view pages | download magazine

Conventional orthography is... a near optimal system for the lexical representation of English words.
- Noam Chomsky and Morris Halle, The Sound Pattern of English

I’m telling you this 'cause you’re one of my friends.
My alphabet starts where your alphabet ends.
- Dr. Seuss, On Beyond Zebra

Anyone who remembers learning to spell in English can probably think of reasons to complain about the English spelling system. The lack of one-to-one correspondence between spoken sounds and written symbols presents some very obvious problems. Since the Middle English period, this has been a subject of real concern to many, and many efforts at spelling reform were made consequently. Some, such as William Bullokar in 1580, suggested new alphabetic characters, while Richard Mulcaster and others preferred to use only pre-existing letters but to combine them in a more regular fashion. Attempts at spelling reform have continued into the present. The most famous efforts were led by George Bernard Shaw, who sought to create an entirely new alphabet for English, and by lexicographers Samuel Johnson and Noah Webster, who sought to apply existing spelling rules more regularly. Shaw was especially vitriolic in his protest against English orthography,[1] claiming, now quite famously, that "fish" could just as easily be spelled "ghoti," so unpredictable is the relationship between sound and symbol.

Much less obvious than its drawbacks are the advantages of the English writing system. Many critics of English orthography argue that the lack of one-to-one correspondence between sounds and symbols constitutes a general failure of the system. These critics assume that such a correspondence is the foremost, or perhaps the only, goal of a writing system. But many linguists and psychologists have begun to argue that there are other purposes served very satisfactorily by English orthography, such as it is.

A Short History

If it seems that our spelling system is badly suited to transcribe the sounds of Modern English, this is partly because that was not its original purpose. Most of our letters were borrowed from the Latin alphabet in order to transcribe Old English.[2] That language was served rather well by the correlation between its sounds and letters.[3] In 1066 CE, the Norman invasion began, changing nearly every aspect of Anglo-Saxon culture, including its language. The Norman influence is generally considered to have caused one of the most dramatic changes in the history of English. English assimilated huge numbers of French words, substantially changing the word stock of the language. Many French spelling conventions were adopted as well.[4] The language was also going through changes unrelated to the invasion, such as natural sound change and the significant reduction of its inflectional system. By the time the Norman Conquest was complete, the language we call Old English had become Middle English. By comparing the language of Beowulf (around 700 CE), which is nearly unintelligible to a speaker of Modern English, to that of Chaucer (1343-1400), which is much more familiar, one can see the vast changes that took place between the Old and Middle English periods. Many of these changes were still well underway when the English merchant and diplomat William Caxton first brought Gutenberg’s printing press to England in 1476. Caxton found himself faced with the problem of trying to represent words whose spellings were highly variable and whose pronunciations were changing rapidly. There was also the problem of settling on a spelling that would be intelligible to speakers of all dialects of English. The necessity of fixing spelling at a time when pronunciation was in such rapid flux did much to widen the gap between spelling and pronunciation that persists in English today.

Spelling Units

We can see, then, that a long and varied history has led to a very complex relationship between spelling and pronunciation. One argument in defense of Modern English spelling, though, is that it is not as irregular as it seems. Modern English has twenty-six letters with which to represent about forty sounds. The representational units of English, however, are not merely those twenty-six letters. When dealing with rules describing the relationship of symbols to sounds, the relevant symbolic unit is not the letter but what he calls the "functional spelling unit," that is, the symbolic representation of a single sound. The pair sh is a compound symbol that represents just one sound, that at the beginning of "shallow." It is an example of a spelling unit, although it is not a letter. Conversely, the single letter x represents two different spelling units, [k] and [s]; that is, x represents a compound spelling unit, whereas sh and b represent simple spelling units.

In order to discuss orthographic rules, we must distinguish between simple and compound spelling units. It is a well known, and relatively robust, rule that English vowels are "long" before a consonant-vowel pair (represented "CV") where C is a single consonant, and "short" in all other environments. In the word kosher, the o is long, because, although it is followed by a double consonant, that consonant represents a simple spelling unit. Conversely, in the word noxious, o is short, because, although it precedes a single consonant, that consonant represents a compound spelling unit. As this pair shows, thinking in terms of spelling units rather than letters reveals a greater regularity in the spelling of English than is otherwise visible.

