The Para-Rigger

The Alphabet
by David Sacks
400pp, Hutchinson, £12.99

In the 1990s there was plenty of work for sign-writers in Baku. Two years ago the government of the former Soviet republic of Azerbaijan declared that it had completed an expensive and disruptive transformation over the previous decade. The country had changed from the Cyrillic alphabet used by Russian to the Roman letters used by English, among other languages. The written language had been liberated from Soviet influence.

Elderly citizens might feel justified in being confused; for them it was the second such shift. Before Stalin imposed Cyrillic lettering in 1940, they had used the notation of Arabic. Coincidentally, the language on which the new Azeri alphabet is closely modelled, modern Turkish, itself switched from Arabic to Roman letters in the 1920s, under Kemal Atatürk's westernising influence.

Alphabets are not native to languages. Their histories reveal the strange chains of artifice and accident that link written letters to spoken words. Such sudden changes as those in Azerbaijan or Turkey may be politically manufactured, but are modern instances of age-old processes. For the most important, extraordinary fact about alphabets is that, as David Sacks puts it, they have "routinely jumped from language to language".

His book sets out to show how our so-called Roman alphabet (though the Romans had no J, V or W) evolved from others. It is not the first such history of the alphabet for the general reader, but it is an especially engaging one. Sacks's clever, simple idea is to follow the individual letters, one by one. He takes us back in time to find how each came to us and how it gained its special properties. In the process, the eccentricities of English spelling and pronunciation become intelligible, even weirdly ingenious.

Shared alphabetical signs do not imply that languages are closely related. When languages have passed from illiteracy to literacy, they have simply needed to find letters, and have taken them from some nearby source or impressively clever group of foreigners. Sets of letters are always purloined from somewhere else. At each stage, as an older alphabet is fitted to a newer tongue, there are changes. New sounds are affixed to old letters. It is extraordinary how far and how clearly we can see back to the origins of our letters. English took its alphabet from Latin (as did many a language that the Romans never heard spoken, from Polish to Zulu to Indonesian). Latin itself was written with letters copied from the utterly dissimilar Etruscan language, a tongue still largely unintelligible to us. A few centuries before this happened, the Etruscans had appropriated the Greek alphabet, even though, again, the languages had little in common. And the Greeks had taken their letters, with minor adaptations, from the Phoenicians, though the two peoples' languages were "as different as Arabic and English".

The Greeks, for instance, made some Phoenician consonants, such as A, stand for vowels. (Phoenician writing, like Hebrew, used only consonantal lettering, leaving vowel sounds to be inferred by the reader.) Thus Greek made its alphabet especially adaptable to other languages. In another example, Old English added two runic signs for the sounds "th" and "w", unrepresented in the Roman alphabet.

Recovering such history makes it clear that one people had a uniquely powerful influence on how others developed writing. These were the Phoenicians, a Semitic Iron Age people based in what is now Lebanon. By 1000BC they had a versatile alphabet; their trade network transported it around the Mediterranean. Commerce made their letters powerful.

The Phoenician system is the ancestor of most of the world's alphabets: not just our own, but others such as Hebrew, Arabic and the Devanagari and Bengali scripts of India. Perhaps 19 of our letters have Phoenician counterparts. The shapes of some of these are extraordinarily intact. So the 12th letter in the Phoenician alphabet stood for an "l" sound. Called lamed (pronounced "lah-med") it meant "ox goad" and imitated the shape of a stick with a crook handle for poking livestock. There, in the very same place in our alphabet, is L, the same ox-herder's stick. There is something as miraculous in these forms as in the most beautiful relics of ancient civilization.

Sacks takes us back beyond the Phoenicians. When his book offers insight into "the archaeology of language" it is not speaking metaphorically. Our familiar 26 letters have distant ancestors in ancient Egypt, legible in writing on rocks in the Sinai desert. Inscriptions unearthed by archaeologists make visible one of the singular intellectual revolutions of human history.

At an ancient turquoise mine, Serabit el-Khadem, foreign workers left us the earliest traces of alphabetic writing. You can see, even in this book's smudgily reproduced photographs, how, almost 4,000 years ago, Egyptian hieroglyphs became adapted to the representation of sounds rather than things or ideas. These ancient etched signs suddenly seem close. The Egyptian hieroglyph for a hand, for instance, became a schematic diagram for the Semitic letter "kaph", which meant "palm of the hand" and stood for the "k" sound. Thence, via Phoenician to Greek and the rest, it became our letter K. You can still see the four fingers.

Follow any letter back through time and you glimpse the extraordinary commerce of language across centuries. After an introductory overview, Sacks gives a chapter to each of our letters, seeing how its shapes and sounds might have changed. Many of our letters, like B or E, are mirror images of early Greek originals, for this, like other ancient languages, could be written alternately left-to-right and right-to-left: boustrophedon, "as the plow-ox turns". Some of the irrationalities of how English pronunciation is presented have ancient reasons that you would never guess. The fact that the commonest sound of the letter C is also represented by K or Q turns out to be a relic of Etruscan, which had three slightly different, irrecoverable "k" sounds.

Just think of the strange things that H does. On its own it is but an expelled breath, hardly any sound at all. An authoritative Latin grammar of AD500 declared it to be not a true letter, merely a symbol announcing the pronunciation of the letter next to it. Yet this half-consonant is everywhere in written English, in CH, GH, PH, TH, WH and (oddest of all) RH. It is "the life and quickening", as Ben Jonson put it, of other letters.

In fact, these uses of H are the living traces of an alphabet's movement between languages. The Romans began using the Etruscan H when transliterating Greek words containing sounds without a Roman equivalent. So the Greek theta, representing a sound foreign to Latin, became TH. So too phi, khi and rho became PH, CH and RH, and originally represented breathy sounds that were distinct from F, K and R. Thus we have philosophy, chrome and rhetoric. A thousand years later the Normans used H to eradicate the non-Roman letters of Anglo-Saxon. The letter yogh, representing the breathy English "g" or "y" sound was to be GH. So, too, we got TH for the "th" sound in "then" and CH for the "ch" in "cheese".

This book is rich in such explanations and the essays on the individual letters digress in many fascinating ways. You can find out how the writing of lower-case letters developed via the wonderful varieties of uncial and Carolingian minuscule; how conventions for Roman inscriptions live on in modern typefaces; which vowel sounds appear universal to human utterance; and why X represents what is unknown (the printer of Descartes's 1637 La Géométrie ran short of Ys and Zs). You will also be given gentle lessons in how to describe the way the human mouth makes intelligible noises - discovering aspirates and fricatives, liquids and labials - and why, therefore, Japanese people turn "l" sounds into "r" sounds and the French do not pronounce H.

Sometimes the page layout cannot marry all the information with Sacks's yen for telling a story. There are inset illustrations with their own little essays attached, digressions in different fonts and interpolations on specialist topics. So much to fit in! There are sometimes interpolations within interpolations. Yet, however untidy, it is a wonderfully interesting book, readable yet unafraid of complexity. It is apparently based on a series of weekly columns that Sacks wrote for the Ottawa Citizen newspaper. Those Canadians must enjoy something erudite with breakfast.

· John Mullan is senior lecturer in English at University College London