Introduction and Linguistic Background

Spelling does not only refer to the printed word, but spelling first and foremost means an abstraction of information from the printed form that leads to a correct result. In this, people make use of phonological, orthographic, and morphological types of information. If there was a one-to-one correspondence between sounds (phonemes) and letters (graphemes) to produce accurate spelling would be simple. However, in English there is no direct one-to-one mapping of sounds to letters. The following are some examples to illustrate this problem: The sound /k/ may have the graphemic representations k, c, ck, or ch, although this might be dependent on the position in the word. However, the position itself is not a reliable cue as can be seen from the example of the grapheme s which may sound /s/ or /z/ (as in save or zoo, respectively). Vowels are even more problematic as their graphematic counterparts may consist of many different letters (e.g., hey, say, weigh, mate, gait, straight; Venezky 1979). English has much more sounds than letters and these sounds may change according to context and may influence each other. There are graphemes that have no corresponding sound in certain contexts as for example the g in sign, (though: signature) or some of them only transport a signal, not a sound, as for example double consonant letters that usually shorten the preceding vowel. 

Another difficulty are words that can only be identified from the context. Words like tear  and tear  are called homographs, they look the same in writing and their pronunciation can only be drawn from the context.. The equivalent in sounding are homophones, as for example in broach  and brooch . (An often cited poem by an unknown artist may illustrate this difficulty best.)

There exists quite a number of scientific studies dealing with the question of how spelling can succeed. Where does this spelling proficiency come from, how does it develop, and what kind of problems will it face? After all, according to Steffler (2001:169) one fact seems to be certain: Due to the complexity of written language, it is not probable that spelling development is determined by explicit learning alone.
 

Factors Guiding Language

Apart from the mechanisms behind spelling there are also other influencing factors that are of interest. Putting it rather naively, languages develop along two competing factors. One wants as much linguistic contrast as possible, the other wants processing as simple as possible. In their extreme, these factors would mean on the one hand: 'Be as precise as possible in your pronunciation so that your language may realise the most possible number of contrasts.' In this way the smallest sound change would result in a change of meaning, as can be traced in the example of the homographs from above (tear  vs. tear ). In a similar way boot [bu…t] and root [ru…t] form a minimal pair, as well. Those sounds of a language that may change the meaning of a word make up the phoneme inventory of  a language. And the number of phonemes and the number of combinatory possibilities makes up the degree of linguistic complexity of a language. The opposing factor, however, demands a structure to be as unmarked as possible and therewith to be as effortless in processing as possible. The most unmarked form, apart from no form at all, would possibly be a form like táta with two simple syllables in a trochaic foot and the sounds being maximally distinct from one another. These two factors mentioned at last lead to the variability among languages and to a diachronic variability within languages. 

That English script has not undergone a change towards its pronunciation for over a thousand years, on the one hand may be seen as a characteristic of the enormous capability of the cerebral processing mechanisms in humans. But it has also lead to this amount of discrepancy of sound-to-spelling correspondence and the problems therein. (For further reading on attempts to change or regularise English spelling and why this has not taken place so far, please read 'Popular Movement Against Traditional Spelling'.) However, apart from all difficulty in processing, scientifically seen the English language is a kind of goldmine in that it offers complex possibilities of investigation to this matter.

The following paragraphs give a short introduction to the spelling mechanism and development. As there is, however, an enormous amount of literature on this topic, readers interested in further reading will find a list of links.
 

Spelling Mechanism and Development

As regards the complexity of the English writing system the following result may be regarded as somewhat reassuring. Hanna et al (1966) counted the probability of spelling a word correctly if one applies the letter string that most often corresponded to the phoneme. Their findings were that 73% of the phonemes in their dictionary would be spelled correctly by that method. However, the probability that a word would be spelled entirely correctly in this way was much smaller. The question that will be asked most often at this point is if readers and spellers therefore must rely on memorization of whole words or is there some kind of middle ground in between?