It is also important to note that, when two spelling units correspond to the same sound, they almost always occur in complementary distribution: that is, of the spelling units in question, which one is used in the spelling of a word is determined almost completely by the surrounding combination of sounds in the word. Even if a sound can be spelled two ways, then, it is often clear from the context of other sounds in the word, or phonological environment, which is the correct spelling. The other sounds in a word, then, are a source of orthographic regularity.

Phonological Environment

Having said this, we can now look at the characteristics of current orthography. This survey of English spelling will begin to reveal a greater regularity than we find when our only criterion is a one-to-one correlation between sound and symbol.

When a sound can be spelled with either of two symbols, or when a symbol can represent one of two different sounds, we often say that the resulting spellings are irregular. Many of these three-way correspondences, although variant, are actually quite regular. For example, c commonly corresponds to both [s] and [k], but the correspondence is far from irregular. Which sound the letter corresponds to is determined almost entirely by its position within a word and by which

letter follows it. In this case, c almost always represents [s] before e, i, or y, and [k] before a, o, u, all consonants, and the end of a word. Similarly, where a sound occurs within a word or syllable can give information about its pronunciation. For example, syllable structure forbids that both k and n be pronounced at the beginning of a syllable, and thus we know not to pronounce the k in knight. This rule applies regularly.

In addition to a letter's position within words and the sounds adjacent to it, non-adjacent sounds can also play a role in determining what sound the letter makes. A good example is the silent e that changes the sound of a preceding and non-adjacent vowel (compare can and cane).

Rules concerning what sound a letter can make in a particular position within a word actually make it entirely impossible that "fish" could be spelled "ghoti" under the current rules of English orthography, since gh never corresponds to [f] in at the beginnings of words, ti never corresponds to [sh] in at the ends of words, and o only corresponds to [Í] in unusual and exceptional circumstances, such as in women.

Even sound-symbol correspondences that are, in fact, irregular, are still not haphazard. Sometimes these irregularities occur in words borrowed from other languages, such as the c (pronounced [ch]) in the Italian borrowing, fettuccine. In such cases, the seemingly irregular pronunciation of a letter can be explained by an alternative set of rules that apply predictably.

Morphology and Spelling

Other seeming irregularities have to do with the structures of morphemes, or the units of meaning that make up words.[5] For example, when two symbols normally interpreted as a spelling unit occur next to each other but belong to separate morphemes, they should be treated as separate spelling units. For example, the letters t and h that happen to meet at the morpheme boundary in hothead must not be taken as the digraph corresponding to [th] (the sound beginning the word think), but as two separate spelling units representing [t] and [h].

Sometimes the interactions of sounds at morpheme boundaries depend on the part of speech of one or both of the morphemes. That is to say that specific, sound-related rules govern the behavior of certain classes of morpheme. The unit ng may be pronounced [ŋ] (the sound at the end of thing) or [ŋg], as in finger, depending on the type of morpheme it touches. At the end of a word, it is always pronounced [ŋ]; before -er, if this morpheme is the comparative suffix for adjectives (found in faster and bluer), the unit tends to be pronounced /ŋg/, as in stronger.

The above rules and examples of their applications are intended to give an idea of the sources and the extent of regularity in the English spelling system as it exists today. In general, those things that appear irregular are still governed by a definite system of rules. Neighboring letters, position within a word or syllable, relation to certain non-adjacent letters, and the source language of a borrowed word, all bring in a specialized set of rules to modify as necessary any English sound. The relation of sounds to morphemes also affects how these sounds are spelled. These factors all increase the complexity of English spelling without affecting its regularity.

The give-and-take between regularity and variation in English orthography is a result of the fact that instead of representing only sounds, it tends to represent both morphemes and sounds (phonemes). English orthography is classified as morphophonemic for this reason. The interactions between morphology and phonology[6] are a major reason why spelling often deviates from pronunciation in English, and examining the effect of these interactions can reveal some of the advantages of the spelling system.

