English Orthography Reform

[Xonen]

I went to the enemy board

You again! Always with the fraternizing.

and found yet another thread with a spelling reform proposal. See also the linked website.

What do you think?

The same as sangi39, mostly. As a matter of personal taste, there's nothing wrong with wanting to get rid of digraphs or using "Romance" values for vowel letters - but the creator's argument concerning the practical usefulness of such changes isn't very convincing. Also, if the goal is to make the orthography more intuitive for learners, then how come some of the consonants are so weird? <x> for /Z/, seriously?

[Glossaphile]

I came here after being re-directed from "Rite yor wayǃ" in hopes of sharing my own reform scheme, but much to my pleasant surprise, it's already being talked about via the debates on Zompistǃ

To answer this latest critique, the re-assignment of a few familiar letters to not-so-familiar sounds was, as stated on the website, aimed at making better use of traditionally redundant graphemes in order to get the most out of the current 26-glyph alphabet. This keeps the number of unconventional characters to a minimum, which is inevitably a plus when it comes to public persuasion.

I think <x> for /ʒ/ and <q> for /ʔ/ (that second one being admittedly marginal and included almost solely for dialectal flexibility) are really the only two that could really be called counter-intuitive, and both of the two phonemes happen to be comparatively rare, so it works out well.

On the others...

The lone <c> for /tʃ/ should be easy because the lone letter is already a member of the corresponding traditional digraph <ch>.

A <ç> is visually derived from <c>, and its sound (/ʃ/) is also conveniently (and thus mnemonically) related to that of unmarked <c>.

The Spanish import <ñ> for /ŋ/ also bears a visual connection to a the symbol <n> for the closely related sound /n/. One might even think of the tilde as a kind of reduced <g>.

All other consonants are either brand-new (to a modern layman) or pronounced in familiar ways. So, I may have a couple of consonants that take some getting used to, but with most or all other consonants retaining their usual sounds and the re-calibration of several vowels and diphthongs, I suspect that it adds up to a net gain in intuitivity.

Regarding the practicality of restoring a Romance-like vowel/diphthong paradigm, I guess my question is: why not? Assuming for the sake of argument that we all agree to bite the bullet and undertake the arduous process of reform, why not get just a little bit more out of our efforts by making the world's lingua franca a little more universal? It seems like a waste not to, and however slight they may be, there are some practical benefits to learnability by foreigners (as well as to English speakers studying other Roman-script languages). As for the cost to public opinion, I don't think that, all other reform system parameters being equal, the proportional effect of choosing continental over insular vowels would be all that notable.

[Ossicone]

The Spanish import <ñ> for /ŋ/ also bears a visual connection to a the symbol <n> for the closely related sound /n/.

That is incredibly confusing and unnecessary.

[Glossaphile]

What would you prefer, then, Ossicone? Hopefully not the traditional <ng>, since that would be ambiguous.

(/sɪŋə˞/ versus /fɪŋgə˞/)

<singer> versus <finger> - distinction lostǃ

<sìñør> versus <fingør> - distinction retained!

[Xing]

Regarding the practicality of restoring a Romance-like vowel/diphthong paradigm, I guess my question is: why not?

A random thought: Couldn't it perhaps make things worse, to make the pronunciation values of the English vowels more similar to those of Romance languages? The thought is that if language A and language B are similar in some respects, it is more likely that native speakers of, say, language A unconsciously copy structures from that language when they try to speak language B. If language B, on the other hand, is more distant from language A, speakers of language A might be more careful to learn the structures of language B from scratch.

Even if your system might make English the vowel letter values more similar to those of of Romance languages, they are hardly identical - since English simply doesn't have the same vowels as Spanish, Italian or other Romance languages. (Or, an anglophone who copies his English vowels to Spanish would probably speak Spanish with a terrible English accent...)

I don't think this is a very strong objection. As I noted, it more like a random though. In any case, I think the actual benefits of adopting Romance-like vowels in English would be extremely tiny - but it seems like this has been discussed at the ZBB, and there is little reason to repeat the arguments.

Hopefully not the traditional <ng>, since that would be ambiguous.

What would the problem be?

[Ossicone]

What would you prefer, then, Ossicone? Hopefully not the traditional <ng>, since that would be ambiguous.

