Web hosting Custom Email SiteBuilder

The Nature of and Alternatives to Standard English Orthography

Abstract:

Most people are not fully aware of all the problems with English spelling. They confuse the language with its traditional orthography and have a misplaced reverence for the antiquated spellings preserved in the dictionary.

This article reveals some of the absurdities of the spelling system used in English speaking countries since 1800, explains the difference between a language, a script, and a spelling system, and chronicles some of the attempts to "break the spell"* advanced by advocates of simplified and regularized spelling. 

In the abstract the solution to irregular spelling is simple. Just make it regular: Select a system of symbols that correspond on a one to one basis with the significant sounds or phonemes of the spoken language and offer it to the younger generations. Spelling simplification reforms have been successful in other countries which is why their scripts are so much more phonemic (i.e., consistent with the alphabetic principle)ⁿ.

In the past, while there was widespread agreement on the need to regularize spelling and simplify English orthography, there was little agreement on how to achieve this. Today, there is some convergence of opinion on how to bring the spelling system used by English speakers more in line with the alphabetic principle. This article explores some of the proposals that would remove the major roadblock to reading and to advancing English as an international language.

The Problem with TO (Traditional English Orthography)

English spelling is notoriously hard to master. While people have forgotten the problems they had learning to read, they continue to be aware of the problems they have trying to spell. What most people don't realize is that their problems are man-made. The horrendous orthography is a historical accident. Around 1700, the spelling of English became standardized before was regularized. Other countries have recovered from such accidents by implementing spelling reforms.

Today, spelling problems are almost unique to English. Its traditional orthography is more of a disguise than a guide to pronunciation. ¹ Compared to the spelling systems used in other countries, many of which use nearly the same Latin script, the code for TO (Traditional English Orthography) is unnecessarily complex, inconsistent, and irregular. According to Laubach,

"English is the world's worst spelled language." ²

 

Some of the absurdities of TO are captured in the following limericks³:

To paraphrase Paul Hanna, the basic problem with TO is an overabundance of graphemic options. The number of graphemically correct spellings depends on the length of the word. A 4 letter word can typically be spelled over 10,000 different ways, only one of which is orthographically (or rather lexically) correct. A two letter word, such as *in can be spelled ( 33x14 ) or 462 different ways. The /short i / sound can be spelled 33 different ways and the /n / can be spelled 14 different ways. In languages with simplified and regularized spelling systems, most four letter word can be spelled only one way. In countries with regularized writing systems, there is no such thing as a spelling bee. If the orthography is phonemic, any word that is correctly pronounced can be easily spelled.⁴

According to Ellis (Dewey, 1971 ), the eight letter word "scissors" can be spelled phonetically only one way [sizerz]. In English, it can be spelled 596,580 different ways. The initial /s/ sound can be spelled 6 different ways. The /i / can be spelled 16 different ways. The table in Fig. 5 lists the number of different ways each sound can be spelled in English. These are multiplied together to arrive at the over half million different graphemically correct options. Only one of the options is lexically correct.

English and Spanish use the same Roman alphabet. Spanish is consistent and highly phonemic. English isn't. How can this be if the Roman letters share more or less the same sounds in both languages?

A writing system consists of both the set of characters (the visible script) and the rules for combining them to represent word sounds. A writing system includes both the graphemes and the orthography or spelling system. Any language can be rewritten or transcribed into any script. For instance, English words can be rewritten in hieroglyphics. At uiuc.edu you can type in your name (or any other short word) and have the computer convert it to hieroglyphic signs (uniliteral phonograms). On the chart below, this would be a case of L1-C2-O1. The only thing being changed is the character set. English can also be rewritten using Portuguese orthography LI-C1-O2.

 

 

If any language can be rewritten using a different set of grapheme-phoneme correspondence rules, why can't the problem with English spelling be solved by simply rewriting it using the Spanish, Portuguese, or Finish spelling system? It can. However, in solving one problem this kind of solution creates another. In the following example, an English phrase was transcribed into the Spanish / Portuguese writing system. It solved the regularity problem but it changed the familiar word patterns of TO.

