So it's not a secret that spelling words in English is distinctly sub-optimal, and the American variant slightly more so.
It's irregular, words arn't spelt like they are pronounced, many have silent letters, and ...well, to use a technical term, it's a sack of shit. Fixing this could knock years off the time it takes to learn the basics, freeing up lesson time for other, more useful or interesting stuff.
I've thought for a long time on how the language could be practically rationalised - so that it would actually stick, and not fail like the American attempt. I recently had an epiphany - I wasn't going far enough! In brief, we stop using the "Latin" alphabet, and create a new one based on the International Phonetic Alphabet. Then words can automatically map to the sound they make phonetically.
To be clear, we probably can't just use the standard IPA glyphs directly. They probably need to work as handwriting, be easily discriminated in a variety of font styles, and probably other issues. This is a solvable problem. I'd be happy for linguists to work on the details of how it would work, and the introduction scheme - which would probably have to be spread over a generation or two.
One advantage of this approach is that a great deal of the work is already done - a relatively small set of symbols exist, cover all possible sounds, and are widely used. Some dictionaries spell out words phonetically already. The other- in the end more significant advantage is that this is future-proofed; the IPA is designed to map every sound the human vocal tract can make to a symbol, so although it may need minor tweaking occasionally to track changes in the science, it couldn't drift off course and wreck itself in a sea of contradictions ever again.
The overarching rule is that words should be spelt how they sound, no exceptions. If the way a word sounds changes, so should the spelling, and this can happen organically over time forever.-- Loris, Jan 01 2021 International Phonetic Alphabet https://en.wikipedi...l_Phonetic_Alphabet [Loris, Jan 01 2021] Alternative Option https://en.m.wikipe...ki/Pitman_shorthandPitman Shorthand; alreday complex & phonetic [neutrinos_shadow, Jan 01 2021] //“Let’s all” do something that has been suggested repeatedly for over a century.//
a) It's going to be hard to do /anything/ in spelling which isn't "let's all" on some level.
b) It has? Where? I mean /obviously/ people have been talking about fixing spelling since always - it's just so clearly broken. But have any suggested reforming English completely using a phonetic alphabet derived from IPA?
Well... let's look at the nice list in wikipedia.
a number use the basic English alphabet. Obviously out.
10 "Extend or replace the basic English alphabet". I've eyeballed them all:
Benjamin Franklin's phonetic alphabet - looks like a fairly conservative reform. Not a complete phoneme<=>character mapping Deseret alphabet - "The Deseret alphabet does not have a distinct symbol for the mid central vowel ("schwa")." I know practically nothing about phonetics, but even I can see that's one hell of a flaw in a phonetic alphabet! The Fun Riform - I think this is essentially a 'moderate' reform with diacritics. Initial Teaching Alphabet - "was not intended to be a strictly phonetic transcription of English sounds, or a spelling reform for English as such" - one letter off, nice try. Interspel - very moderate reform, with accents. Apparently intended to be a temporary measure "Until there is a breakthrough to an international script that can cross languages." Iezi Ingglish - not sure on this. Looks like it's aiming to be a sort of compromise between current English and true phonetic english. Perhaps an attempt at phase 1? Musa - phonetic, using a pairing system, presumably representing parts of the vocal system. Claims to be universal. Could work, haven't looked into it. Shavian alphabet (revised version: Quikscript) - looks like a decent attempt at a phonetic alphabet and fairly close in concept, but I don't think it has full IPA coverage; it only covers sounds generically present in English, so would presumably fail on loan-words or as english evolves. Unifon "Unifon attempts to match each of the sounds of spoken English with a single symbol, though not all sounds are distinguished. " : nope
Finally, perhaps the one you were thinking of:
Romic alphabet - The ancestor of IPA, and I think the closest to my proposal by a long way. For all that it's sometimes called the "Romic Reform", I don't think it was proposed as a replacement for English spelling. : "It was intended as an international system of phonetic transcription for oral languages, originally for pedagogical purposes."
So to sum up, you're wrong technically, and also practically, because the one closely IPA-based proposal - I think the only one which aims to have one symbol per phoneme - is so old it's actually a precursor to IPA, and hence hasn't had the years of refinement that IPA has had.
I could honestly live with fully developed derivatives of several of these - but obviously, none of them got legislative support and into the necessary process of being taught in schools.