(/sɪŋə˞/ versus /fɪŋgə˞/)

<singer> versus <finger> - distinction lostǃ

<sìñør> versus <fingør> - distinction retained!

There's nothing wrong with <ng>. The ambiguity you bring up is not really that important since I don't think there any words which contrast [ŋ] and [ŋg].
However, adding your distinction brings up new problems. In my dialect <singer> and <finger> both end with [ŋə˞] and there are other dialects that pronounce both with [ŋgə˞]. In both those cases people would have to have to memorize a spelling which provides no benefit.

[Glossaphile]

There's nothing wrong with <ng>. The ambiguity you bring up is not really that important since I don't think there any words which contrast [ŋ] and [ŋg]. However, adding your distinction brings up new problems. In my dialect <singer> and <finger> both end with [ŋə˞] and there are other dialects that pronounce both with [ŋgə˞]. In both those cases people would have to have to memorize a spelling which provides no benefit.

There will inevitably be a few regional discrepancies of that sort regardless of how /ŋ/ is spelled. If it's not this one that diverges from the local accent, it'll be some other feature. Such is the price we pay for the easy inter-dialectal communication afforded by standardization. Even I know that you can't please everyone all of the time. What we can do is base standard spelling on one or two of the most influential dialects, as even those who don't speak them are likely to hear them frequently (in TV news broadcasts, etc), so they'll always have a ready reference for how to spell words. Just ask yourself, "How would a typical NBC or BBC reporter pronounce this word?"

Regarding the potential crisscrossing of linguistic features due to the orthographies becoming more similar, as suggested by Xing, I can confidently say that my own experience, at least, doesn't really indicate that this would be a noteworthy effect. I've studied both Spanish and Italian, which are about as similar as two separate languages can get, particularly in regards to vowel/diphthong spelling/pronunciation, and I very rarely if ever mix them up.

[ol bofosh]

There's nothing wrong with <ng>. The ambiguity you bring up is not really that important since I don't think there any words which contrast [ŋ] and [ŋg].
However, adding your distinction brings up new problems. In my dialect <singer> and <finger> both end with [ŋə˞] and there are other dialects that pronounce both with [ŋgə˞]. In both those cases people would have to have to memorize a spelling which provides no benefit.

Memorising spellings that have little to do with how we pronounce them is what we do anyway. Even if we adopted a spelling reform based on GA there's be more in common with how I pronounce words than with the current system. In these case of <ng> I do agree that Glossophile really goes into too much detail and could relax some rules a bit. Orthographies will have some ambiguities, not least because pronounciation differs between accents and even individuals.

In Spanish they do it all the time. There's often confusion between <v> and <b> because they share the same sound values, <h> being silent in most dialects can often be left out, and if you have a seseo accent (/s/ instead of /T/) then there can be confusion between <s> and <z>.
Despite the Spanish orthography not agreeing with all Spanish accents, its regularity and relative phonemicity makes it easier to learn. Only the most backwater Spanish accents with little or no access to "prestige" accents would find it most difficult.

This is also, I believe, Glossophiles point in using "prestige" accents to base it on, because RP and GA are accents that most anglophones are exposed to, they are dictionary standards and foreigners usually start learning with one or the other (here in Spain GA (or "Hollywood" for that matter) is very common).

In my dialect <singer> and <finger> both end with [ŋə˞]

Interesting, I never heard of that one. How widespread is it? What with northern England accents saying /ŋg/ in both singer and finger I think that's a good case for keeping them the same. I don't remember as a kid having much trouble with this, so don't think it's the most troublesome feature of English spelling.

[Xonen]

To answer this latest critique, the re-assignment of a few familiar letters to not-so-familiar sounds was, as stated on the website, aimed at making better use of traditionally redundant graphemes in order to get the most out of the current 26-glyph alphabet. This keeps the number of unconventional characters to a minimum, which is inevitably a plus when it comes to public persuasion.

In my opinion, a much better way of keeping unconventional characters to a minimum would be not introducing them in the first place, considering that we already have a bunch of digraphs that do the job just fine. A word like adulthood is so transparently obviously a combination of adult + hood that most people have no trouble at all realizing that the <th> combination is not supposed to be pronounced as a digraph here.

