I must annoy the nieces no end.

[conuly]

They can't ask me the definition of any word and just get it, I have to take them through the etymology of it as well. (Plus I make them guess from context FIRST, and usually do charades as well. I'm easily entertained.)

My ultimate goal is to convince them to stop asking me, but I'll accept "educate them" as a (distant!) second answer. So when Evangeline asked me why I called "that thing" a "structure" I took her through CONstruction and DEstruction and INstruction as well. (One of the few concrete benefits of taking Latin, however poorly I may have done at that, is that I can rattle this stuff off without blinking.)

I also am prone to doing things like this if they ask me to spell stuff for them and telling them "sheesh, sound it out already" (my default answer - we're not supposed to tell them how to spell things according to the schools) isn't likely to work, I run through WHY it has the weird spelling it has (if, indeed, it has a weird spelling).

I'd rather have a sensible orthography, but that's not likely.

And what really bugs me beyond belief is the argument that if we had a sensible, reasonable, rational orthography we'd somehow lose all knowledge of etymology! It's a silly argument to begin with, but it's made even sillier when nobody (well, almost nobody) teaches this to kids to begin with! (Few people really grasp it even as adults, apparently, which is just sad, but that's beside the point.)

I browsed a list of tips for teaching unintuitive spellings the other day, and one of them was about using mnemonics. Well, I can go with that - but the example given was of a teacher who told her students that "grammar comes from Mars".

And that just bugged me. Why not tell them that it's related to the word grammatical (which it is, and also glamor and grimoire, the root concept for all of these being "learning", but neither of those words really is helpful in this instance, and I have the feeling it's a different sort of related anyway), which is equally mnemonic and also teaches them something useful? (That this has to be taught strikes me as strange, but if it were obvious people wouldn't get it wrong so often, would they?)

Interestingly, the case of grammar indicates another issue with spelling reform that opponents never ever mention, the question of whether we'd do everything totally phonemically (for whatever dialect we'd just have to pick or invent to do it all in) or whether we'd do it morpheme by morpheme. The first has the advantage of being really easy to spell and read, the second has the advantage of keeping similar spellings for words that vary only according to suffix (so grammar/grammatical would start off the same way, just like they do now, even though they sound like they have a different vowel.)

Opponents of spelling reform, though, hardly ever seem to have any good arguments. I've noticed that. I don't know why that should be, but I've noticed it. It's not fair that I should have to argue their side as well! (It's probably because it's never gonna happen, so they don't have to bother. But it's still laziness.)

Comments

[rantinan.livejournal.com]

You wont get any arguments from me regarding spelling reformation. Oh, to be able to spell phonetically and just have to worry about getting the letters in the wrong order would be swell.

[pne.livejournal.com]

the word grammatical (which it is, and also glamor and grimoire, the root concept for all of these being "learning"

It is? I would have thought it goes back to Greek γράφω graphô "write" / γράμμα, γράμματ- gramma, grammat- "something written; a letter [of the alphabet] / a letter [an epistle]" - which I think go back to something like "mark, incise, score" even earlier.

*etymonline* OK, I was just thinking further back than you.

And it seems that γράφω is related to "carve", which still preserves the old meaning. Whee!

[conuly]

Yeah, I refuse to go back as far as Greek :)

[pne.livejournal.com]

It's not like I made you go back to reconstructed Proto-Indo-European :)

I guess I tend to think of Greek because I learned modern Greek, so cognates often jump out at me. (I imagine that since you learned Latin, you'd think of Latin roots as well, rather than just - say - Middle French.)

[pne.livejournal.com]

Interestingly, the case of grammar indicates another issue with spelling reform that opponents never ever mention, the question of whether we'd do everything totally phonemically (for whatever dialect we'd just have to pick or invent to do it all in) or whether we'd do it morpheme by morpheme. The first has the advantage of being really easy to spell and read, the second has the advantage of keeping similar spellings for words that vary only according to suffix (so grammar/grammatical would start off the same way, just like they do now, even though they sound like they have a different vowel.)

I think #2 is called "morphophonemical" (or "morphophonological"? something like that).

It's more or less what Korean does, as I understand it.

