Gree/i/ie/k is tri/ee/ie/cky

Last night I dreamt that someone was explaining to me that, in Early Modern written English, things that look like an /i/ sound are actually pronounced as two syllables. I forget which two, exactly.

I blame the dream entirely on having read this post the previous night.

And, slightly more relevantly, I think phonetic spelling is overrated; regional accents and language change ensure that it can never be spot on and standardized. Far nicer to have etymological clues in the spelling, as in English :).

... You're playing with me, right?

Actually it seems that if you know Ancient Greek a lot of these difficulties turn into memory aids. Alas, I don't.

(I'm against phonetic spelling, sure. But phonemic, that might be something to aim for. And if you're going to put etymological clues in the spelling, at least you should try to make them correct...)

I'm half serious about the hopelessness of phonetic spelling; I see the problem on a continuum between pure logography and an utterly faithful phonetic transcript in some kind of hypothetical über-IPA. There are disadvantages at both ends of this gamut, and you end up with a necessarily unsatisfying compromise somewhere in the middle.

For English I think even a phonemic system would be more hassle than it's worth -- too much semi-rule-bound variation in time and space. How would you deal with, say, the Kiwi vowel shift of the past half-century?

As to false etymologies getting fossilized in words -- well, inevitable I suppose, but in general I'd say that English spelling acts as a brake rather than a spur to that kind of thing...

Rule-bound variation in time is easy: change the system. That's what a language academy is for... (Tongue slightly in cheek, but slightly serious too. Languages change. You have to choose: adapt the spelling to go along, with all the hassle that brings, or let your spelling system drift slowly further and further away from the spoken language, which brings along problems of its own.)

I'll admit that English has a big problem with variation in space, especially once it reaches the phoneme level. Although phonemic variation probably isn't any more extreme (if represented sensibly) than British/American spelling variations, I would have thought, and we seem to manage all right with those.

(Which Kiwi vowel shift do you mean? We can agree that pin/pen/pan all have different vowels, even though ours sound funny, so a phonemic system copes fine with that. Are we erasing distinctions other folks make?)

But really, the cruft that we've accumulated... and as I understand it a lot of English spelling is etymologies put back in (sometimes correctly, often enough not). That might be helpful when everyone learns Latin and French and Greek along with readin' 'ritin' and reckoning (a false etymology is just as good as a real one, if it helps you remember the spelling), but it's not so handy these days.

Re: Kiwi vowels -- yes, I meant the short front ones. OK, I agree if we're writing at the phoneme level we can probably ignore that kind of variation (to me pin and pen often sound identical, but I'm willing to believe that Kiwis can hear the difference even without context). But I think there is enough non-rule-bound variation that a phonemic system would end up almost as complicated as what we've currently got. Say we try to pin down a phoneme A for the sound in cAt -- "If in the UK, pronounce A as /æ/; if in NZ, as /ε/; unless it's the first syllable of "pasta", in which case it's /a:/..."

I think you're right, too, about English etymolography (heh) being to some extent retrofitted -- early modern spelling was of course notoriously variable and I think Johnson (who was far from infallible) was responsible for much of the standardization. But I still think it works far more often than not.

It could well be that I'm just scraping up arguments to support a status quo that I'm familiar with. But from a practical point of view it also seems that spelling reforms are very hard to implement. To my knowledge, Swedish, Norwegian, Czech, and German have all tried spelling reforms in the past century or so. The result in all cases seems to have been conflict, confusion, and failure -- the Swedish academy dictionary is still working on purging the results of an aborted spelling reform from its entries, and Norwegian has ended up with at least five written forms, two of them officially recognized. There are 1.8 billion English speakers as against 4.7 million for Norwegian, so I suspect there's little hope. Of course, if you're just saying that phonemic spelling reform would be a good thing rather than that it would ever stand a chance of working, this doesn't really bear on your argument.

About pen/pin: that's exactly why I stress phonemic as opposed to phonetic. We know that these are "different vowels" even if we in fact pronounce 'em exactly the same. If we can't hear the difference without context that doesn't matter: "phoneme" is a conceptual/cognitive label not an auditory one. (I don't mean to lecture and I'm sure you know this, but it's essential for my point.)

And as for the kind of variation you're worried about ("how should we pronounce the A in cat?"), why is that something for the spelling system to deal with?

There are two principles you might want to stick to, if you're trying for phonemic spelling: (1) a single phoneme is always spelled the same, (2) a single glyph (perhaps glyph combination) always represents the same phoneme. You're not worried about either of these here, but about how that phoneme gets realised phonetically. (I think the 'a' of "cat" might be a different phoneme to the first 'a' of "pasta" -- say "cart" without post-vocalic R and you've got damn near a minimal pair, apart from some length issues. But if that's the case it should be the case in most every English. More problematic is the suggestion, if I'm reading it right, that distinct phonemes are merging in Kiwi. That causes problems for this sort of approach, and it does happen, just not nearly as often.)

The most important of those rules seems to be the second: that would let second language learners learn spoken vocabulary from writing. (That's assuming the phonetic realisation of phonemes is rule-based enough to get to grips with, which it sort of must be by definition.)

I'm talking mainly here about second language learners, by the way, because we lucky native speakers seem to manage fine as it is. Although apparently the Chinese character system is holding back national literacy levels something wicked, but we're not as bad as all that.

Finally about actually making the changes... not very likely. Although besides your examples I'd like to hold up (a) the Netherlands (constant low-grade spelling reform, basically successful), and (b) the US of A (as I understand it Webster was responsible for lots of stuff they do wrong, er, differently). There are several practical problems: making it stick across the billions in different jurisdictions is one, deciding which reforms to pursue when is another (you don't overturn everything in one go, of course, you make small and gradual changes... unless you spend all your time arguing about which ones to start with).

