Fixing English writing
English writing is hard to learn because the letters often do not match the sounds. One starts with a w sound, but we do not write w. Colonel sounds like kernel. The letters ough are not one sound: they change from word to word, as in though, tough, cough, plough, and through.
This is not a small problem for children. Italian writing is close to one letter, one sound: about 33 letter groups for about 25 sounds [1]. English is full of exceptions: one count is about 1120 letter groups for about 40 sounds [2]. In a study of 13 European writing systems, children in regular systems such as Italian could read simple familiar words well before the end of the first school year (more than 95% correct). English children were at 34% in year 1. They needed more than twice as long to reach the same level [3]. Another study found Italian children already reading about 94% of real words at the end of year 1 [4]. Spelling stays hard for longer too: Italian children were accurate after 2 years of school, while English children were still not accurate after 5 years [5]. The extra time is not because English children are slower. It is because English letters do not keep one sound [6].
Start from the sounds
A better way is to start from the sounds, and write those. This is called phonetic transcription.
The International Phonetic Alphabet already does this, but it uses many new symbols. People want the familiar letters they already use, without having to change the keyboard of their phones and laptops, and without having to learn a new alphabet.
We can keep those letters, and add small marks on the vowels.
The DoAc system
Hungarian already does this. It has five vowel letters, and a few kinds of marks:
- no mark: a e i o u
- one dot: ȧ ė ị ȯ ụ
- two dots: ä ë ï ö ü
- one accent: á é í ó ú
- two accents: a̋ e̋ i̋ ő ű
We call this the DoAc system, short for dots and accents. Five letters times five kinds of marks is 25 vowel signs. That is enough.
How many vowel sounds
But how many sounds does English have? Well, it gets tricky. On a standard account of English sounds [7] there are 20–25 vowel sounds in British Received Pronunciation, 14–16 in General American, and 19–21 in Australian English. Some of those extra sounds are two vowels said one after the other, like the sound in night. Some are a vowel plus r, like the sound in here.
If we want one sign for each simple vowel, and we want to cover British English too, we need the full set of marks. For simplicity, we think that when you hear two vowels, you should write two letters. Night is nait, not one special letter. The 25 signs are for the simple vowels. The two-vowel sounds are just two of those signs in a row.
Linguists already have names for these vowel families. Each family is called by one example word in capital letters [8]. STRUT is the family of but and blood. PALM is the family of father and yacht. TRAP is the family of cat. Those names are well known, so we keep them here.
A mapping
Here is one way to use DoAc. Empty signs can be used later, for example for the two different a sounds that some Australian speakers have in bad and lad.
| linguistic group | word | our vowel | transcription |
|---|---|---|---|
| STRUT | but | a | bat |
| PALM | yacht | á | iát |
| TRAP | cat | ä | kät |
| PRICE | night | ai | nait |
| MOUTH | now | au | nau |
| DRESS | bed | e | bed |
| FACE | day | ei | dei |
| COMMA | the | ė | thė |
| NURSE | colonel | ë | kërnel |
| SQUARE | hair | e̋ | he̋r |
| KIT | bit | i | bit |
| FLEECE | see | í | sí |
| NEAR | here | ï | hïr |
| LOT | cot | o | kot |
| GOAT | go | ou | gou |
| THOUGHT | thought | ö | thöt |
| CHOICE | boy | oi | boi |
| FOOT | book | u | buk |
| GOOSE | too | ú | tú |
| CURE | sure | ü | shür |
If you hear an r, write an r. Do not hide it inside the vowel. Sure is shür, not shü. Here is hïr. Hair is he̋r. Car is kár. Some British speakers do not say that r. They can leave it out. That is their accent. It is also one reason American English is listed with fewer vowel sounds: it still writes and says the r.
Consonants are easier. Do not write silent letters. If the sound is k, write k. We can keep th, sh, ch, and ng. The j in jeep stays j. The glide at the start of yes is written i, not y: yes is ies, yacht is iát, you is iú. After a consonant the same i is used, so queue is kiú and beautiful is biútifėl. Then y is spare.
This map keeps LOT, THOUGHT, and PALM apart (kot, thöt, iát). Many speakers merge some of those. A writer who merges them can use fewer marks.
Should one accent be the standard? Canadian English could be seen as a strong candidate. It is rhotic, so every hearable r is written, which matches the rule above. It sits close to General American, already the main model in film, pop music, and many classrooms. It is relatively uniform from coast to coast. However, two familiar Canadian facts would hide splits that other large accents still make. Many speakers merge LOT and THOUGHT, so cot and caught would look the same, and PALM often joins them. Raising of PRICE before voiceless consonants would also split night and ride, so they would no longer share one spelling. Learners and dictionaries already name General American or Received Pronunciation. The shared standard here is the Wells set, not a national accent. The letters on the page then record one rhotic pronunciation, close to General American, while still leaving cells for British distinctions.
Examples
Then hard words become easy to read.
Though the tough cough and hiccough plough him through
becomes
thou thė taf köf ėnd hikap plau him thrú
Those words look alike in normal English. Here they have different vowels, because they have different sounds.
The rain in Spain stays mainly in the plain
becomes
thė rein in spein steiz meinli in thė plein
You can see the same two letters, ei, in rain, Spain, stays, mainly, and plain.