(For example, bada "sea" is spelled "ba-da" while bada-, the stem of "receive" [IIRC] is spelled "bad-a" even though it's pronounced the same, because the -d- belongs to the stem. Or eoptta is spelled eobs-da even though the -s- is not pronounced, but it is pronounced in other forms of the verb such as eobseoyo, spelled eobs-eo-yo: you can see the common eobs-.)

I'd argue for going the whole hog and doing phonemic spelling (your #1); then people will simply learn to recognise typical sound alternations between related words such as "vain-vanity, serene-serenity" etc. Much like Finns learn to recognise typical sound alternations (http://en.wikipedia.org/wiki/Consonant_gradation) in inflection forms such as pp/p, kk/k, tt/t, k/v, k/j, k/ng, t/d, etc.

[conuly]

I also think phonemic spelling is the way to go. If people can't put together word pairs on their own, that's not my problem.

But there is an advantage to the other system, and it's worth bringing up even if I think the advantage doesn't outweigh the HUGE advantages of a more straight-forward orthography.

Also, thankies for the information :)

[redbird]

The problem with spelling phonetically is, what accent. I used to know someone who spelled phonetically with a Missouri accent, and I found it a lot slower to sound out than standard English spelling. If I'd also been dealing with other people doing the same in five or six other accents, including say Edinburgh, Boston, Mumbai, Jamaica, and Melbourne, it could have been downright painful.

English isn't in anything like the situation of the Chinese ideograms, but we are using the same set of spellings for a lot of accents and even dialects. As is, I can edit a book for Texas nine-year-olds, and not worry about how they pronounce the words; that's for them, their teachers, and their parents.

[conuly]

Ah-ha! I've already tackled that one! BEHOLD, the spelling reform post of DOOM!

To sum up:

1. It doesn't really matter for most things "what accent". Almost everything follows consistent sound change rules. So if you say "I" in the same way that I say "ah", so long as we both agree that the letter representing that sound (whatever sound it is!) is the same, we can both consistently spell the same way.

1a. There is some trickiness when it comes to mergers and splits (where one dialect has more sounds than another), we'd have to decide whether to make it easy to spell or easy to write. We also have some trouble where one word has two pronunciations and it does NOT follow a consistent rule, like the word "orange". However, this leads to point 2.

2. In cases like that, we'd have to just pick a method and stick with it. Make one pronunciation the standard and let it go. This would inconvenience many people, but no more than "two/to/too" already does. (I would suggest that we simply invent a new standard, using the old standard as a base but making sure, in cases of disagreement, to pick the option preferred by most speakers as much as possible. So if most people say sorry with the same vowel as in or, go with that. If most people say orange with the same vowel as in are, go with that. If it's about even, flip a coin.)

[pne.livejournal.com]

Argh, I would hate having to spell in an orthography based on an accent that makes "weird" mergers.

But you know what? That's me speaking as someone who's been literate in standard spelling for about thirty years.

If I had grown up writing phonemic-for-General-American, I'd be used to that. And the fact that it doesn't differentiate between some vowels that I keep separate wouldn't be qualitatively different from the fact that th is ambiguous between the sound in "thy" and the sound in "thigh", or that oo is ambiguous between the sound in "foot" and the sound in "boot", or .....

I'd just have different ambiguities (probably father-bother and/or cot-caught mergers, possibly Mary-merry-marry) in my spelling.

I'm sure I'd survive.

[conuly]

Any spelling reform would annoy and aggravate a huge segment of the population used to the old spelling. (Well, of the literate population. Plenty of people already can't spell.)

And yeah, nothing could ever be perfect for everybody. But! We can definitely improve off what we already have!

[pne.livejournal.com]

Any spelling reform would annoy and aggravate a huge segment of the population used to the old spelling.

True.

But if it's done properly, that'll be two or so generations who'll be inconvenienced, and the youngest will simply grow up with the new system.

And reading documents in old orthography will look weird and provincial, I suppose, but shouldn't be insurmountable.

What I'm worried most about is getting a half-baked spelling reform, especially if it's due to committee politics, because if we ever got around to reforming the spelling, we'll probably only have one chance, so it's important to get it, if not perfect, then at least decent, or you cause more problems than it's worth.

Funnily enough, you guys on your side of the pond already have the benefit of a (minor) spelling reform/simplification/regularisation! You have "draft, color, theater, program" where we have "draught, colour, theatre, programme", etc., where your spellings match the pronunciations better than ours.