So, because you missed it (and ended up flailing off at a tangent), I'll try to phrase it briefly and explicitly: Take the IPA phonemic mapping, and standardise spelling using it, so all words are phonetically spelled. Legislation is obviously required, just as it was for decimalisation. Effectively, this needs to be taught in schools (because - duh, it's not going to take otherwise.) This is a more significant long-term project, and will need to be done in stages. Get experts to plan it out, because it's too complicated for a layman to specify in a small text box on a website where ideas are expected to be halfbaked.-- Loris, Jan 01 2021 [Loris] You might be able to do something with High-enjoyment unicode as the character per phoneme you mention.
Make a website where people upvote characters from unicode based on how much people like them. Now that's simultaneously vague and literal:
"on a scale of 1-11 how much do you like this character" (
Then after you have enough for your version of English, You could use people's favorite letter shapes. Think about it, graphically these are the 26-40 characters that just look the most wonderful to people out of many thousands of glyphs. You get to pick what the heart and recycle symbol sound like.
Ok, I'm not such a big fan of this concept either I'm just fond of Angstroms. All the IPA characters are already at Unicode.-- beanangel, Jan 01 2021 I think this idea is based on a number of fundamental misunderstandings, of the nature of and relationship between, written and spoken languages.
Written words don't make a sound. They correspond to a certain range of sounds made by different humans.
In a way this reminds me of discussion of musical notations - the more comprehensive you make your notation system the more useless it becomes for practical use, and in the end you have an audio recording.
Spoken English comes in a wide variety of different standardised dialects which sound quite different from each other and handle different groups of phonemes in different ways. They are usually mutually intelligible but not always.
There intelligable pronunciations even within one dialect.
Consider the words "grass" meaning, a green cereal plant etc.
The "g" sound is fairly stable. The "r" sound is wildly unstable and can be pronounced in any number of different ways from a voicouced "w" kind of sound at the front of the mouth to a gutteral gg sound at the back, and anything inbetween. Then the "a" sound can be pretty much any vowel you can imagine, or can be a dipthong that pushes to wards sounding like "are". Then the "s" sound can range from "sss" to "zzz" to "sh".
I think English is different from other languages in having this consonant stability, while vowels can drift in surprising ways.French, for example, tends in my experience to supress or minimise the importance of consonants while placing a lot more immprtance on very precise vowel form.
Umberto Eco's book "The search for the perfect language" does not directly address the question of synthetic spelling systems for natural languages, but he does deal fairly comprehensively in a very witty and accessible way, the differences between the different forms of a langauge and the fundamental reasons why ad-hoc or artificial language systems are almost always foolish and impossible to implement.-- pocmloc, Jan 01 2021 Reforming the spelling of English has been suggested so many times in the past and it always ignores the fundamental point that English is primarily a written, not spoken language. Thus, the relationships and echoes of meaning between words become apparent in written text rather than speech - e.g. sign/signal and know/knowledge - relationships that would be lost if spelling was changed to reflect pronunciation. If you’re going to go down this road it would be more logical to change English pronunciation to consistently reflect current spelling. This would give you (for a short period of time, before pronunciation changed again through natural evolution) a bit of order but would ruin most historic poetry and literature.-- hippo, Jan 01 2021 Maybe take a step back, & base the symbols not on the sound itself, but how the human vocal tract produces the sound. I think Pitman Shorthand (linky) is sort-of like this. So, a set of symbols corresponding to tongue position, to be combined with symbols for mouth position, & other symbols for breath control, etc etc. It would be more complex, but able to represent any language on earth.-- neutrinos_shadow, Jan 01 2021 Lots of parallel discussions; I'm posting as a block, separating out each one.
~ beanangel
//You might be able to do something with High-enjoyment unicode as the character per phoneme you mention. Make a website where people upvote characters from unicode based on how much people like them.//
I kind of like that. There are languages with some beautiful glyphs in there. But I think this would need to involve experts in the field at a number of stages of the process, to make sure the characters we ended up with were as good a set as possible, and not just attractive in isolation.
~ hippo
//[...] and it always ignores the fundamental point that English is primarily a written, not spoken language. //
People sometimes say that, and I wonder why - because it's just wrong. Protohumans spoke before they wrote, children speak before they read and write (mostly), English as a language was spoken before it was written; individual words are almost ubiquitously spoken before they're ever written down, an authoritative English dictionary didn't even exist until Dr Johnson made one, effectively defining the spellings as he did so. Information is conveyed on the TV (where it theoretically could be either) predominantly in spoken form. English is primarily spoken.