More examples:
colonel → kërnel
queue → kiú
women / woman → wimin / wumėn
one, two, eight, four → wan, tú, eit, för
knight → nait (same as night)
choir → kwaiėr
yacht → iát
Leicester → lestër
beautiful → biútifėl
knowledge → nolij
yes → ies
I owe you → ai ou iú
here / hair / sure → hïr / he̋r / shür
cot / caught / cart → kot / köt / kárt
A connected sentence uses the same map:
I told the colonel the choir would wait in the yacht.
becomes
ai tould thė kërnel thė kwaiėr wud weit in thė iát
There is an old joke that fish could be written ghoti. The joke takes the gh sound from enough (that gh is an f sound), the o sound from women (that o is an i sound), and the ti sound from nation (that ti is an sh sound). Put f + i + sh together, and you get fish. The joke works only because the same letters change sound from word to word. If each letter had one sound, ghoti could not mean fish. In DoAc, fish is written fish.
This would take time to type. The marks are not a new keyboard: a long press on a phone already offers á and ä. Writing with no dots or accents can be treated as fine for speed: thot for thöt, the for thė. A reader who already knows the map can recover the vowels. That page is still formally incomplete. The marks are what keep the vowel families apart. A fast note can drop them. A primer or a public text should keep them.
A first test would take people who are learning to read English for the first time and split them at random into two groups. One group is taught with DoAc. The other group is taught with ordinary English spelling. After one day of lessons, both groups read the same list of familiar words and simple invented words aloud. An evaluator who cannot see the page, and does not know which group the reader is in, scores only from the audio: which group reads more words correctly. Until that trial exists, this is a claim about signs, not a claim about classrooms.
Different accents would not pick the same marks for every word. People who like old spellings would not like it. Children already spend years on the old spellings. We have 25 signs, and British English has 20 to 25 vowel sounds. We do not lack signs. We lack a writing system that starts from the sounds.
Earlier attempts
This problem is old, and there have been many attempts to fix English writing. They fall into a few groups, and each group falls short in a different way.
Some people changed only a few words. Noah Webster [9] gave American English color and center. That is easier to print, but the vowels are still a mess. One, colonel, and ough did not get fixed.
Some people invented a new alphabet. Benjamin Franklin did this in 1768 [10]. The Deseret alphabet did it in the 19th century [11]. George Bernard Shaw left money for a new alphabet with at least 40 signs, not based on Latin letters. The result was the Shavian alphabet [12]. One book was printed in it, Androcles and the Lion, in 1962. Almost nobody used it. A new alphabet can match the sounds. People still have to leave the letters they already know.
Some people kept Latin letters, but added extra shapes for school, then asked children to switch back. The Initial Teaching Alphabet [13] was used in British and American schools in the 1960s. Children often learned to read faster at first. Then they had to learn ordinary English spelling as a second system. The extra letters were also hard to type. The method fell out of use.
Some people kept the 26 letters and respelt words, like thru and enuf. The Simplified Spelling Society worked on this for many years, for example in New Spelling [14]. Cut Spelling [15] only drops spare letters, so accommodation becomes acomodation. These look more like English, but 26 letters are not enough for 20 to 25 vowel sounds. The extra sounds then need piles of two-letter and three-letter groups, so the exceptions come back in a new form.
So the old attempts miss at least one of these: they leave the Latin alphabet, they do not have enough vowel signs, they change too little, or they make children learn two systems. DoAc keeps the familiar letters, adds marks so there are enough vowel signs, and is one system, not two.
| kind of reform | Latin letters | enough vowel signs | one system for life |
|---|---|---|---|
| Webster-style tweaks | yes | no | yes |
| new alphabet (Shavian) | no | yes | yes, if adopted |
| Initial Teaching Alphabet | partly | yes | no |
| 26-letter respelling | yes | no | yes |
| DoAc | yes | yes | yes |
What do you think? Would you read English in DoAc? Some signs are still free, like ị ȯ ụ i̋ ű. That gap is for the difference between 20 and 25 vowel sounds. Please comment below.
References
- A. L. Lepschy and G. Lepschy, The Italian Language Today (Hutchinson, 1981).
- J. Nyikos, “A linguistic perspective of functional illiteracy,” in The Fourteenth LACUS Forum 1987 (Linguistic Association of Canada and the United States, 1988).
- P. H. K. Seymour, M. Aro, and J. M. Erskine, “Foundation literacy acquisition in European orthographies,” British Journal of Psychology 94, no. 2 (2003): 143–174. doi:10.1348/000712603321661859
- G. Cossu, M. Gugliotta, and J. C. Marshall, “Acquisition of reading and written spelling in a transparent orthography: Two non parallel processes?” Reading and Writing 7 (1995): 9–22. doi:10.1007/BF01026945
- C. V. Marinelli, C. Romani, C. Burani, and P. Zoccolotti, “Spelling acquisition in English and Italian: A cross-linguistic study,” Frontiers in Psychology 6 (2015): 1843. doi:10.3389/fpsyg.2015.01843
- J. C. Ziegler and U. Goswami, “Reading acquisition, developmental dyslexia, and skilled reading across languages: A psycholinguistic grain size theory,” Psychological Bulletin 131, no. 1 (2005): 3–29. doi:10.1037/0033-2909.131.1.3
- “English phonology,” Wikipedia.
- J. C. Wells, Accents of English (Cambridge University Press, 1982).
- “Noah Webster,” Wikipedia.
- “Benjamin Franklin’s phonetic alphabet,” Wikipedia.
- “Deseret alphabet,” Wikipedia.
- “Shavian alphabet,” Wikipedia.
- “Initial Teaching Alphabet,” Wikipedia.
- “English-language spelling reform,” Wikipedia.
- “Cut Spelling,” Wikipedia.