How was Linear B deciphered?

Linear B was deciphered in 1952 by Michael Ventris, a young British architect. He worked from a grid of signs that earlier scholars had made possible, then used guesses about place-names to fix the first sound values. The script turned out to record an early form of Greek, which Ventris had not expected. John Chadwick, a Cambridge classicist, was the first scholar to accept the decipherment, and he later collaborated with Ventris. [1] [2]

Nobody could read the script or knew its language. Proposals ranged from an Anatolian language to Etruscan, and few scholars took Greek seriously, largely because Arthur Evans had dismissed it. Two things were clear. Linear B has 87 phonetic signs, too many for an alphabet and too few for a logographic system, so it was probably a syllabary. The tablets were also administrative records, with ideograms for people, animals and commodities that allowed scholars to infer roughly what a tablet was about. [2]

Emmett Bennett Jr. published the Pylos tablets in 1951 and made the first systematic classification of the signs and their variant forms. Only after that could valid statistical analysis be done. Alice Kober, who taught Classics at Brooklyn College, studied groups of words that differed by a single sign and used them to show that the language was inflected, with endings changing for properties such as gender and number. This allowed her to begin building a grid showing which signs shared consonants or vowels, without knowing their sounds. [2]

Kober refused to speculate about the language or the sounds. She recorded sign frequencies and combinations by hand on more than 180,000 index cards cut from scrap materials. She died in 1950, before the decipherment. [3] [2]

Ventris corresponded with other researchers, circulated a 1950 questionnaire-based report and then a series of Work Notes, and refined the grid. In early 1952 he made several guesses. One sign, common at the start of words, might be the pure vowel a. Comparison with the already deciphered Cypriot Syllabic script suggested another sign was ni. He also guessed that some words in tablet headings were place-names. Looking for Amnisos, the harbour of Knossos, he found a word that fit, which gave values for mi and so. The grid then let him deduce others, and he arrived at ko-no-so, Knossos itself. [2] [4]

As more words became readable, the language proved to be Greek, though Ventris had long thought Etruscan or a related language most likely. On 1 July 1952 he announced on the BBC that the tablets were in a difficult and archaic Greek, about 500 years older than Homer. [2] [5]

Chadwick heard the broadcast and was convinced within days; his wartime code-breaking experience and knowledge of Greek helped him recognise the solution. The decipherment was confirmed in 1953 by newly found Pylos tablets. One has the word ti-ri-po-de, Greek tripode, meaning two tripods, followed by a three-legged vessel ideogram and the numeral 2. [2]