//Thus, the relationships and echoes of meaning between words become apparent in written text rather than speech - e.g. sign/signal and know/knowledge - relationships that would be lost if spelling was changed to reflect pronunciation.//
I don't really get what you're saying here. I imagine that there are many more misleading written matches than the clues you seem to be alluding to. On the other hand, if pronounciation were to be changed to match spelling in some cases to improve clarity I wouldn't be bothered.
//If you’re going to go down this road it would be more logical to change English pronunciation to consistently reflect current spelling. This would give you [...] a bit of order [...]//
Well no, because that would only work~
//for a short period of time, before pronunciation changed again through natural evolution//
and also
//would ruin most historic poetry and literature.//
~ neutrinos_shadow
//Maybe take a step back, & base the symbols not on the sound itself, but how the human vocal tract produces the sound. //
This is what IPA does. I think Musa (I mentioned in a comment above) is doing that basically literally, with a combination scheme, while IPA characters represent the vocal system state in aggregate.
~ kdf
I did mention that an american spelling reformation attempt failed. I must say I'm getting the impression you haven't really read the idea. I humbly suggest that the fact of past failures in an area /probably/ shouldn't be used to rule out the possibility of other attempts out-of-hand. Things are often tried multiple times in different ways before they're massively successful.
~ pocmloc
//Written words [...] correspond to a certain range of sounds made by different humans.//
The thing is - /no they don't/. Not unless you consider it to be a mapping from every individual word to some set of "certain range of sounds made by different humans", that is. The /whole point/ of this is to make it so the word corresponds to the certain range of sounds.
//In a way this reminds me of discussion of musical notations - the more comprehensive you make your notation system the more useless it becomes for practical use, and in the end you have an audio recording.//
It might be worth noting that (as I understand it) IPA has a broad and a narrow set of definitions. Using a broader set is probably the way to go, at least in the basics.
//Spoken English comes in a wide variety of different standardised dialects which sound quite different from each other and handle different groups of phonemes in different ways. They are usually mutually intelligible but not always.//
I'm not denying it. I'm not sure whether being able to easily write in your dialect is an advantage, but it would be possible. At the very least, it could make fiction more immersive.-- Loris, Jan 01 2021 You contradict your self.
//Then words can automatically map to the sound they make//
//to make it so the word corresponds to the certain range of sounds//
So which is it? Does the word unambiguously and directly represent a single pronounciation or a range of possible alternative pronounciations?
If the former, the choice of which becomes arbitrary, as well as appearing confusing or perverse to most people
If the latter, how do you indicate the multiple alternatives?-- pocmloc, Jan 02 2021 [Loris] //I don't really get what you're saying here.// - in the examples I gave and in much of the rest of written English there are important clues in their spelling as to the relationships between words which enrich their meanings - all of which you’re proposing to throw away, just to make written English have a phonetic relationship with spoken English in which, in any case, pronunciation changes over time.-- hippo, Jan 02 2021 // all of which you’re proposing to throw away, just to make written English have a phonetic relationship with spoken English //
"Oldthinkers unbellyfeel Ingsoc ..."-- 8th of 7, Jan 02 2021 //Does the word unambiguously and directly represent a single pronounciation or a range of possible alternative pronounciation?//
Both. As with all analogue samples, there is some inherent range. Apparently IPA has the ability to write broad or narrow transcriptions. The narrow transcription adds an additional layer of detail where it might be necessary.
Analogy: colour descriptions. You can describe a colour broadly, or narrowly. Both are unambiguous, but the latter is more precise. "Green" is clear, but broad. "Avocado green" is a much more specific subset.
Generally, people would write words using a fairly broad transcription, because the point of writing is to easily communicate information, not to document the intricacies of how someone is speaking. If someone mispronounces a word but you knew what they meant, you'd write what you understood. If someone had a speech impediment you'd generally write the words as normal, just as you do now. However, people with one accent might choose to write some words differently to those with another. So for example some people in the uk say bath with a short 'a' and some a longer 'ar' sound. Obviously I can't use non- ascii chars here, which is a bit of an impediment, and also I'm not a linguist, but they might well write the equivalent of b\a\th and b-ar-th.
//If the former, the choice of which becomes arbitrary, as well as appearing confusing or perverse to most people//
No, that describes the current situation.
//If the latter, how do you indicate the multiple alternatives?//
You don't. You write it how it sounds. If you say a word differently, you can write it differently.
