A mere week after producing the third grid, Ventris’s imagination was clearly in overdrive. The second volume of Scripta Minoa was yet to arrive, but still he wrote briefly to Myres about a particularly intriguing experiment he had tried over the weekend. For some time he had suspected that the Knossos tablets studied by Kober in the mid-1940s contained place names, and he knew that Myres agreed with him. In addition, he had noticed that Kober’s particular ‘triplets’ occurred only in the tablets from Knossos, never in the tablets from Pylos. Could each ‘triplet’ refer to a different town in Crete? Now he noticed that if he made ‘only a little adjustment’ to the phonetic values that he had guessed in the grid, and then substituted them for the corresponding signs in Kober’s ‘triplets’, a very interesting result emerged: what appeared to be archaic Greek names of three well-known Cretan cities, including Amnisos (the port of Knossos) and Knossos itself! ‘This is one of those guesses it’s best to keep up one’s sleeve, because there’s an extremely good chance of its being completely wrong’, he told Myres cautiously. If correct, it suggested (but did not prove, since the words were place names and not common nouns), that the language of Linear B might be Greek. Ventris must have felt like Evans with the foal tablet (page 36): each man looked the ‘Greek solution’ in the face for a moment but pulled back on the brink of accepting it because he preferred another solution – ‘Minoan’ for Evans, ‘Etruscan’ for Ventris.
Yet he could not ignore his hunch. At the end of April, after receiving and immersing himself in Myres’s massive edition of Evans’s tablets from Knossos (and also dashing off some immaculate architectural plans for his new house), Ventris told Bennett: ‘It would be a wonderful thing if one could sit down on the hill at Knossos and know just what the names of all the surrounding towns and villages were in Late Minoan times; because I’m sure some of them must occur [in these tablets]. I’m still rather intrigued by AMNISOS for [one of Kober’s “triplets”].’ As May wore on, he could not resist making further experiments with some phonetic values using his grid.
He began by allotting a to the ‘double-axe’ , pure vowel 5 on the grid. Its already-mentioned high initial frequency in Linear B and similar behaviour in other languages suggested this identification to Ventris (and before him, Kober and a Greek scholar, K. D. Ktistopoulos). For his second guess he turned to the Cypriot clue once tried by Evans: comparison of the shapes of Linear B signs with those of the Cypriot script. Ventris had studiously avoided using this comparison until now – at least until the making of his third grid in early 1952 – because (like Evans) he distrusted it; but he had continued to believe in some kind of historic link between the languages of Crete, Cyprus and the Aegean area. He now hazarded that Linear B was equivalent to -na in Cypriot, and that Linear B was equivalent to -ti in Cypriot. (He also adduced evidence from Etruscan, just to confuse right with wrong reasoning!) If these guesses were right, then consonant 8 on the grid must be n and vowel 1 must be i, which automatically meant that was ni, according to the grid.
Ventris’s next step was inspired: a ‘leap in the dark’ of the kind mentioned by Chadwick (see the Introduction). He decided to pursue his February hunch about the names of Cretan towns, and see where the hypothesis would lead him.
The town of Amnisos, being the ancient port of Knossos, he knew was likely to be mentioned in the tablets. But Amnisos was its classical Greek name. Its name at the time when it was written syllabically in Linear B would, Ventris proposed, be A-mi-ni-so, with no final s – the ‘s’ being the classical Greek ending also found, as we saw before, in the later Cypriot script. (Evans had had the same idea about Linear B spelling when he speculatively equated Linear B ‘polo’ with classical Greek ‘pōlos’.) In introducing this assumption, that Linear B spelling might have omitted final s in nouns, Ventris was undoubtedly sticking out his neck, since he had no supporting linguistic evidence at all for such a spelling change between archaic Greek and classical Greek – but he plunged on.
Written in Linear B, A-mi-ni-so would be:
The first word in one ‘triplet’ was . If it meant Amnisos, then
Then, according to the grid, consonant 9 must be m, and vowel 2 must be o. This would mean, in turn (the so-called ‘chain-reaction’) that = no.