Words and morphemes are said to be in their base forms if they have no prefixes or suffixes attached. The base forms of words tend to be spelled very nearly as the word is pronounced. When base forms combine with affixes, or when they combine with one another to form compounds, they often retain their earlier spelling while changing pronunciation in response to new environmental conditions. We say, then, that base forms are spelled phonemically, and compounds and forms with affixes morphemically. In morphemic writing, the spelling of words serves more to remind us of the connection with their base forms than to tell us how they are pronounced. As an example of morphemic spelling, consider the affix -ed, used for the past tense of regular verbs: this morpheme is always spelled -ed, whether it is pronounced [d] as in dragged, [t] as in walked, or [əd] as in added . The same is true of a morpheme when combined with an affix: in the word photography, the root morpheme, photograph-, is spelled the same as the base form photograph, pronounced [fótəgræf-]; but in the affixed form photography, it is pronounced [fətágrəf-], in spite of the common spelling. In compound, affixed, and conjugated forms, the spelling refers not to a word’s pronunciation—which is in a sense more superficial—but to its underlying morphemes. The orthography of English is therefore called a deep orthography, as opposed to the shallow orthographies of Finnish or Spanish. In a shallow orthography, spellings change to reflect the altered pronunciation of inflected forms. Consider Spanish sacar and saquen, where the spelling changes to preserve the sound [k], which would otherwise be lost in *sacen.[7]

The processes of inflection and derivation do sometimes bring about spelling changes to better reflect pronunciation. This is especially common when a morpheme is pronounced like its base form but the spelling created by the process of affixation would suggest a different pronunciation. An example is the insertion of k in words such as trafficking and picnicking, intended to avoid the impression that the c should be pronounced [s] because of its position before an i.

Theoretical Bases for Evaluation

In their seminal work on phonology, The Sound Pattern of English, Noam Chomsky and Morris Halle outspokenly defend the English writing system. They outline some general claims about orthographies:

Conventional [English] orthography is... a near optimal system for the lexical representation of English words. The fundamental principle of orthography is that phonetic variation is not indicated where it is predictable by general rule. Thus, stress placement and regular vowel or consonant alternations are generally not reflected.

The key idea here is that orthographic systems tend toward economy rather than completeness or specificity in representing sounds. The linguist Edward Sapir writes:

Orthography is a system designed for readers who know the language, who understand sentences and therefore know the surface structure of sentences. Such readers can produce the correct phonetic forms, given the orthographic representation and the surface structure, by means of the rules that they employ in producing and interpreting speech. It would be quite pointless for the orthography to indicate these predictable variants.

Orthographies depend on a reader’s knowledge of a language. The aforementioned economy of the orthographic system depends on the reader’s ability to apply a set of rules to the information conveyed in writing in order to transform this information into the corresponding spoken sounds.

The idea of surface structures and underlying forms is crucial to Chomsky’s linguistic theory. We have encountered the idea already with respect to morphology: recall the morpheme -s that gives regular nouns plural number and its three possible pronunciations. Each of the possible pronunciations, [s], [z], and [əs], is a surface form. This surface form results from the application of phonological (sound-related) rules to the underlying form. According to Chomsky and Halle, orthographies work by representing underlying forms: readers are familiar with the surface structures of their language and will therefore be able to derive those surface structures by applying rules, which they know natively as speakers, to the underlying forms stored in the brain.[8] This is the basis for the claim that "phonetic variation is not indicated where it is predictable by general rule" and moreover, that "orthography is a system designed for readers who know the language." The orthographic system of English evolves from the way sound is naturally represented in the brain. It is therefore well suited to the language.

Although Chomsky’s and Halle’s remarks on the advantages of English orthography have caught the attention of many readers and become rather famous, The Sound Pattern of English is not a book about orthography, and it says little else on the subject. However, many others have written extensively on this topic. The question of the advantages of English spelling is of particular interest to education specialists who study reading and spelling pedagogy. The pursuit of effective techniques for teaching these subjects has motivated a good deal of scholarship on the subject.