//in the examples I gave and in much of the rest of written English there are important clues in their spelling as to the relationships between words which enrich their meanings - all of which you’re proposing to throw away, just to make written English have a phonetic relationship with spoken English//
I guess so, then. I just don't think it's at all important that sign and signal share some letters. Because so does sign and signorina. So does haze and hazel. So does rum, ruminant, rumour, rump and rumple. So does polish and Polish. It seems like one of those egregious lies like "I before E except after C" which people who've memorised all the cases and exceptions tell themselves.
//[...] in which, in any case, pronunciation changes over time.//
But that's the beauty of this scheme. As pronunciation changes, the spellings can also change. Provided the phoneme scheme is taught as the basis of spelling, everyone will adjust automatically.-- Loris, Jan 02 2021 OK so in that case how is this different from IPA?-- pocmloc, Jan 02 2021 Plenty of rocks in the solar system. A new reaction drive is a must. No native languages to muddy the waters, though, the environment will still have a lensing effect.-- wjt, Jan 03 2021 //OK so in that case how is this different from IPA?//
Well, IPA is a phonetic notation system, used as a standardized representation of speech sounds in written form by linguists. This is an idea that involves IPA. I'm proposing a root and branch transformation of written English to a phonetic system derived from IPA, still with its full power but potentially tweaked to improve metrics like writability, legibility, etc. For example some of the glyphs may benefit from being less similar. It probably wouldn't need much, because it's already been heavily worked over, but it's been used predominantly in a scholarly and printed context, and this would take it outside that.
Just as the UK didn't decimalise properly until the decimal system was taught in schools as the primary units of measurement, and those children grown up, so with this. We're now reaping the benefits of the superior measurement system, and the old schemes are only used seriously by pensioners and idiots.
Obviously this would be a larger project than that, so it might have to proceed in generation-long stages, but the benefits would be much more significant, and just as long-lasting.-- Loris, Jan 03 2021 So the idea could be recast as "use IPA instead of Latin alphabet for writing English"
What about other languages? Is there a reason to restrict this to the English language? Most countries using English have other official or semi-official languages, so it would seem odd for them to have two official systems (like selling beer in pints and milk in litres for example)
I'm sure Eco's book has a discussion about optimum numbers of glyphs in a written linguistic system. I have a feeling that IPA has too many characters, and that the Latin alphabet is about right. We can argue about diacritics if you like. But there are issues with increasing or decreasing the number of glyphs.-- pocmloc, Jan 03 2021 //What about other languages? Is there a reason to restrict this to the English language?//
Absolutely. I imagine that the first part taught in primary school would be the phoneme subset required for speaking English, but this would be followed up by the rest (useful for other languages, and rare and loan-words) later in education. As English became regularised, people would naturally put their improved phonetic fluency to good use in their dealing with other languages. Other countries may well observe the benefits and join in, and some uncommon languages might be dragged along for the ride.
//Most countries using English have other official or semi-official languages, so it would seem odd for them to have two official systems (like selling beer in pints and milk in litres for example)//
There would probably be a tipping point. If it did progress in stages, then it might occur by assimilation anyway. Britain apparently has a fairly significant TEFL (teach english as a foreign language) diaspora, which wouldn't hurt.
//I'm sure Eco's book has a discussion about optimum numbers of glyphs in a written linguistic system. I have a feeling that IPA has too many characters, and that the Latin alphabet is about right. We can argue about diacritics if you like. But there are issues with increasing or decreasing the number of glyphs.//
The English~Latin alphabet has a number of wierd features. Why do we even have capital letters? We're practically doubling the number of glyphs for no good reason. Some of them are too similar to each other or other symbols.
I don't know what subset of IPA is needed for English, but it's definitely not all of it, and in any case it seems like a cheat-sheet can fit it on a single page, so it's not /awful/. It looks about the same level of complexity as one Japanese syllabary (about 50 characters) - of which to learn Japanese you need two, plus kanji, which there are thousands. So it's not unmanageable.
I agree though that IPA as it stands doesn't look up to the task of being a basic writing system. Hence the need to modify it first. As I say I'm not an expert, and the right thing to do would be to get a sensible group to work on it in depth. But speaking for myself, I'd try to minimise the amount of out-of-line notation (superscript, subscript, diacritics) necessary for typical use, and make sure the glyphs were as distinct as possible. I think the thing to do would be to work out the common sounds used in English, the same for several other diverse languages, then try to form those up so the commonalities all worked well in combination, then put in the rarities around that.-- Loris, Jan 03 2021 random, halfbakery