The first word in another ‘triplet’ was . Using the grid, this transliterated as ?-no-so. If = ko, the name could be Knossos itself. In due course Ventris was to extract from the five ‘triplets’ the names of three further known Cretan towns:
Tu-li-so (Tulissos)
Pa-i-to (Phaistos)
Lu-ki-to (Luktos)
An entire ‘triplet’ could now be transliterated as:
A-mi-ni-so (Amnisos)
A-mi-ni-si-jo (Amnisian men)
A-mi-ni-si-ja (Amnisian women)
These proposed meanings of the second and third words, though as yet unproven by Ventris, reminded him of words with similar inflections in Homeric Greek, and were therefore promising. It looked as if Kober’s ‘triplets’ were not, after all, noun declensions but something similar to the ‘alternative name endings’ he had originally proposed back in 1948: proper names of towns and their ethnica (e.g. London/Londoners, Paris/Parisians).
Thus proper names, vital in the Egyptian hieroglyphic decipherment, now looked as if they would prove to be vital in deciphering Linear B too. Instead of Ptolemy and Cleopatra, there was Amnisos and Knossos. But everything was still speculative, dependent on the initial assumptions for the phonetic values.
Having worked out various other ‘Greek’-looking words in the tablets, Ventris was unable to contain himself. On 1 June, he sat down to type what would turn out to be his final Work Note, number 20, and boldly titled it: ‘Are the Knossos and Pylos tablets written in Greek?’ At the beginning, he called the short, five-page note ‘a frivolous digression’, ‘not intended to prejudice’ the serious analysis of the newly published Knossos tablets. Then he gave the exciting evidence about the Cretan town names and a handful of other words. At the end, he cautiously ‘covered’ himself: ‘If pursued, I suspect that this line of decipherment would sooner or later come to an impasse, or dissipate itself in absurdities; and that it would be necessary to revert to the hypothesis of an indigenous, non-Indo-European language.… But this fantasymaybe the excuse for us once more to ask ourselves the question: Which is historically more incongruous, a Knossos which writes Greek, or a Mycenae which writes “Cretan”?’ During the next year, the Knossos, Pylos, Mycenae and other Linear B tablets would be compelled to deliver up an incontrovertible answer.
6
Breakthrough
‘During the last couple of days I have been carrying on with the fantasy I discussed in my last Note; and though it runs completely counter to everything I’ve said in the past, I’m now almost completely convinced that the…tablets are in GREEK’
Michael Ventris in a letter to Sir John Myres, mid-June 1952
We cannot speak of a precise instant of breakthrough when Ventris deciphered Linear B and revealed its secrets – as when Howard Carter’s candle suddenly illuminated Tutankhamun’s concealed tomb. But neither was the crucial insight a long drawn-out process. The key period was undoubtedly two or three weeks in late May and early June 1952. Lois Ventris, Michael’s wife, remembered being woken by him in bed in the Highpoint flat at about 2 a.m. ‘with a long story about place names like Amnisos and symbols for chariots and so on, all of course with illustrations’. In early June, there was another memorable moment. It happened when the Ventrises invited to dinner an architect friend, Michael Smith, and his South African-born wife Prudence, a post-war classics student at Somerville College, Oxford (where she had attended the famous and eccentric lectures of Sir John Myres), who in 1952 was a BBC radio producer. These two knew Ventris as a skilful and hard-working architect when he was at the Ministry of Education, who also had a ‘hobby’, the Minoan scripts, which they regarded as his amusement – ‘rather as, at Somerville, it had seemed rather amusing that the
brilliant philosopher Iris Murdoch should be “trying” to write a novel,’ Prudence Smith wrote in a lively memoir finished just before her death in 1999.
