the table. The examples especially are essential to show what contexts certain allophones are used. As it stands, only the diaphoneme column remains visible. FitzCorb (talk) 11:42, 30 March 2026 (UTC)

This is something we've fixed in the draft revision of this article! ^{[citation unneeded]} (talk) 12:24, 31 March 2026 (UTC)

Oh I'm quite pleased to see that there is indeed an overhaul planned! I started off by wanting to fiddle with the abbreviations and the ordering of them, but I realized it would get more complicated from there. So I'll leave that be and in the meantime, I'll just make the change for abbreviations without changing the ordering.

• ScE → SSE (Scottish Standard English) per the Scobbie, Gordeeva & Matthews source that is cited the most for Scottish English in this article
• SSE → SgE (Singapore English) since most sources don't abbreviate it (and also don't specify that it's the standard variety) – and this one has it as SgE.

This also appears to be consistent with general practice among sources. I could not find much usage of ScE online except to mean Scots. ‍ –‍ Kilvin77 • 👾  05:53, 17 May 2026 (UTC)

Changing the consonant table

It would be more useful to change the consontant table so that it would similar to the vovel charts. So first Diaphonemes, then Examples and then each variety of English or the varieties we know and then just a general 'others' column for these varieries we don't know enough ~2026-22093-04 (talk) 07:41, 10 April 2026 (UTC)