Also, as you point out, your most unintuitive choices concern uncommon sounds... So if you're going to introduce "unconventional" characters anyway, why not use them for these and reserve <x> and <q> for something more common? There are real-world languages out there that use <x> for /ʃ/, for instance.

The lone <c> for /tʃ/ should be easy because the lone letter is already a member of the corresponding traditional digraph <ch>.

I'm not too sure about this. People have internalized completely separate pronunciations for <c> and <ch>; removing the <h> from the latter and expecting people to still associate it with the old <ch> rather than the old <c> would probably work about as well as removing the ascender from <d> and expecting people not to see it as <ɑ>.

In my experience, the notion that <c> is /k/ before a back vowel can be extremely difficult for people to override once they've internalized it (even though - or perhaps because - many of them are probably only subconsciously aware of the whole "before a back vowel" thing in the first place). If I got a nickel for every time I've heard students of eastern European languages (where <c> tends to stand for /ts/ regardless of what follows) pronounce <c> as /k/ despite being way too far in their studies not to have gone over basic pronunciation rules, I'd have... well, perhaps a couple of dollars at this point (the best example might be [lä:tsäkot] for lazacot from one girl after about six months of studying Hungarian; this stuff is hard for some people). So it wouldn't exactly make me rich, but I'd say it's still indicative of a trend.

A <ç> is visually derived from <c>, and its sound (/ʃ/) is also conveniently (and thus mnemonically) related to that of unmarked <c>.

Eh, I guess. Still, if you want a single-letter representation for /ʃ/, I'd suggest something like <š>, which is also conveniently related to a letter with a related sound, and, again, has the advantage of actually already being in use in several real-world orthographies. (Another possibility would be the aforementioned <x>.)

Regarding the practicality of restoring a Romance-like vowel/diphthong paradigm, I guess my question is: why not? Assuming for the sake of argument that we all agree to bite the bullet and undertake the arduous process of reform, why not get just a little bit more out of our efforts by making the world's lingua franca a little more universal?

Because any potential advantage this has would 1) be too small to justify the negatives and 2) only affect new learners - and of them, only second-language speakers. Spelling reforms pretty much always cause difficulty and controversy; assuming for the sake of the sake of the argument that one were possible for English at all, why needlessly increase the difficulty?

Also, as Xing points out, keeping things more distinct can actually help people remember to mind the differences. Despite all the /k/ for <c> and /ts/ for <z> confusion mentioned above, I've never heard anyone pronounce English with Finnish values for the vowels (except as a joke or a quick way to spell out a word).

What would you prefer, then, Ossicone? Hopefully not the traditional <ng>, since that would be ambiguous.

It's not, for the same reason that <th> isn't in adulthood. Also, even if it is to some people, the contrast is so marginal that it hardly warrants introducing a whole new letter (and with an unusual value for said letter, to boot).

[Glossaphile]

In my opinion, a much better way of keeping unconventional characters to a minimum would be not introducing them in the first place, considering that we already have a bunch of digraphs that do the job just fine. A word like adulthood is so transparently obviously a combination of adult + hood that most people have no trouble at all realizing that the <th> combination is not supposed to be pronounced as a digraph here.

You can hear a phoneme. You can't hear a morpheme boundary. When it comes to teaching children, I think we should keep things as concrete as feasible.

Also, as you point out, your most unintuitive choices concern uncommon sounds... So if you're going to introduce "unconventional" characters anyway, why not use them for these and reserve <x> and <q> for something more common? There are real-world languages out there that use <x> for /ʃ/, for instance.

Is voicing that /ʃ/ really that much of a leap, especially considering that a fair number of traditional uses for <x> are voiced, including a few instances where it's just /z/ and not /gz/? We already voice <s> into /z/ all the time, and I don't think chalking it up to morphophonological alternation is a 100% reliable account for that.

People have internalized completely separate pronunciations for <c> and <ch>; removing the <h> from the latter and expecting people to still associate it with the old <ch> rather than the old <c> would probably work about as well as removing the ascender from <d> and expecting people not to see it as <ɑ>.