That evening, for what seemed to her a very long time, she, her architect husband and Lois Ventris sat chatting on the Breuer furniture in the main room at 47 Highpoint, getting a little hungry and drunk on sherry while waiting for Michael, with Lois apologizing somewhat anxiously for his absence at fairly frequent intervals. He was in the study, she said, and would come out as soon as he could. Eventually, he burst into the room, his normally neat hair ruffled, ‘full of apologies but even more full of excitement’. ‘I know it, I know it. I am certain of it,’ he told them.
After dinner, Michael took Prudence into the study. During the previous year, he had shown her some of the Work Notes as he had written them, and so she was able, to some extent at least, to follow the complicated business of vowel frequencies, syllabic spellings, inflectional endings and cross-references to other ancient languages in other ancient scripts, and the phonetic values set out on the various grids – which she nevertheless found ‘as challenging, in their own way, as the Minoan labyrinth to its victims’. But she hung on listening to him, for even though she knew little about the Ventris methodology, she knew enough Greek not to doubt that if his complex positionings and suppositions were valid, then the revealed language was indeed a form of Greek. ‘[His] work still had a long way to go, but the road he had travelled entirely persuaded me, on that strange evening, of his achievement.’
Within days of the ruined dinner party, Ventris felt confident enough to write to Myres and Bennett – the two leading scholars in Linear B studies. ‘Dear Sir John.… During the last couple of days I have been carrying on with the fantasy I discussed in my last Note [Work Note 20]; and though it runs completely counter to everything I’ve said in the past, I’m now almost completely convinced that the Pylos tablets are in GREEK. It’s a pity there’s not a new language to study, but it looks as if we must go to Linear A for that.’ Then he listed a series of words and phrases with his proposed Greek transliterations and English translations, glossing some of them with explanations, and finished up: ‘It may still be a hallucination, and you may well say that the Knossos forms just don’t fit. But the thing that staggers me is that whenever I go to the Greek dictionary to check a word I seem to have found but which is unfamiliar to me, it generally seems to exist and to make sense.’ To Bennett, the list of words he sent was longer and the explanation extremely terse: ‘I have, I think, great news for you. You must judge for yourself, but I think I’ve deciphered Linear B, and that Knossos and Pylos are both in Greek.’ Apart from a lengthy list of tablet numbers with his transliterations of their inscriptions in Greek, the only other evidence in the letter was a grid dated 18 June 1952, with 10 consonants and 5 vowels labelled with their phonetic values and more than 40 Linear B signs placed on it (see below).
At the same time, Prudence Smith, conscious of having an intellectual scoop on her hands, was persuading her colleagues at the BBC’s highbrow Third Programme that Ventris must give a talk on his discovery. ‘No, I had to tell them, he did not work in a university, or a museum, he happened to be a friend…a young architect. They were sceptical – rightly so, for I too was very young; but they gave in, I suppose, to my fervour.’ She also had to persuade Ventris himself. For if he was wrong, he would not only make a fool of himself (like the Czech scholar Hrozný who had claimed to decipher several undeciphered scripts, including Linear B, in the 1940s), he would also discredit his methodology. But since his whole purpose in circulating the Work Notes to other scholars had been to open up discussion on the Minoan scripts, and since he knew that a scholarly exposition of his Greek theory in a journal might easily take a year or two to develop and publish (if it was accepted at all), Ventris too agreed to the talk.
Ventris’s Mycenaean syllabary, as sent to Emmett Bennett Jr in June 1952, with Bennett’s pencilled emendations.
(From the archives of PASP, Program in Aegean Scripts and Prehistory, University of Texas at Austin, courtesy Tom Palaima)
‘Deciphering Europe’s earliest scripts’ was broadcast on the BBC on 1 July – astonishingly soon after Ventris’s breakthrough. It is the only recording we have of his voice: urbane, clear, unemotional and precise, as one might expect, but also a curious combination of firmness and diffidence, reflecting the brilliant but still unproven nature of his discovery, and probably betraying something deeper too. In the most quoted passage, Ventris declared: ‘For a long time I, too, thought that Etruscan might afford the clue we were looking for, but during the last few weeks, I have come to the conclusion that the Knossos and Pylos tablets must, after all, be written in Greek – a difficult and archaic Greek, seeing that it is 500 years older than Homer and written in a rather abbreviated form, but Greek nevertheless.’ And this in turn meant, Ventris concluded the broadcast, that the language of the Linear B tablets from Knossos, like those from Pylos on mainland Greece, should no longer be referred to as ‘Minoan’ but as ‘Mycenaean Greek’ – that is, as the language not of Evans’s Minoans but of the mainland Greek civilization based at Mycenae that had preceded the civilization of the classical Greeks.
