During the Iron Age, the island of Cyprus did not use the common Greek alphabet, but rather its own syllabary, which was deciphered in the 19th century thanks to bilingual inscriptions in Greek in this script and in Phoenician. This syllabary seems to have been designed for a language with a consonant-vowel syllable structure, like Japanese, which does not work well for Greek, which has words like anthropos; the script accommodated such words by using silent vowels. In the early 20th century, excavations in Crete found tablets from the Bronze Age in two scripts, named Linear A and Linear B; later Linear B was also found on the Greek mainland. Linear B had obvious numerals (1, 2, 10, 20, 100, 200...), units of measure, obvious logograms (man, woman, horse, pig, chariot, tripod vessel...) and about 90 characters - too few for ideographic writing, and too many for an alphabet or an abjad. Matching them with the Iron Age Cypriot script did not lead anywhere; it was not clear, what language these tablets were written in, but suppose the two syllabaries were built on the same principles? Then an American woman scholar realized that many words in the script are the same except for the last few characters, which may indicate a language with case endings (e.g. nominative - anthropos, genitive - anthropou...). What if the first characters of different endings have the same consonant but different vowels? In the early 1950s, the author and another classicist, an architect who was a navigator in the RAF during World War II, built up a grid of classes of syllables thus related; there were five such classes corresponding to five vowels. When they found likely toponyms such as Knossos and plugged them into the grid, what emerged was an early form of ancient Greek, as far from classical Greek as Beowulf is from Shakespeare! Unlike the Iron Age syllabary, it dropped the final consonant. Many classical scholars questioned their analysis, but then another batch of tablets was excavated; on one tablet the picture of a tripod was accompanied by syllables ti-ri-po (tripos in classical Greek), two tripods - by ti-ri-po-de (tripodes, the dual number in classical Greek), a vessel with four handles - by qe-to-ro-we (in classical Greek, four is tetra, but in other IE languages the first consonant is k; -owe is owes, or ear), a vessel with three handles - by ti-ri-o-we-e and another - by ti-ri-jo-we, and a vessel with no handles - by a-no-we. Although the corpus of writing in this script is limited to administrative records of Bronze Age palaces, this is the language spoken by Homer's heroes. It has the consonant q in some places where classical Greek has p (e.g. man is a-to-ro-qo, horse is i-qo, but then the Latin aqua became apă in Romanian). This was a less complicated operation than the almost contemporaneous breaking of the Enigma, but it was carried out by fewer people with less resources. Linear A remains undeciphered to this day.