In my experience, the notion that <c> is /k/ before a back vowel can be extremely difficult for people to override once they've internalized it (even though - or perhaps because - many of them are probably only subconsciously aware of the whole "before a back vowel" thing in the first place). If I got a nickel for every time I've heard students of eastern European languages (where <c> tends to stand for /ts/ regardless of what follows) pronounce <c> as /k/ despite being way too far in their studies not to have gone over basic pronunciation rules, I'd have... well, perhaps a couple of dollars at this point (the best example might be [lä:tsäkot] for lazacot from one girl after about six months of studying Hungarian; this stuff is hard for some people). So it wouldn't exactly make me rich, but I'd say it's still indicative of a trend.

Interesting. I wonder what's unique about <c> that makes it harder to re-learn for some than <z>, which that girl had apparently mastered with its non-English affricate pronunciation.

Still, if you want a single-letter representation for /ʃ/, I'd suggest something like <š>, which is also conveniently related to a letter with a related sound, and, again, has the advantage of actually already being in use in several real-world orthographies. (Another possibility would be the aforementioned <x>.)

The problem with that is that it's not available on the US-International keyboard layout, to which I relegate myself as a sort of compromise for skeptics worried about easy typability. This keyboard layout can be activated within minutes on any Windows PC without any new hardware or software. If reform were ever implemented, it probably wouldn't matter that much in the long run, as manufacturers would probably figure out there's money to be made in upgrading their physical keyboards. But in the initial stages, I think such an appeasing concession to conservatives could be crucial in getting the ball rolling.

Regarding the practicality of restoring a Romance-like vowel/diphthong paradigm, I guess my question is: why not? Assuming for the sake of argument that we all agree to bite the bullet and undertake the arduous process of reform, why not get just a little bit more out of our efforts by making the world's lingua franca a little more universal?

Because any potential advantage this has would 1) be too small to justify the negatives and 2) only affect new learners - and of them, only second-language speakers. Spelling reforms pretty much always cause difficulty and controversy; assuming for the sake of the sake of the argument that one were possible for English at all, why needlessly increase the difficulty?

While I agree that the benefits may not be huge, I don't think the downsides would be all that big either. On the other forum, the point was raised that it's relatively easy to learn new sounds for old letters, so if you accept that, then continental vowels/diphthongs shouldn't seem nearly as disruptive.

Plus, English speakers are not complete strangers to these more Latinate correspondences. They may not be the dominant ones, but they do occur in English, and we don't necessarily have to look at really obscure terminology to find them. Consider "father," "herd," "machine," "or," and "rude" for examples where the vowels are pronounced as they would be in RLS (assuming an RP accent). Most other Roman-script languages, on the other hand, have absolutely no words that uses English-like vowels (except for obvious borrowings). So really, the native English speakers are the ones making the lesser adjustment.

Also, as Xing points out, keeping things more distinct can actually help people remember to mind the differences. Despite all the /k/ for <c> and /ts/ for <z> confusion mentioned above, I've never heard anyone pronounce English with Finnish values for the vowels (except as a joke or a quick way to spell out a word).

But by the same token, if they're similar enough in the first place, the perceived accent or effect of any mistakes decreases significantly. If we assume your view is correct, greater differences mean less frequent but also more egregious mistakes. Smaller differences would mean more frequent but also less severe mistakes. Does either approach have any net edge, then?

What would you prefer, then, Ossicone? Hopefully not the traditional <ng>, since that would be ambiguous.

It's not, for the same reason that <th> isn't in adulthood. Also, even if it is to some people, the contrast is so marginal that it hardly warrants introducing a whole new letter (and with an unusual value for said letter, to boot).

See above. Also, how is /θ/ an unusual value for <þ>? The letter itself may be new to many people, but the phoneme assigned to it would make perfect sense to those few who are already familiar with it.

[Xing]

Consider "father," "herd," "machine," "or," and "rude" for examples where the vowels are pronounced as they would be in RLS (assuming an RP accent).

I have some questions regarding your vowel assignments - as presented on your website.