Another problem somewhat unique to our point in history is the upheaval a reform would involve. We've got heaps of words lying around on paper, that would have to be rewritten. A few hundred years ago there weren't that many of the damn things, and in a few hundred years they'll be stored electronically and we will just use find-replace. Right now though... tricky.

Sorry, my irrelevant pen/pin aside probably confused matters -- yes, I know what you mean by phonemic vs. phonetic, and I can certainly see the advantages. But re: "is this something for the spelling system to deal with?" -- I thought that the main motivation for phonemic spelling was a simpler mapping between spelling and pronunciation. The ideal, as I understand it, is that the phonemes are fixed and each dialect corresponds to a well-defined bijection between phonemes and sounds. But cat/pasta doesn't fit there: if Kiwis are pronouncing cat as "ket", they should be pronouncing the first syllable of "pasta" as "pest". So we have to say they're different phonemes, and spell the word differently depending on where we are -- or, more likely, accept that no scheme will be perfect. But perhaps such anomalies are rare enough that the system would still be worthwhile. As you say, phoneme mergers are a worry too, but probably rare.

Tangential aside: from the Wikipedia talk page on the Droste Effect, on the pronunciation of "Droste": "... it's roughly what your car says when you try to start it and the battery is dead".

Cat/pasta: I'm confused. Are you saying that in your head they are the same phoneme? Then, indeed, different dialects have different phonemic representations for the same word (which isn't surprising really, but which we hope will be fairly rare).

But really, forget cat/pasta and look at through/trough/enough/thorough... I think we can agree there are bits of English spelling that are more complicated than they ought (ought!) to be.

(Another aside: I think dialects will vary much more on the vowel system, certainly phonetically and perhaps also phonemically. So why not start by trying to regularise the consonants? And maybe some of those are problematic too, but start with the simple and obvious ones... Only please, o universal English regulatory body, start somewhere!)

And re Droste: fantastic! And people say Wikipedia isn't a replacement for a good encyclopedia.

Cat/pasta: that's exactly what I'm saying. It was probably a mistake to start this kind of discussion without some kind of audio channel to clarify matters.

"through" reminds me that I haven't yet dragged out the argument that spelling reform will increase the number of homographs... it's probably not very compelling anyway -- we have tons of homographs already and context disambiguates nicely for the most part.

My staunch support of the status quo is weakened whenever I dip into the OED for something and am reminded that most English words have been spelled in every conceivable fashion over time. Often in several inconceivable fashions too. Of the 70 spellings listed for "through", I think my favourite is "thrwch" -- did I not know better, I'd incline to pronounce it like the noise your car makes when you hit reverse instead of second.

Cat/pasta: you're right, this might be futile. I have a glimmer of a sense of what might be going on, though: is "your" pasta what you'd think of as the Italian pronunciation? Because I think I'd argue that there are two distinct ways of pronouncing the word, the "Italian" and the "American", say (to be offensively stereotypical), which ... use different phonemes ^_^ But yeah, without audio it's hard to tell if we're really talking about the same thing.

I'll continue to maintain, though (now that it's occurred to me) that focussing on vowels is like complaining about the impossibility of herding those cats out of the room... you're ignoring the elephant.

Homographs... A cheap shot: we manage fine with homophones, and those would be exactly the homographs if the reform is done properly. It's cheap, of course, because some of the audio context we use isn't present when written (punctuation is a poor substitute for intonation). I do think though (gut feeling) that these wouldn't pose much of a problem to new learners of the new system. They'd be hell, I'll admit, to people having to transition: "Is that new 'thru' the 'through' I knew or the 'threw' I knew?"

"Thrwch" is marvellous, but actually not much more bizarre than "through" -- the "w" is puzzling, but the "ch" is I guess the same sort of remnant as "gh". And even the "w" is only odd in a very phonemic sense: we don't use it that way but we could have chosen to, systematically. If I try to add the Dutch "g" to "through" I can appreciate the claim that the vowel is the same as the start of "wood" or "will". (Welsh maybe does systematically?)

Btw, "dough" ought to go in that list as well.

Also btw, the Wikipedia page for Welsh cites an essay by Tolkien which is full of gems, including this particularly nice reason to impose spelling reform: "C for K, because the printers have not so many as the Welsh requireth."

[...] having a debate with my southern hemisphere evil twin at the moment, about the merits of spelling reform. No conclusions to be expected, of course, but it’s leading to all sorts of delightful [...]

Final, probably futile, attempt to clarify pasta: I distinguish (perhaps incorrectly) two English pronunciations: UK /'pæstə/, Kiwi /pɑːstə/ -- I think US may be different again but I don't know what it is. The Italian pronunciation as near as I can get it seems to be /'pasta/. I think that `-ast-' in English is almost always /æ/ (spastic, raster) or /ɑː/ (master, bastard) and never /a/, so the English variants have just jumped different ways in their approximations. But then I'm a very amateur phonologist.

And yeah, forget homographs, I only mentioned them to mention that I wasn't going to mention them.

Cat/pasta: finally, oh finally, I'm on the same page. Because now I understand why this is a problem: if you're going to indicate the difference then Kiwis will spell "pasta" differently to Brits. (That's if we can find a minimal pair for "-ast-", which surely must be possible.)

To which I reply, yes. But maybe we'll have to relax the rules for vowels anyway.

Once we're making the rules, I mean. When we're in charge.