One very successful account of the merits of English orthography is Brett Kessler’s and Rebecca Treiman’s article, "Is English Spelling Chaotic? Misconceptions about Irregularity." The authors acknowledge that English spelling does not have close correspondence between sounds and symbols, but such a correspondence, they suggest, is only one of many possible purposes of a writing system. They list other advantages of English spelling, namely conservatism, faithful spelling of loan words, and representation of morphological information.

Conservatism refers to the relative lack of change in spellings over time, something made complicated by the fact that natural sound change has dramatically altered the pronunciation of English words, creating a gap between the once consistent spelling and pronunciation of words. In the sense of creating this gap, conservatism is a disadvantage. However, it has some very concrete benefits as well. As the authors point out, conservative spelling allows for a unified writing system that can be understood by any speaker of English, regardless of accent, dialect, and idiolect. A speaker, they continue, would have no motivation to try to learn a new, highly phonetic spelling system that does not represent the dialect he or she speaks.[9] Conservatism also has the related advantage of allowing a single writing system to represent both current and older forms of English. The universal inconvenience of an orthography that represents both sounds and morphemes yields the universal convenience of a writing system that can be used and understood by any literate speaker of English. In this way, no generation of speakers has had to undergo the difficult and artificially imposed transition to a new writing system, which would be nearly impossible to bring about from a practical standpoint.

For reasons related to conservative spelling and faithful spellings of loan words, spellings can often communicate other kinds of information besides how to pronounce a word. For example, spellings can help distinguish between homophones (words with different meanings that are pronounced the same). Kessler and Treiman give the well-chosen example of sight, cite, and site, which would be spelled the same were it not for the orthographic conservatism: this tendency has preserved a now silent gh in sight, and if the original spellings of the Latin roots in cite and site had not been retained, the sound beginning both words would presumably be written with the same symbol. The ability to make this distinction is one of the advantages of the spelling system for readers of English.

Kessler and Treiman also point out that spellings can help distinguish between function words, such as but and in, and content words, such as table and jump. Among other things, content words almost always have at least three letters. The authors point out that the spelling of inn would seem irregular and look strange were it not for the fact that the double consonant saves it from being unusually short.

It is clear enough that shallow orthographies are more convenient for spellers, young children, sufferers of reading disorders, and non-native speakers of a language. In order to best evaluate the deeper orthography of English, it is necessary to figure out whether the benefits of the system to fluent readers outweigh these inconveniences. One question is whether readers of English have a better understanding of English grammar because of their frequent dealings with the morphemic structure of their language. Perhaps an easier thing to measure is the speed and efficacy of reading in English. There is some evidence that English spelling, due to its deep structure, resembles a logographic system[10] to some degree, in that readers take in whole words rather than individual sounds. Because whole clusters of letters represent units of meaning in English rather than individual letters simply representing sounds, it is easier to read words by merely glancing at them rather than actually parsing each letter. Studying the movements of a reader’s eye can yield information about how written words are processed. This sort of whole-word processing would also explain, for example, the fact that a person can read words written with all the vowels omitted.[11]

This sort of investigation has particularly important implications for teachers of reading and spelling. If we read through bottom-up processing, by noting each letter in succession, figuring out what sound it makes, and then piecing them together into a whole word, then a phonics-based approach to spelling and reading makes sense. This approach emphasizes the predictable correlations between sounds and letters. If, on the other hand, readers use top-down processing, taking in groups of letters that collectively represent a particular morpheme, then a whole-word, memorization-based approach to reading is in order. Most teachers use elements of both methods, since top-down and bottom-up processing both take place to different degrees. Reading and spelling pedagogy has benefited greatly from research on the interaction between these two kinds of processing and the circumstances in which they take place.

Thanks to Professor John Huehnergard for providing the inspiration for this paper and for helping me to write it.

---

---

© 2007 Tuesday Magazine / a student-run organization at Harvard College
The Harvard name is a trademark of the President and Fellows of Harvard College and is used by permission of Harvard University.
faith h. zhang / webmaster