In search for the latter, Venezky (1970) has shown that there seem to be context conditioned rules as well. In most cases where a pattern is disambiguated by an adjacent letter, the influencing letter is to the right of the letter(s) in question.

All in all, the mastery of English spelling presents a greater challenge than learning to read and most people are able to read more words than they can spell accurately even with advanced learners (Holmes & Carruthers 1998). When getting repeated exposure to the written word people begin to develop an orthographic representation in memory that contains the word's spelling as a kind of code. According to Barry (1992), people refer to this kind of knowledge base in spelling even in very regular languages rather than solely referring to rules. A number of models in this field come to the conclusion that there might be a complete separation between reading and spelling as there are people who are able to read and not to spell and also people who can spell but not read. 

Misspelling a word itself may have various consequences. According to Stanback (1992), about 75% of the words used in text are monosyllabic. Misspelling of a monosyllabic word will more probably lead to a confusion of the target word with other words. A failure to code the l of the word fold  would mean that it would be confused with fond, food, and ford. The orthographic representation for short words, especially monosyllables, must therefore be accurate. Longer words like separate contain letters whose identity is not essential for their identification. A spelling like seperate would not lead to a confusion with another word. Hence, in normal readers the stored information may be inaccurate or incomplete with regard to perfect spelling but still adequate enough to support identification in reading. The same contrasting models of spelling vs. reading assume that there is an input orthographic lexicon which will be used in reading and an output orthographic lexicon which functions for spelling (Ellis 1993). 
 

Educational Advice for Poor Spellers

When mature learners produce spellings under optimal and careful conditions, reading and spelling processes are occurring simultaneously. Adults spellers also monitor their spellings visually because they are aware that they may sometimes make mistake. Young children, however, will rather not verify their spellings to see if what they have written matches with what they meant to write. 

When children start writing, certain strings will regularly recur and thereby become familiar. In this way spelling becomes automatic. It has been shown that visual memory plays an important role especially in learning irregular sequences and that this may be one factor underlying poor spelling (Bradley 1983). The author's advice for dealing with backward readers is to help them use not only their auditory sense, i.e. to phonemically analyse the target word but to provide them with visual and tactile stimuli, as well. She found that plastic script letters may be a method of learning to generalise from one word to another that different words can share common units. Another advice in the same direction has been given by Rourke (1983:229), who found that first and second grade children strongly rely on a phonemic analysis at the expense of other skills. Children showing a degree of spelling disability at this age should preferably be taught on orthographic relationships.
 

References

Barry, C. (1992). Interactions between lexical and assembled spelling (in English, Italian, and Welsh). In: Sterling, C. M. & Robson, C., eds. (1992). Psychology, spelling and education. Clevendon, UK: Multilingual Matters. 71-86.

Bradley, L. (1983). The Organization of Visual, Phonological, and Motor Strategies in Learning to Read and to Spell. In: Kirk, U., ed. (1983). Neuropsychology of Language, Reading, and Spelling. London: Academic Press.

Ellis, A. W. (1993). Reading, writing and dyslexia: A cognitive analysis. Hove, UK: Erlbaum.

Hanna, P. R., J. S. Hanna, R. E. Hodges & E. H. Rudorf (1966). Phoneme-grapheme correspondences as cues to spelling improvement. Washington, DC: U.S. Department of Health, Education, and Welfare.

Holmes, V. M. & Carruthers, J. (1998). The Relation between Reading and Spelling in Skilled Adult Readers. Journal of Memory and Language 39. 264-289.

Rourke, B. P. (1983). Reading and Spelling Disabilities: A Developmental Neuropsychological Perspective. In: Kirk, U., ed. (1983). Neuropsychology of Language, Reading, and Spelling. London: Academic Press.

Stanback, M. L. (1992). Syllable and rime patterns for teaching reading: Analysis of a frequency-based vocabulary of 17,602 words. Annals of Dyslexia 42. 196-221.

Steffler, Dorothy J. (2001). Implicit Cognition and Spelling Development. Developmental Review 21. 168-204.

Venezky, R. L. (1970). The structure of English orthography. The Hague: Mouton.