In the Spanish and Portuguese orthography, [e] = the /short e/ and that's about it. One letter, one sound. The /long e/ sound is represented by the letter [ I ]. In English, the /short e/ and the /long e/ are represented by the same grapheme. English adds to the confusion by letting the letter [e] represent just about any vowel. As shown in Fig. 5, the letter [e] can be pronounced 7 different ways.

In Spanish, the /long a/ sound is spelled [ei] as in *vein. In English, the /long a/ is spelled 36 different ways. It can be spelled [ei] as in Spanish and Portuguese, but it can also be spelled [a, ae, a..e, ai..e, aigh, au, aye, ah, au, a..u, au..e, a..ue, ea, . . .]. In Portuguese, the /sh/ sound is spelled [ch] as in *machine. In English, [ch] is one of 14 different ways to spell /sh/. To represent /ch/, Portuguese adds a T to their /sch/ sign yielding [tch]. Thus,

If regularity were the only problem, then it could be solved by using a highly phonemic script such as Spanish or Portuguese. However, converting to Portuguese orthography achieves regularity and the expense of familiarity. What many spelling reformers want is an orthography that will be both consistent and retain familiar word patterns. The problem, illustrated in Figure 3, is that English is so irregular that 30% to 60% of the words look strange no matter what phonemic writing system is used. Follick spelling, shown above, seems particularly wierd to TO readers.

As early as 1793, people such as William Thornton were writing,
 

English aut too kontaen a singul distinct mark oer kaeractur, az dhu representativ ov eech simpl sound hwich iz posibl for dhu human vois and breth too utter. Noe mark shuud represent too oer three distinct soundz, noer shuud eni simpl sound bee represented bie too or three different kaeracturz.

Inglish ot tu ka'ntein a' singl di.stinct mark or kera'ktr, az th' repra'senta'tiv ov ich simpl saund `wi.ch i.z posibl for th' humn vois a'n breth tu, u'tr. No' mark shu.d repra'sent tu, or thri di.stinct saunds, nor shu.d eni si.mpl saund bi repr'sent'd bi tu or thri di.f'ra'nt kera'ktrz.

The first paragraph above, written in WES (World English Spelling), is an example of a highly phonemic script which preserves many of the spelling patterns of TO. "A chair" would be written "Ae chaer" in WES. WES does not disrupt old sight reading habits as much as the Follick spelling and CCS. Anglic, a variation of WES, comes even closer to TO by accepting 40 common irregularly spelled words such as *of and *the.

In Chekt Spl. apostrophe and [a'] is a sound-sign representing the unstressed /uh/ .

In the next chart, the different English spellings of one phoneme /oo/ are enumerated. Nine different spellings of the sound common to rule, tool, and flue were listed by Soffietti . His list, shown in column one is not exhaustive. Another nine spellings were added in column two. According to Dewey (1971) an exhaustive list from the American College Dictionary would contain 29 different spellings. As shown in Figure 3., a phonemic script such as WES (World English Spelling) would provide only one spelling option: /oo/ is always spelled [oo]. This means that *wooed would be spelled [wood] and *wood would be spelled [wuud]. The sound common to *wood, *could, *book and *foot would be spelled [uu].
 

In traditional English orthography (TO) the same sound is spelled an average of 14 different ways

	TO Graph	TO TraditionaL Orthography	WES World Eng Spl OGD	Cut Speling	Chekt Speling Follick		TO Graph	TO Traditional Orthography	WES World Eng Spl	Cut Speling	Chekt Spelling Follick
1.	u..e	rule	rool	rul	rul	10.	u	ruby rooby	roobee	ruby	rubi
2.	ue	flue	floo	flu	flu	11.	ou..e	troupe	troop	trup	trup
3.	ui	fruit	froot	frut	frut	12.	ough	through	throo	thru	thru
4.	eu	maneuver	manoovr	manuvr	manuvr	13.	ooe	wooed	wood	wooed	wud
5.	ou	group	groop	grup	grup	14.	oo..e	ooze	ooz	ooz	uz
6.	ew	grew	groo	grew	gru	15.	heu	rheumatism	roomatism	rumatism	rumatism
7.	o..e	move	moov	move	muv	16.	ui..e	bruise	brooz	bruse	bruz
8.	oe	canoe	cunoo	cano	canu	17.	wo	two	too	to* two	tu
9.	oo	moon	moon	moon	mun	18.	o	do	doo	do	du

 Figure 3. Alternative Spellings of the /u:/ or /oo/ sound.  latest chart

The table shows 18 of the 29 different ways that the /u:/ sound is spelled in TO. Cut Spelling reduces the number from 18+ to 4. Phonemic systems such as WES, CS, and NF reduce the number of different ways to spell a sound to 1. See Dewey, 1971, for a complete list words illustration the different possible TO spellings for every sound.