[ihcoyc.livejournal.com]

Because English spelling is etymological rather than phonemic, it has the advantage that foreigners can be taught to read and write it with very little adjustments. The Latinate language of scholarly prose moves the feature forward; master a handful of grammatical function words and the jargon of a field, and foreigners can produce text like that fairly easily without acquiring the ability to understand most spoken English.

Our promiscuous lexicon makes back-of-the-mind etymology necessary for comprehension. Próduce the noun is obviously related to the verb prodúce; a phonemic spelling would indicate which vowel is stressed and which reduced in each of the two forms, obscuring the relationship. We have dozens of basic but borrowed words that work similarly.

[conuly]

I don't think that giving an advantage to foreigners is worth giving a severe disadvantage to beginning readers.

Also, we wouldn't necessarily have to indicate stress even with a phonemic orthography, nor indicate that vowels reduce (that's an automatic feature of the language, and one morphological thing I'm okay with). However, even if we did, it would not really obscure anything. Both forms sound pretty similar to me, likewise present and present sound similar even though the stress and first vowel are different and so on.

[pne.livejournal.com]

One thing I really liked about Greece from the point of view of a foreigner is that it marks the stress on all words of more than one syllable.

But it depends on what the optimisation goals of the spelling reform are: make it easier for foreigners, or for native speakers.

Native speakers nearly always know where the stress goes anyway, except where the placement of stress makes a difference. So the question would be whether to mark those exceptions, or whether not marking it is enough.

For what it's worth, the Shaw alphabet does mark reductions to shwa, and in doing so, implicitly marks stress in many cases. Thus, doublets such as perfect/perfect and present/present would be unambiguous. (It could be argued that the STRUT sound is not a separate phoneme but merely the stressed allophone of the COMMA souns. Shavian treats them separately; similarly with the rhotacised versions NURSE and LETTER, respectively.)

[conuly]

Yeah, but firstly then you have the debate of "mark schwas only, or also schwis", which apparently some people don't have, and secondly you have the fringe groups declaring that it's all a plot by mad linguists to encourage lazy speaking.

[pne.livejournal.com]

Hopefully any such debates would be held in a circle of competent experts (probably mostly linguists, but presumably also some teachers or other pedagogues for input on what's easily teachable and what's tough) that's as small as possible, and the result presented to the public as "this is it". Then whatever that group comes up with, we use, whether that means shwis or not.

I think the value of *a* decent (even if not 100% "perfect", whatever that means) standardised phonemic orthography are greater than continually tinkering with it in order to achieve perfection *or* simply doing nothing.

[conuly]

(I mostly made that comment as an excuse to use the word schwi. It's really just fun to say, and funner to type!)

Also, apparently "funner" counts as a word for Spellcheck. Ashkenazi doesn't, but funner does. Go figure.

[janewilliams20.livejournal.com]

Having got part-way through the comments in said Post of Doom, I see much discussion of the problem that immediately came to my mind: some accents treat certain sounds as identical when in others they're entirely different, and some treat letters as silent when others don't. You mention vowels in particular as causing a problem, but it isn't just that. I remember a college book about some language issue or other that used the "fact" that "rider" and "writer" were supposedly pronounced identically to prove some point or other, causing total confusion because none of our tutorial group thought they were even similar.
"Where" and "wear" - no, the "h" is NOT silent.
And so on....

[conuly]

Splits and mergers, basically.

Yeah, it'd be a problem. I'm inclined to go with "easier to read, harder to spell" and include as many options as possible (and certainly there's a good argument for "write + er" to continue to keep a t, even without the pronunciation difficulty!), but the other way is also doable, it just means some people have to get used to the fact that not all sounds are represented.

At the bare minimum, I really would just want to get rid of the most obviously meaningless of the silent letters - all kns except for knish and knuffle, for example, could become plain n without much trouble to anybody. And the silent e in have or love is pointless, really.

[pne.livejournal.com]

Would you split "haff too" (as in "I haff too go too skool") from "hav too" (as in "I hav too books")?

And "yooced too" from "yoozd too"? ("I yooced too go too the beech when I woz yung" vs "I yoozd a scroodriver too pry open the windo")?

Just curious. Because those are spelled the same now, but they're arguable separate lexical items in people's brains, so they might as well be spelled the way they're pronounced.