Would it not make more sense to let the checked <a> represent <æ> rather than /ʌ/? It is true, that in some accents, /ʌ/ and /ɑː/ are distinguished primarily by length - but that's far from universal. In many English dialects, /ʌ/ has a more closed realisation. In some dialects, it's more like a short version of /ɜː/. And then there is of course the northern English dialect area - where /ʌ/ and /ʊ/ are not distinguished at all.

Also, why do you have free <e> representing /ɜː/? AFAIK, this phoneme occurs only (perhaps with a few, marginal exceptions) only before an (underlying) /r/ in the coda.

[Zontas]

I went to the enemy board and found yet another thread with a spelling reform proposal. See also the linked website.

What do you think?

You guys should kiss and makeup already. We've been trying for the past month to remove/ drop down the condescension rate.

We conlangers and conworlders should unite, not form schisms in an already rare hobby.

[Zontas]

And nowwww, the non-awaited part of the X English trilogy...........

Part 3 of 3: Other

•Decapitalization of mid-sentence I, days of the week, and sentence-initial word-initial consonants.
•Dedoubling of consonants word-finally after a vowel (optional).
•Etymological respelling:

• buy > bye > bie
• buoy> buey
• could > cood
• die (v.)> dye > diey
• eke > eak
• eye> ieye
• lye > liey
• mote (v.) > moet
• newt > neut
• ptarmigan > tarmiggan
• sastruga > zastruga
• sky > skie
• weigh > waih
• you > yu
• your > yor

[Glossaphile]

Would it not make more sense to let the checked <a> represent <æ> rather than /ʌ/? It is true, that in some accents, /ʌ/ and /ɑː/ are distinguished primarily by length - but that's far from universal. In many English dialects, /ʌ/ has a more closed realisation. In some dialects, it's more like a short version of /ɜː/. And then there is of course the northern English dialect area - where /ʌ/ and /ʊ/ are not distinguished at all.

There are a couple of reasons for that assignment. First, in my experience, speakers of other European languages (particularly Romance) who can't quite master /ʌ/ seem to use [a] (or something close) as their approximation. Secondly, if I let checked <a> represent /æ/, it leaves /ʌ/ in need of a symbol of its own, and using <æ> for that instead just seemed like too much of a stretch even for me. Most if not all other options were unavailable on the US-International keyboard, too bizarre even for my tastes, or both. I made a judgment call, basically, and this seemed to be the all-around best solution.

Also, why do you have free <e> representing /ɜː/? AFAIK, this phoneme occurs only (perhaps with a few, marginal exceptions) only before an (underlying) /r/ in the coda.

The more obvious choice, which I presume would have been /eɪ/, is a diphthong, and one of my firmest design principles was that of calling a spade a spade when it came to monophthongs versus diphthongs. Incidentally, with that principle in place, you wouldn't necessarily need an explicit Romanized-vowels policy to arrive at something very similar to what I did, because four out of the five so-called "long vowels" in the traditional system are actually diphthongs and would thus have to be reassigned on that basis anyway.

Also, assigning /ɜː/ to free <e> contributes to the dose of dialectal flexibility of RLS, since one can easily represent both rhotic and non-rhotic forms without changing any rules or even resorting to the Non-Standard Dialect Representation Toolkit (a peripheral repertoire of optional-use symbols for a few well-known dialects besides GA and RP, such as Australian, Canadian, Southeast American, and even Cockney). Consider, for instance, <feðø> /fɜːðə/ and <férðør> for /fɜ˞ðə˞/.

Thanks for your questions! I appreciate you showing enough interest to ask them!

[Xing]

You can hear a phoneme. You can't hear a morpheme boundary.

In that case, you can't hear a phoneme either. You hear streams of noise. Phonemes are abstract entities that our brains use to make sense of the streams of noise reaching our ears.

There are a couple of reasons for that assignment. First, in my experience, speakers of other European languages (particularly Romance) who can't quite master /ʌ/ seem to use [a] (or something close) as their approximation.

In my experience, speakers who have not yet mastered English pronunciation often merge various vowel phonemes (since English has an unusually large vowel system). And further, it has been a trend many English dialects that /æ/ lowers towards [a]. Not writing [a] with <a> would therefore make English spelling less similar to that of the Romance languages - and thus violate one of your design goals.

And dos speakers of Romance languages usually identify <e> with /ɜː/?

