HALF a century ago this year the cracker of the Knossos code, Michael Ventris, died. The distinguished British anthropologist Sir Mortimer Wheeler said of him: "He died like all geniuses at the age of 34." Ventris (1922-1956) radically changed the world's views on when the Greek language was born.

The story starts in 1900 with Arthur Evans excavating Knossos in Crete and finding clay tablets marked with indecipherable signs. He called the script 'Linear B' to distinguish it from Linear A - whose ideographic code is still uncracked - and decided that it could not be Greek. Minoan Crete - which lasted from 2500BC until 1400BC - had, he insisted, nothing whatever to do with the civilisation of mainland Greece that succeeded it. The existence of alloglot (speaking a different language) Cretans is even attested by Homer who wrote in The Odyssey:

Out in the middle of the wine-dark sea there is a land called Crete, washed by the sea on every side; and in it are many peoples and ninety cities. There, one language mingles with another... Among the cities is Knossos...

In his fascinating biography The Man Who Deciphered Linear B, Andrew Robinson describes how Michael Ventris, at the age of 14, first came across the puzzle that would dominate his short life. He and some schoolfriends visited an exhibition of Greek and Minoan art at Burlington House in London. By chance Evans was there and he volunteered to show the boys around the exhibition. Michael, the youngest of the party, intrigued, said: "Did you say the tablets have not as yet been deciphered sir?" "Yes," answered Evans. Fourteen years later Ventris cracked the code with the help of John Chadwick, a professional scholar, who explained the whole story in his book The Decipherment of Linear B, published in 1958.

Ventris had a head start in matters linguistic. His half-Polish mother taught him her native language; he also went to school in Switzerland for two years and thus from a tender age became fluent in English, French, German and Polish. His family background was less than happy, a fact that weighed heavily on his short life. His father - British in spite of his foreign-sounding surname - was an officer in the Indian army who retired early because of tuberculosis that killed him in 1938. The young Michael was deeply affected by his parents' divorce when he was 13 and even more so by his mother's suicide when he was barely an adult. Robertson portrays the young Michael as very bright, fascinated by all things cryptic, timid, gentle and unconventional, with a wayward sense of humour and somewhat forlorn at times. The dominant influence on him came from his mother whose social circle included the architect Marcel Breuer, the sculptor Naum Gabo and the painter Ben Nicolson.

After school, Ventris trained as an architect but never ceased to battle with the Linear B challenge. In order to explain how he succeeded where many a classical scholar had failed, Robinson suggests that Ventris was not only free of prejudices but was also blessed with the gift - useful to architect and decipherer - of understanding the underlying structures of complex patterns, of how to solve problems within narrow constraints and how to be as meticulous and thorough as necessary in order "to combine the functional with the visible".

The first hurdle Ventris had to overcome concerned the language written in the Linear B script. With no "Rosetta stone" - conveniently carved in 196BC in Egypt in three scripts (Hieroglyphic, Egyptian demotic and Greek) - to help the decipherer, Linear B remained an enigma. Hittite, Etruscan, even Basque were mooted as possible languages while most classicists ruled Greek out. The reason was the accepted belief that Greek - the earliest Indo-European language and script - emerged for the first time in the eighth century BC with Homer's Iliad. Then in 1939 Carl Blegen, an American archaeologist, found 600 tablets in Linear B in the Peloponnese on mainland Greece in the ancient palace of Pylos.

The next development that pushed Ventris in the right direction was the discovery by the American scholar Alice Kober - who studied the Linear B tablets closely between 1943 and 1950 - that the ending of the words changed in a predictable way. This meant that the language was inflected, as Greek is. Ventris tried first to squeeze as much information as possible from internal study before trying out any languages to see if they fitted. He concluded that the script was syllabic - each symbol representing a consonant plus vowel combination. This form differs both from an ideographic script in which one symbol represents one word - such as Chinese with its thousands of characters - and from an alphabetic script such as English in which a small number of characters represent the sounds that make up words. In 1952 Ventris tried out an idea that seemed revolutionary at the time: He assumed that certain words in the Knossos tablets might be place names. He then substituted his experimental phonetic values for the signs and found himself reading the familiar names of Cretan towns in Greek.

Thus in 1953, Robinson tells us, three important things happened. The structure of DNA was explained, Mount Everest was climbed and Ventris broke the Linear B code. His discovery did not reveal any new Homeric epics, only palatial business, but it showed that 14th century BC Crete was administered by Greek-speaking monarchs, a complete reversal of received opinion. Ventris died three years later when he drove his car into a lorry. Did he commit suicide because he could not face going back to architecture having run out of a major purpose in life? Some think so.