[conuly]

That's a really good question.

I'd certainly like to. They mean different things. But that sort of thing would be later in the game. Do it too fast and people might revolt. This language is revolting sometimes.

(But that's not at all the orthography I'd go with! Too ugly!)

[pne.livejournal.com]

Have you given thought to a possible orthography?

I was extrapolating from present-day (Great-Vowel-Shifted) values of the various vowel letters and combinations.

What would you suggest?

Or would you devise entirely new symbols - either to complement existing Latin letters (à la Unifon or Pittman ITA), or by replacing the entire thing with a whole new alphabet (à la Shavian)?

[conuly]

Some thought. Not too much serious thought other than "Well, I think I know how many letters I, personally, would like to see" which... you know, I don't know how typical my own speech is!

Unless I want to go the Klingon route of mixing caps and lowercase (in which case I'd mix the letter shapes, not the letter heights) we'd pretty much have to add a few new letters, or else a lot of digraphs and diacritics. I'm not a fan of digraphs unless they're blatantly transparent (so "ow" as in "how" is RIGHT out, we're now two nieces in and both of them think that sound should be written with an a!), and people think diacritical marks are funny looking, so that probably means adding a few letters.

We can get away with reintroducing the thorn for th, and repurposing letters like c, x, and q (because there's no point to them in any logical system, if you ask me), but it's not enough. My thought, though, is to be lazy and just steal letters from other alphabets. Why re-invent the omega?

[pne.livejournal.com]

My thought, though, is to be lazy and just steal letters from other alphabets. Why re-invent the omega?

That would make it easier to adopt on computers, too, since the letters would already be available in Unicode and you wouldn't have to wait for them to be encoded!

And, if you only stole from Cyrillic and Greek, you'd already have great font coverage and the styles would likely match the Latin letters.

[conuly]

Exactly!

Also, they already look familiar, and Greek letters, at least, are already used in our culture, for mathematical symbols and for the names of fraternities and so on, so people have some experience writing them.

Integrating them into our cursive might be tricky, but who writes cursive anymore, really?

[pne.livejournal.com]

I've been thinking about this, and the real issue is with the number of vowels. English simply has really many vowels (especially if you include the diphthongs as single phonemes).

Plus you can't raid Greek and Cyrillic as much as you might like because some of the letters are the same shape; for example, you might like to use eta as an extra vowel, but while the lowercase η is distinct, the capital Η looks confusingly like capital h (H). Similarly, Cyrillic capital У looks different from Latin capital Y, but the lowercase Cyrillic у looks too much like lowercase Latin y. (In Greek, it's the other way around: here, uppercase Greek Υ looks like uppercase Latin Y but lowercase Greek υ looks different from both Latin y and u.)

Also, not using diacritics reduces the number of possibilities; both Latin and Cyrillic tend to add diacritics or hooks and squiggles rather than creating a new letter shape.

However, various African orthographies have been useful, as they took IPA letters and created capital letters for them.

Anyway. Here's what I came up with so far - following the order in http://en.wikipedia.org/wiki/Shavian_alphabet#Letters , and with a sample word where possible:

peep: P p, Pиp.
bib: B b, Bib.
tot: T t, Tot.
dead: D d, Ded.
kick: K k, Kik.
gag: G g, Gag.
fee: F f, Faf (= faff [around])
vow: V v, Vərv (= verve)
thigh: Þ þ, Þoþ (= [Egyptian god] Thoth)
they: Ð ð

so: S s, Sɔs (= sauce)
zoo: Z z
sure: Ʃ ʃ, Ʃuʃ (= shush); alternatively: Ш ш
measure: Ʒ ʒ; alternatively: Ж ж
church: C c, Cərc; alternatively: Ч ч
judge: J j, Juj; alternatively: Џ џ
yea: Y y
woe: W w
hung: Ŋ ŋ
ha-ha: H h

loll: L l, Lol
roar: R r, Rɔr
mime: M m, Maym
nun: N n, Nun
if: I i
eat: IY iy; alternatively: И и
egg: E e
age: EY ey; alternatively: Є є or Э э
ash: A a
ice: AY ay; alternatively: Æ æ

ado: Ə ə
up: U u; alternatively: Ʌ ʌ
on: O o
oak: OW ow; alternatively: Ω ω
wool: Ʊ ʊ
ooze: ƱW ʊw; alternatively: Ȣ ȣ
out: AW aw
oil: OY oy; alternatively: Œ œ
ah: ???
awe: Ɔ ɔ

are: ???
or: ƆR ɔr
air: ER er
err: ЪR ъr
array: ƏR ər
ear: IR ir
(Ian: IƏ iə)
yew: Ю ю

In general, I used the Latin vowel letter for the so-called "short" sound (as in "bat, bet, bit, bot, but") and analysed diphthongs as vowels+y/w and rhotic vowels as vowels+r. (This means you can't distinguish between merry/Mary.)