Secondly, if I let checked <a> represent /æ/, it leaves /ʌ/ in need of a symbol of its own, and using <æ> for that instead just seemed like too much of a stretch even for me. Most if not all other options were unavailable on the US-International keyboard, too bizarre even for my tastes, or both. I made a judgment call, basically, and this seemed to be the all-around best solution.

I don't se why <æ> for /ʌ/ would be "too much of a stretch", at least not as compared to many of your other suggestions. I mean, if you can use individual <e> and <a> for schwa-like vowels, why can't you use a ligature of them for the same purpose?

If the spelling reform is meant to be global, have you checked other keyboard layouts than the "US-International" one, for what they can and cannot easily type?

The more obvious choice, which I presume would have been /eɪ/, is a diphthong, and one of my firmest design principles was that of calling a spade a spade when it came to monophthongs versus diphthongs. Incidentally, with that principle in place, you wouldn't necessarily need an explicit Romanized-vowels policy to arrive at something very similar to what I did, because four out of the five so-called "long vowels" in the traditional system are actually diphthongs and would thus have to be reassigned on that basis anyway.

Whether /eɪ/ is a diphthong or a monophthong is highly dialect dependent. I pronounce it as a diphthong - something like [ɛɪ̯] - but it's not uncommon to pronounce it as a monophthong either. The same is true for /oʊ̯/ - which have both monophthongal and diphthongal realisations. And what about /iː/ and /uː/? Those are also diphthongs for many English speakers. (This might become clear if you compare them with the purer monophthongal i's and u's of many other languages.)

And by the same logic, shouldn't something similar apply to affricates? (They should have one glyph indicating the stop phase, and one indicating the frication phase?)

[Salmoneus]

I've given up trying to argue with you on the thing in general, but I am puzzled over one particular detail: " Consider, for instance, <feðø> /fɜːðə/ and <férðør> for /fɜ˞ðə˞/." Huh? What words are these meant to be? "fairther"? There's no such word.

[Click]

Further?

[Xing]

Some other questions regarding pre-rhotic vowels. How would you spell:

Mary
merry
marry

furry
hurry
ferry

sorry
lorry
story

[sangi39]

I get the feeling that Sal was indicating that he wouldn't immediately associate <e> with /ɜː/.

On a similar note, and to go back to something Xing asked, i.e. "And do speakers of Romance languages usually identify <e> with /ɜː/?", although this doesn't cover romance speakers, but rather Polish ones I've worked with and a Lithuanian guy I went to university with, they seem to pronounce /ɜː(r)/ as something closer to [u:r], and the German speakers I know pronounce it as if it were <ör>. They never seem to handle <er> as /ɜː(r)/ and almost exclusively pronounce it as /e(:)r/ in isolation.

As others have pointed out, /ɜ:(r)/ distribution is fairly predictable, deriving from older /ɛr ɪr ʊr/. Some dialects, typically found in Scotland, maintain the distinction, while others further merge this vowel with the /jɜ:(r)/ of words like "cure" through yod-dropping wihile other merge it with an entirely different vowel, notably /eə(r)/ in Liverpool.

[ol bofosh]

A rough guideline to my girlfriend's pronounciation. It's to show that learning to spell/pronounce well isn't always necessary to speak English, and that spelling doesn't have to be highly detailed.
The sets below aren't predictable according to native English accents as she has a few spelling pronounciations, has influence from French and Spanish and stress patterns differ.

TRAP-PALM-BATH /ä/
DRESS /e/ (possibly /E/ as in French)
KIT-FLEECE-happY /i/ (with some advice some FLEECE words are lengthened so she now distinguishes sheet and shit)
CLOTH-LOT /o/
FOOT-GOOSE /u/ (I've not noticed difference in length, if there is, it's incidental)
STRUT /V/ (more or less, at least she distinguishes it from /{/ or /A/ words, unlike the Spanish)
THOUGHT - /O/
commA /@/ (usually words spelt with <a> are not reduced and remain /ä/; French influence perhaps)
PRICE /äj/
MOUTH /äw/
FACE /ej/
GOAT /ow/
oi /oj/
START /är/
SQUARE /er/
NEAR /ir/
CURE /ur/
NURSE-lettER /@r/