The Riddle of the Labyrinth (39 page)

The discovery was Ventris's first meaningful advance in assigning sound-values to Linear B characters. It gave him grist for his grids, of which he would eventually produce two more. One, in Work Note 15, from September 1951, plotted fifty-one signs. (In this grid, Ventris prudently refrained from assigning sound-values.) The other, in Work Note 17, from February 1952, also plotted fifty-one signs and again posited sound-values. These values, Ventris wrote, “represent the values which seem the most useful in giving an ‘Estrusciod' character to Pylos names, words and inflexions.”

Histories of the decipherment have focused extensively on Ventris's grids, but what is truly significant is that they helped speed the decipherment
despite the fact that they were only partly correct
. Even in his third grid, Pope observes, Ventris has assigned wrong values to 25 percent of the signs. But he was nonetheless able to crack the code—an even more extraordinary achievement in light of the imperfect road maps he had constructed.

How, then, did Ventris manage to do it? The answer lies in his making a single intuitive leap, which would provide the catalyst for the whole decipherment. It was a deeper version of the kind of leap needed to equate the symbol
with the English word
merry
in the Blissymbolics problem we examined earlier.

Ventris's leap centered on the sets of related nouns he called “Kober's triplets.” When he first read her articles in 1948, he wasn't persuaded that the triplets truly made the case for inflection. “I must say I was slightly disappointed,” he wrote Myres then. “I envy the orderly presentation, but I don't feel it results in conclusive proof that the ‘terminations' are actually inflexional.” Ventris thought that the triplets more likely represented “alternative name-endings,” on the order of
Brooklyn
/
Brooklynite
/
Brooklynese
, and so forth.

Little by little, Ventris came around to Kober's point of view, and before long he was combing the Pylos data for additional triplets and “near-pairs”—words or phrases, including
that differed in only one or two final signs. By the summer of 1951, he wrote, he had collected 160 such groups, “which on the face of it show inflexion.” These related forms helped Ventris isolate still more “bridging” characters of the kind Kober had described, and to add the corresponding sound-relationships to his grid. Ventris at last seemed to realize that Kober had handed him the key—the critical intersection of inflected language with syllabic script, and the “sharing” of consonants and vowels that results. Now he had only to figure out which door the key unlocked.

Before he did so, he made time for a trip to Crete and the Greek mainland in the fall of 1951. In Istanbul for a conference on the ancient Near East, he traveled on to Crete, where he stayed at the Villa Ariadne and saw Knossos. “I was frankly rather disappointed in the Palace & in the frescoes in the Museum which seemed to me to have been restored in such a way as to kill most of their charm & atmosphere,” he wrote to Myres. “Mycenae, on the other hand, where there's little else but rubble, the atmosphere is stupendous.”

Returning to London, Ventris began to take hesitant steps toward his second major advance. Now that he'd been through the Pylos data, he realized something vital: While it contained many “triplets” of the
kind
Kober had identified, the
specific
triplets she had isolated were found only at Knossos. The words unique to Knossos included a set she had published in her 1946 paper on declension:

What sort of words, Ventris wondered, might be particular to one place but not another? For the answer, one has only to look to any modern municipal document. The municipal documents in my life come from Manhattan. The diverse official papers I've accumulated here over the years, from marriage certificate to local tax forms to a string of jury summonses, all have one thing in common: Somewhere on each of them, prominently featured, appear the words “City of New York.” The Cretan tablets were the administrative documents of the palace at Knossos. Perhaps, Ventris conjectured in early 1952, the words found exclusively there were the names of Cretan towns. With that in mind, he tried the following experiment:

Ventris homed in on the simplest forms in Kober's paradigm, the three words—
,
, and
—in the bottom row. Examining
, he already assumed the first character was the pure vowel “a.” By making “only a little adjustment” to his grid, as he told Myres, he was able to plug in reasonable guesses for the values of other signs, much as Champollion had done in the famous “Ramses” cartouche. What Ventris wound up with looked an awful lot like the Greek names, spelled syllabically, for three major towns of Cretan antiquity: Amnisos, Tulissos, and Knossos.

That didn't mean, of course, that the language of the tablets was Greek: Though all three names were attested in Classical Greek writings, they were known to be of pre-Greek origin. Nevertheless, even the barest hint of Greek raised a distasteful enough specter for Ventris to reject his results out of hand. He still believed firmly in an “Etruscan solution” and remained adamant, as he'd written as an eighteen-year-old in 1940, that “the theory that Minoan could be Greek is based . . . on a deliberate disregard for historical plausibility.” He dismissed the place-names as a mirage and cast the idea aside.

I
N THE SPRING OF 1952,
Scripta Minoa II
was published at last. (Though Myres had written Kober in 1948 that “all your help in the B volume will of course be acknowledged in the preface to it, very gratefully,” the published book gave scant indication of the full extent of her years of hard labor.) The volume, with its hundreds of inscriptions, gave Ventris still more data. In May, he tried his place-name experiment a second time, with one significant difference: He allowed himself to make use of a clue he had previously shunned—the only external clue in existence to the possible identities of some Linear B characters.

The clue had been available to investigators from the very beginning, but as they all knew, it was a risky one. It came in the form of an Iron Age writing system known as the Cypriot script. Also a syllabary, the script was used to record the indigenous language of Cyprus between about the seventh and the second centuries B.C. (The Cypriot syllabary was a descendant of the Cypro-Minoan, on which John Franklin Daniel had done important research in the early 1940s.) The
language
of the Cypriot script remained a mystery, but the sound-value of each character was known: After the Hellenization of Cyprus, the syllabary was retained for a time to write Greek, and the island teemed with Greek syllabic inscriptions on monuments and coins. Thanks to the discovery in 1869 of a bilingual inscription in Cypriot and Phoenician, the Cypriot syllabary was deciphered in the 1870s.