I couldn't think of any good symbol for the "ah" sound as in "palm". Perhaps the American dictionary style "Ä ä" might be reasonably even if it does have a diacritic on it. Greek lowercase α is distinctive but the uppercase Α is not. And IPA ɑ has a capital Ɑ, but that's really new in Unicode and font support is likely to be bad. (Or look to Armenian and borrow "Ա ա"?)

What do you think? Would you suggest any changes? Would you add any sounds?

For example, if you want to distinguish "which" and "witch", you could add Ɯ ɯ for the former, perhaps. (Unfortunately, IPA ʍ, which would be more accurate, doesn't seem to have a capital equivalent - yet?)

Or for horse/hoarse, I suppose you could go with "or" vs "ɔr" or "ωr".

[conuly]

I definitely wouldn't consider o-as-in-oak and so on to be worthy of digraphs. People speaking English don't think of them that way at all! You can convince them that ay and oy and even I are diphthongs, but not o and ee and so on. (I can't type IPA. I can read it, can't produce it at all!)

As for sh and so on, I'd do it by stealing c for that purpose, and then just caving and making ch a digraph, because it already IS a digraph in English, so it's a little change.

And otherwise, I'd want to avoid repurposing digraphs with new meanings. Repurposing letters works - after all, people can grasp that if you're not going to use c for city and cook, it can be used for something else - but telling everybody "From now on, /ay/ doesn't mean day, it means die"? It makes logical sense, because that's how it's said, but you can't convince people. If it came to that, I'd rather stick with diacritics, it's not quite as tough a sell.

[pne.livejournal.com]

I'm inclined to go with "easier to read, harder to spell"

This makes sense to me.

and include as many options as possible

In moderation, this does, too - and it kind of follows from the other.

But if we distinguish anything distinguished in any native dialect of English anywhere, we probably end up with something similar to what we have now.

Limiting the number of sound changes that people have to perform in their heads to read a text is probably good.

*Where* to draw the boundary may well, of course, prove contentious - and some decisions (e.g. wait/weight) will probably be less controversial than others (e.g. which/witch).

[conuly]

But if we distinguish anything distinguished in any native dialect of English anywhere, we probably end up with something similar to what we have now.

We wouldn't have two anymore.

[pne.livejournal.com]

That's possible.

But you'd probably have to retain historical post-vocalic "gh" just as you would almost certainly retain historical post-vocalic "r" (e.g. spelling "lore" and "law" differently, or "Bart" and "baht").

[conuly]

*pokes*

That's not historic. Those words have rs in them because they have rs in them. It's not my fault half the English speaking world gets it all wrong! :P

[pne.livejournal.com]

Ah - that's because lots of people are rhotic! A majority of English speakers, even.

For some, though, they're merely historic.

The thing with "gammatic" speakers (to coin a word) is that they are so few, most people probably aren't even aware they exist, so for a vast majority, "gh"s are solely historical.

[conuly]

But yeah, for "possible" I meant "reasonable". Pick a percentage of the population, and anything with that percentage or more having that particular sound, include it in the alphabet. (Unless, I guess, it's really very rare even in the groups that have it.)

[pne.livejournal.com]

Yeah.

Preferably with a bit of wiggle room for people to decide but not too much that it ends up completely to personal preference on the part of the reformers which dialects to include.

A bit of present-day significance should probably play a part, but if most of the decision is by a percentage, that would make sense to me.

[pne.livejournal.com]

And the silent e in have or love is pointless, really.

I just had to think of a Finn on the CONLANG mailing list who uses a moderately reformed spelling of English - basically, very cautious changes that still make the text easily comprehensible. Leaving off such useless final e's is one of the things he does. (So he'd write "giv", for example.)