The talk was a masterly exposition of complex, rarefied material for a non-specialist audience – and also, to quote its producer Prudence Smith, ‘an essay in modesty’ which even managed to preserve the speaker’s ‘rather surprised and grateful air of…a revelation which had somehow or other happened to him.’ Old Myres wrote encouragingly afterwards to Ventris who replied that the ending had to be rewritten four times, ‘starting from a decidedly Etruscan bias!’ The single serious failing (which hardly anyone listening would have noticed) was that he omitted to mention Alice Kober – the scholar whose work had influenced Ventris the most – while he gave credit to many others, including of course Evans, Myres and Bennett. In his moment of triumph, even Michael Ventris was not above the innate human tendency to downplay the contribution of a rival, though in all his subsequent writings he seems to have realized his lack of generosity and given Kober her due.
There can be hardly any doubt that Kober, had she still been living in July 1952, would have treated Ventris’s decipherment with scathing scepticism. Myres, though he may have been positive about the BBC talk, did not endorse Ventris’s chief conclusion, neither did Bennett. Both scholars would take many months to come round to it; a solid consensus in favour of the decipherment would require two or three years to develop; and a few serious scholars would never accept it and would even revile it (as we shall see). Something similar happened with Champollion’s decipherment of Egyptian hieroglyphs in 1823, which was not fully accepted until the 1860s; and again, in our own time, with the decipherment of Mayan glyphs started in the early 1950s, which was the subject of acrimonious debate for two or three decades and is yet to make it into some reference books.
To persuade the ‘experts’, three fundamental obstacles had somehow to be overcome by Ventris in July 1952, as he moved into the ‘substitution’ phase of the decipherment. First, he had to explain how he had arrived in a logical way at his phonetic values for the Linear B signs. Second, he had to show that ‘Mycenaean’ Greek related to classical Greek in a way that was both internally consistent throughout the Pylos and Knossos tablets and also consistent with the reconstructions of earlier, simpler forms of Greek made by classical philologists according to linguistic ‘laws’ governing sound changes over time and place. Third, he had to show how the many awkward words he had transliterated into forms which were not found in Greek dictionaries, might be interpreted plausibly as Greek.
On the first question, we already know from the previous chapter that Ventris was only partly successful. He could never completely justify his thought processes during the analysis phase of the decipherment, precisely because some of them were not logical but intuitive. Those who supported the decipherment were inclined to overlook this, and to argue that it was not the methods that mattere
d but the end result: did the phonetic values produce recognizable Greek words? If so, never mind how the values were arrived at – to put the argument baldly. But in dealing with the second and third questions, Ventris would be substantially successful over the next months and years, having taken the help of specialist scholars.
The first thing he did was to draw up a long vocabulary for Mycenaean Greek, showing all the words he had been able to interpret, including proper names like Amnisos, Knossos and Pylos. Overleaf is the first page (which does not show any proper names).
The first page of Ventris’s Mycenaean vocabulary list, July 1952.
Let us have a look at a mere five words in his list:
a. Number 7, pa-te in Mycenaean Greek, is equivalent to ‘patēr’ (father) in classical Greek. Final ‘-r’ is omitted in Mycenaean Greek; and the length of vowels is not noted (contrast classical Greek which has both epsilon, ɛ, for ‘e’, and eta, η, for ‘ē’).
The Man Who Deciphered Linear B Page 9