By removing the letters which are irrelevant to pronunciation, Cut Spelling (SC), shown in column 5., reduces the number of possible spellings from 18 to 8 (u, oo, o, ew, ...). CS has two advantages: (1) It does not disrupt existing visual habits. Literate readers are not slowed down by the elimination of redundancies.⁵ (2) It conserves space. WES may use 2% fewer letters than TO; Cut Spelling uses up to 20% fewer letters (18.6% above). New Follick uses about 12% fewer letters.

How irregular is English?
English can't be spelt. - G.B. Shaw

English was not regularized before it was standardized. As a consequence, over 25% of the words have irregular spellings and 10% don't fit any logical pattern whatsoever (Hanna, 1971). The situation is actually a little worse than the 25% statistic suggests but not quite as bad as the 92% figure that some have used.

The extent of irregularity one finds in English depends on ones unit of analysis. One could argue that since we have 500 symbols (including digraphs) to represent 40 or so sounds, English is 100%-(500-40)/500% or 8% phonemic. This would suggest that English is 92% irregular.

If one examines the phonemes (or single sounds) used by Lincoln in his Gettysburg Address, 75% are phonemic or regular. However, if one examines the 265 word address at the word level, only 106 (or 40%) are phonemic. What this means is that if the Gettysburg Address were transcribed to a more phonemic script, up to 60% of the words would look strange, e.g.,

 

Forskoer and seven yeerz ugo our faathers braut fortth on this kontinunt u noo naeshun... - Cf. Dewey, p. 16 -  Full Address words highlighted in red transcribe correctly Only 3 out of 15 or 20% are regular in this passage the rate goes up to 40% for the full 265 word address.

60% of TO spellings are not alphabetical Phonemes Syllables Words 75% 50% 40% Fig. 4.  Three Answers to the question "How Phonemic is English?" the unit of analysis can change the result

 In OGD-positional spelling the segment would transcribe with more correctly spelled words:
 Forskor and seven yeers ago our faathers braut fortth on thiss continent a nu naytion...
 OGD is a systematic but not a phonemic (1 sound-1 spelling) orthographic system.

Hanna's study of 17,000 English words showed that about 50% could be spelled correctly on a phonological basis (p. 122) and that 84% were spelled according to a regular pattern. For example, The /long e/ in *eel is spelled 20 different ways. However this study indicates that it is spelled [e] or [e...e] (72% of the time), [ee] (10%) and [ea] (10%). These four regular patterns account for 92% of the long-e spellings found among the 17,000 most frequently used English words. If regularity is defined in terms of absolute consistency or one and only one spelling pattern, then only 40% of the words are regular. If, on the other hand, it is defined in terms of having no more than four different spelling patterns, then about 84% of the words are regular.

How English came to be written the way it is
 

The heart of the trouble is with our foolish alphabet,  it doesn't know how to spell and can't be taught. - Mark Twain, 1899

Mark Twain took some literary license in his chautauqua speech. As should be clear by now, it is not the alphabet (or the character set) but rather the orthography or spelling system that is at fault. The English spelling system does not adhere to an alphabetic principle. In this sense, Twain can argue that "it doesn't know how to spell." A transcript of the full speech appears on the Web.

