Friday, September 24, 2010

Thoughts on English Spelling Reform

There are half as many dyslexic Italians as there are English ones (source). Why? Because, while dyslexia is a symptom of differences in the brain, the way it manifests depends on how complicated the spelling system is, and Italian spelling is much simpler than English.

American schools spend 12 years teaching spelling, only to have half of American interweb writers misspell separate and definite. Italian schools spend two years on spelling. Imagine what kids could be taught if spelling only took two years! Maybe they'd have time for some statistics.

Cartoon showing student correcting teacher's spelling on the chalkboard
Because of these facts, I'm trying to put aside my crotchety resistance to English spelling reform. Like any endeavor, it's a complex undertaking with many approaches and advocates. My take on it, though, is that a new, more regularized spelling system would have to meet these requirements:

  • The new words would have to maintain as much similarity as possible to current spellings to aid the transition from current spellings to reformed ones.
  • No diacritical marks (accents, umlauts, etc.) and no introduced characters should be used.
  • The new spellings should be based as much as possible on the meanings of the words rather than just the sounds, although sounds will clearly be taken into account in many cases. For example, most plurals end in s. The fact that plural s is sometimes pronounced ess (as in cats) and sometimes z (as in dogs) doesn't mean the plurals should vary their endings to s or z, depending on pronunciation. Consistently signaling pluralness is more important than making the letters represent the sounds. (Whether we should get rid of the variant plurals such as oxen is another question.)
Two proposed systems that look interesting are called SoundSpel and Cut Spelling. In addition to regularization, they also shorten texts (4 percent for SoundSpel, 8 to 15 percent for Cut Spelling), which is an important side benefit in terms of paper-usage.

When transforming English spelling, the biggest challenge is the vowels. We only have 5 (or 6), used to represent over 20 sounds. Of course, there are pronunciation variations across English-speaking countries and regions, mostly in the vowel sounds. But one standard has to be chosen.

SoundSpel has been in development for about 50 years and is endorsed by the American Literacy Council. It uses all of the existing letters, and deals with the vowel shortage by creating two-vowel combinations. The Wikipedia entry lists all of the vowel assignments, including the basic five as a = sat, e = set, i = did, o = dot, and u = cut. Long vowels are made by adding an e to to the basic vowel: ae = sundae, ee = see, ie = die, oe = toe, ue = cue. Double vowels are used for some of the others sounds, such as aa = alm, oo = moon, uu = book.

All of the consonants are kept as is, including the wh in why, with the addition of zh to create the z sounding azure. C and K are both kept (not the simplest choice, but obviously better for the transition), and there is still no differentiation of the th sound in the vs. thing. Qu is simplified to Q.

Some exceptions are made, again for aiding the transition, I assume. No change is made to the short, common words was, as, of, the, he, she, me, we, be, do, to, and off. The letter s is used for plurals, whether it is pronounced ess or z (yay!).

Here's a sample:
It was on the ferst dae of the nue yeer the anounsment was maed, allmoest siemultaeniusly frum three obzervatorys, that the moeshun of the planet Neptune, the outermoest of all planets that wheel about the Sun, had becum verry erratic. A retardaeshun in its velosity had bin suspected in Desember. Then a faent, remoet spek of liet was discuverd in the reejon of the perterbd planet. At ferst this did not cauz eny verry graet exsietment. Sieentific peepl, however, found the intelijens remarkabl enuf, eeven befor it becaem noen that the nue body was rapidly groeing larjer and brieter, and that its moeshun was qiet different frum the orderly progres of the planets. – H.G. Wells (from the Wikipedia page)
Cut Spell, on the other hand, removes letters more than it reassigns them. Unpronounced letters are generally zapped, although the silent helper e at the end of words is kept and made more consistent (so peace becomes pece). Unstressed vowels (the source of all those misspelled examples of separate and definite) are also removed, so symbol becomes symbl, waited becomes waitd, and most disconcertingly to me, the becomes th. Most double consonants are also dropped, ph becomes f, g as in judge becomes j, and igh (as in high) becomes y.

Here's a sample:
Wen readrs first se Cut Spelng, as in this sentnce, they ofn hesitate slytly, but then quikly becom acustmd to th shortnd words and soon find text in Cut Spelng as esy to read as Traditionl Orthografy, but it is th riter ho really apreciates th advantajs of Cut Spelng, as many of th most trublsm uncertntis hav been elimnated. (from the Wikipedia page)
Of the two reform systems, I think I prefer SoundSpel because it appears to be more consistent in its changes, and less reliant on knowing traditional spellings. Some oddities from the Cut Spelling sample:
  • who becomes ho. How would anyone know how to pronounce that correctly?
  • Readrs isn't changed to Redrs? What good is that extra a?
  • The ys in slytly are pronounced differently. How would a new reader have any way of knowing that?
Despite this, reading the passages of Cut Spelling and SoundSpel is not horrendously hard (although it's also not comfortable). But I think I could get used to reading this way within six months or a year, maybe less. It would take me much longer (or maybe never) to write them.

But automatic translation such as this is what we have computers for, right?

No comments: