Learning to read and write

The process of learning to read and write a language like English has to start by unpicking part of the process by which alphabetic writing was invented, a very difficult historical process over some 3,000 years. The history involved some giant leaps, each hundreds of years after the last. So necessarily this was the work of a series of scholars, all building on the work of others before them. It is often noted systems of writing were invented separately in China and Central America. But that is true only of the first step of representing a word by a symbol. The whole history only happened once. That gives an idea of how difficult that history really was.

For some children it is not easy to carry out a mental recapitulation of this history. It wasn’t for me. I still remember my utter confusion at being told that three quarters of a circle drawn in the air had something to do with a cat. There were things called letters. And they could be read in order. But the same principle did not apply to the word the. Why not? I asked myself. It took me two years to get the idea. I still remember the bright, sunny day the penny dropped.

Unpicking history

With the exception of four clever capital letters, not conventionally taught first, the alphabet is essentially a set of completely abstract and arbitrary representations of speech sounds or ‘phonemes’. What makes this hard to understand is that there is a three-way relation here – between

· The position of a phoneme in a word;

· The form of the letter representing that phoneme;

· The way the letter is formed by hand.

This three-way relation has to be unpicked in the mind of the learner as a kind of instantaneous insight before the learning process can begin. The scale of the unpicking should not be under-estimated.


Acoustically, human speech is a continuous stream of sound with no forcible separation between the words. Where the individual phonemes begin and end is not well-defined. 

Pictures for words – or ‘hieroglyphs’

The idea of pictures for words occurred at least three times, in China, in Egypt, and in Mexico, seemingly independently in each case. In the grand scheme of what a modern child has to learn, there is the idea of a word being representable. But there is a limitation in the fact that words like I and not and would cannot be illustrated in any perspicuous way. In the cases of  I, not ,and would , there is a work around by using characters for words that sound the same –‘homonyms’ – in these cases eye and knot and wood, and letting the reader work out the meaning from the context. But in as many cases there is no homonym.

Essentially, the now-abandoned ‘Look and say’ approach to the teaching of reading treats written words as though they were bad hieroglyphs, with no resemblance to what they seek to represent. But like a hieroglyph, the totality of a written word recapitulates the idea that a spoken word can be represented.

Arbitrary sets of wedge marks – ‘cuneiform’

Somewhat stretching the notion of a leap, the next giant leap, again in the Middle East, was taken over a period during which the representation of real world entities gradually disappeared, being replaced by ‘cuneiform’ characters, made in wet mud by a pointed stick. But like the hieroglyphs from which the system developed, the system was limited.

Letters for consonants

The greatest of the great leaps was the recognition somewhere in the Middle East between 3,000 and 4,000 years ago that a language could be represented by characters for the phonemes, taking the abstraction one step further. It is possible that the forms of these first letters (in the modern sense) had once been used to represent words beginning with a particular sound, a natural enough move, making the characters easier for the first users to learn, teach and remember. So if beth meant house, what had been a picture sign for house became the letter B. As an index of the cleverness here, the representation of phonemes only seems to have occurred once in human history. The characters for the sounds in M and N are recognisable as early forms of the modern capital letters. The new letters were conceived as having an order, probably for the sake of teaching and learning.

Then as now, the languages spoken around the South East corner of the Mediterranean, and in Somalia, Eritrea, Ethiopia, known as ‘Semitic’, nowadays including Arabic, Hebrew, Tigrinya, Amharic, and Maltese, have the vowels ‘inserted’ on the fly by rules according to the position and role of the word in the structure of the sentence. We have something a bit like this in English with sink, sank and sunk, as in “This ship is about to sunk” or “The ship sank in shallow water” or “The ship was sunk by one silly mistake.” But whereas in English, the root is clearly sink, and the vowel changes, in the Semitic languages, most or all of the roots consist only of consonants, and the vowel or vowels are added, rather than changed.

Somehow this unusual property of the Semitic languages may have facilitated the discovery of phonemes as the building blocks. By the original system, only the root consonants were represented because then, as now, for native speakers of a Semitic language, that was all, or nearly all, that readers needed.

Consonants and vowels

The new writing soon came to be used by traders to record their trades all over the Mediterranean. But it did not work for non-Semitic languages in which the vowels could not be omitted from the written form. Among these languages were early varieties of Greek. So between 2,700 and 2,800 years ago, system for showing consonants was reworked for a language in which the vowels were intrinsically part of the word structure. Letters for phonemes which did not occur in Greek were used for Greek vowels, including what was now easily recognisable as modern A. This was now the first modern alphabet, but with only one letter height. Soon the highly cultured Etruscans adapted the new alphabet as a way of writing their language. It may have been an Etruscan who systematised the design of what became the four capital letters, M, N, D, B. Then the up and coming Romans, who would eventually crush the whole Etruscan civilisation, adapted the alphabet for Latin with just 21 letters, later adding Y and Z for Greek words containing those phonemes, which did not occur in Latin.

The look of text

In the time of the Roman empire, people writing informally in Latin started forming some letters in ways going outside the ‘x-box’, in a style known as ‘minuscule’ or what we now know as ‘lowercase’, as opposed to ‘majuscule’. 

As the Roman empire collapsed and Christianity grew with a devotion to ‘holy writ’, monasteries in Northern Europe started producing beautifully executed and decorated bibles. The Kells Bible is one spectacular example. But these bibles were meant to celebrate the wealth, power and dignity of their royal or lordly owners, not to be read or studied. In 781 Charles the Great or Charlemagne, then the most powerful ruler in Europe head-hunted Alcuin, a famous poet and scholar from York, to run his bureaucracy, join his court, and teach him and his sons to read and write. Charles clearly understood that the running of the largest empire in the known world needed good, clear communications. Alcuin brought a tradition with him. But he developed it in ways that bear the clear stamp of one personality. It is now known as ‘Carolingian’ or ‘Caroline’ minuscule. It was (relatively!) easy to write clearly and fast. The readability was improved by short lines, spaces between words, the equivalents of full stops, commas, and question marks, sentences beginning with capitals, and some letters having ascenders, and others having descenders. Because Alcuin’s thinking was backed up by Charles, it was broadly followed.

700 years before the invention of printing, Alcuin laid down the look of modern text.

Syllables as ‘syllabaries’

For some languages, scholars have devised systems of single characters for the equivalents of key, cow, and car, pea, par, and paw, tea, tar and tore (in London English). One such language is Bengali. But this would be hard or impossible to implement in respect of a language as many possible syllables as English, with perhaps five thousand. (The doubt about the number of syllables partly hinges on syllables like scloo, which some speakers judge impossible as a word of English even though it occurs in the words exclusive, which most speakers can say without difficulty.)

But there are syllables in English, tion in station, tle in little, and for some speakers tre in metre, unstressed at the ends of words, all of them written in ways which might just as well be regarded as single characters because they have no other pronunciation and in some ways violate the normal rules, such as these are.

After the unpicking

Obviously, the unpicking of history is only the very first step in learning to read and write. Teachers have found clever ways of squeezing the extraordinarily long, complex, and difficult, historical process into just a few words spoken once – in my case on a single day by one memorably clever teacher.

But the difficulty of this is easily underestimated, particularly by the vendors of commercial schemes. The difficulty is not removed by the notion of phonics, useful as it is not to lose sight of the relation between the sound of B and the first letter in bat. By a consistent phonics, bury and berry should both be written beri. But even with that sort of consistency, there is still an irreducible problem for learners of English spelling in the fact that English has an inventory of around 44 phonemes, only 26 unaccented letters to write them, and almost 300 relations between letters and phonemes, known as ‘graphemes’.