To represent the 40 or so sounds (or phonemes) of English requires 40 to 50 sound signs (or graphemes): at least 18 more than what is available in the 26 character Roman alphabet which has 3 redundant characters. (See Fig. 3)

The Latin alphabet was adopted by the Celtic speaking population during the Roman occupation (ca. 400 ad). When the Romans left, the Anglos and the Saxons, Germanic tribes from Denmark arrived. Eventually they to adopted the Roman alphabet which replaced Futhark. Initially, the Latin alphabet was augmented with some of the Futhark characters notably the ash, thorn, and eth. Germanic tongues have 40 some phonemes compared to 20 some for Latin. There was nearly twice as many sounds as there were letters in the Latin alphabet. From 1000 AD to 1600 AD, English and Latin speaking scribes came up with a wide variety of individualistic solutions to cope with this graphemic deficiency. The most common solution was to augment or expand the Roman alphabet with digraphs, letter combinations used to represent the missing phonemes.

The problem was that they created too many digraphs and letter combinations. Instead of one digraph for each unrepresented sound, they created a surplus of 5 to 20. Even the sounds that were represented, such as "s", were spelled six different ways. (s, ps, sw, c, sc, sch). What began as a way to cope with the problem of graphemic deficiency was the genisis of today's problem, that of graphemic overabundance or diversity.

In the 14th Century, Chaucer lamented the "grete dyversite in English and in writyng of our tongue." Over the centuries, one type of diversity has been effectively abolished but the diversity of ways to spell a particular sound has remained. We can no longer use a [y] for the /long I/ in "diversity" but we can still use this letter to represent the /long I/ in other words such as "fly".

The property of a written sign to have more than one value is called polyvalence (Coulmas, 1996, p. 413). The typical goal of a spelling reform is to reduce the degree or level of polyvalence. Prior to 1760 there were reforms in the way that English was written but most of them increased polyvalence.

While writing tends to remain the same, pronunciation changes over time. The 15th century was a period of rapid change referred to as "the great English vowel shift." This was followed by a rise in the affordability and popularity of the printed word. In the 16th century, the demand for reading material was largely met by foreign (e.g., Dutch) printers. The printers not only dropped 4 of the old English characters, such as the thorn (þ) , but also failed to keep pace with the sound changes that were affecting the language.

With the publication of the King James Bible in 1611, a standard or authority for some uniformity and consistency in spelling was established. After the publication of Johnson's Dictionary in 1755, only one of the 50 or more ways that a short word could be spelled was deemed correct. Unfortunately, this standardization came about before regularization. Strict adherence to the alphabetic principle would not have permitted one to spell fish "G-H-O-T-I" or retained the confusion of most lost ...

In the abstract, the solution to irregular spelling is simple. Just make it regular: Assign one and only one sound sign to each of the 40 or so significant sounds or phonemes of the spoken English and allow the younger generations to use it in place of TO.

While such spelling reforms have been successful in other countries and while Ben Franklin, Noah Webster, and others believed the benefits would clearly outweigh the temporary inconvenience, support for a major reform has never enjoyed widespread political support. It is as if tampering with the spelling system was tantamount to tampering with the Constitution.

People are more likely to change their religion
than change their writing system.
--Charles Hockett, 1952

Webster, due to the popularity of his dictionaries, was able to drop the silent [u] in colour and implement a few other piece-meal reforms. However, Teddy Roosevelt's support of *thru for *through wasn't enough to sway public opinion.

From a practical standpoint, resistance to change, even change for the better, is almost insurmountable. Most scholars agree that having a spelling system similar to the one enjoyed by the Finns would increase literacy and speed the attainment of a 5th grade reading level by more than 25%. Due to the lack of awareness and a variety of mostly invalid reasons, few people beyond those who have studied the topic are ready to advance or endorse any proposal for a spelling reform.

The convergence among the proposals that have been advanced in the last 50 years can be seen in the following table. Note the close similarity of the three digraphic scripts, NS, WES, and ITA (The Initial Teaching Alphabet).⁶
Several of these systems,e.g., SpelRiet, Fanetic, and Inglish, World English, are well described on the Web.

Larger 40 character matrix   Sounds of English

For the last 50 years there has been convergence of opinion regarding the pragmatic solution to the problem of graphemic diversity. There are now three solutions depending on how much of the traditional system one wants to preserve and the level of phonemicity one wants to achieve:

(1) streamlined traditional,

(2) digraphic, and

(3) unigraphic augmented Roman.

Next Page

Next Page