The Riddle of the Labyrinth (20 page)

“Dear Professor Myres,” she wrote on November 20, 1946:

In closing, she wrote, with characteristic quiet ardor, “I must confess, strong as the statement sounds, that I would gladly go to the ends of the earth, if there were a chance of seeing a new Linear Class B inscription when I got there.”

She signed the letter, “Apologetically, yet hopefully, (Miss) Alice E. Kober.” Then she settled down with her cigarettes and her slide rule to wait for an answer.

BY THEN, KOBER—working with only two hundred inscriptions—had already taken a significant step forward. The thousands of hours spent sifting her cards had begun to pay dividends, and, piece by piece, the picture on the puzzle box was starting to emerge. In 1952, in his triumphant announcement of the decipherment on BBC Radio, Michael Ventris discussed the steps by which a decipherer makes his way through the forest of symbols:

What Ventris did not say was that it was Alice Kober, at her dining room table, who had first done those very things.

In 1945, Kober published what would be the first of her three major articles on Linear B. Her great contribution to the decipherment—elegantly on display in all three papers—lay not only in
what
she found but also in the means
by which
she found it. Several things about her method stood out:

To begin with, she assumed almost nothing, disregarding nearly every idea about the script that had been put forward by others. (Practically the only thing Kober did not doubt was that Linear B was a syllabary. “If all the Minoan scripts aren't syllabic, I'll eat them,” she wrote privately. “That's the one thing I'm sure of.”) Second, she refused to speculate on the language of the tablets. Third—and this was the most radical departure of all—she refused to assign a single sound-value to the Linear B characters.

One of the greatest temptations an archaeological decipherer faces is to assign a phonetic value to every character at the start—to speculate, that is, on which sounds of the language the various characters stand for. With an unknown language in an unknown script, there is little way to do this besides random guesswork. The great danger of doing so, as Kober repeatedly warned, is that it sets up a cycle of circular reasoning: The characters of a script correspond to the sounds of Language X; therefore, the script writes Language X.

To Kober, assigning sound-values at the outset was the refuge of the careless, the amateurish, and the downright deluded. By contrast, she treated the symbols of Linear B as objects of pure form, looking for patterns that might lead her, all by themselves, into the structure of the Minoan tongue. In the Blissymbolics problem above, we began with form
and
meaning: the curious symbols and the set of English translations. In Linear B, investigators faced a far more brutal situation: They were forced to inhabit, as Kober evocatively wrote, a world of “form without meaning.” Of all the would-be decipherers, she was the one most willing to dwell there for as long as it took.

Methods like these inform all of Kober's writing about Linear B. They come to full flower in her three important papers on the script, published in the
American Journal of Archaeology
between 1945 and 1948. In her 1945 paper, “Evidence of Inflection in the ‘Chariot' Tablets from Knossos,” she examined a series of tablets, reproduced by Evans in
The Palace of Minos
, that deal with chariots and their constituent parts. (The “Chariot” tablets are noteworthy among the Knossos tablets for containing complete sentences of the Minoan language rather than mere word lists.) The tablets in question, as hand-copied by Kober from Evans's book, included these:

Kober's paper was significant for two things, both of which would have a profound effect on the course of the decipherment. First, it showed that it was possible to analyze the Cretan symbols without attaching sound-values to them. Second, it proved that the language of Linear B was
inflected
—that is, that it relied on word endings, much as Latin or German or Spanish does, to give its sentences grammar. Though on its face this discovery may look like a small find, it would eventually furnish the long-sought “way in” to the tightly closed system of Linear B.

Roughly speaking, inflections are word endings, also known as suffixes. Languages make use of inflection to varying degrees: In English, inflections include
-s
(which indicates a third-person singular verb);
-ed
(past tense);
-ing
(participle);
-s
(again, this time indicating a plural noun); and a few others. Modern English has only about eight inflections. Other languages, like Chinese, have almost none.

Some languages, however, use far more, as anyone who has sweated over Latin, with its constellations of word endings, can attest. In Latin, verbs can be inflected for a spate of grammatical attributes, including person (first, second, or third person) and number (singular or plural), depending on who is performing the action of the verb. Tiny but powerful, these inflections impart the kind of information that in English usually requires extra words. Consider the Latin verb
laudāre
, “to praise.” When any of the following suffixes is attached to its stem (
laud-)
, the result is an entire small sentence, elegantly bound up in a single inflected word:

The nouns of Latin are also richly inflected. In the table below, inflections mark the noun's “case,” telling us what role it plays in a given sentence: subject, direct object, indirect object, possessor, or addressee:

For the analyst of an unknown script, demonstrating that the language of the script is inflected is a major diagnostic triumph, for it narrows the field of likely candidates in a single stroke: One can immediately eliminate inflection-poor languages like Chinese and focus the search on inflection-rich ones like Latin. But there is a catch: Conclusive proof of inflection is extremely hard to come by—much harder than it seems.

When Evans first studied the Knossos tablets, he noticed small sequences of symbols that seemed to recur at the ends of words. Among them were these:

He decided that these sequences were inflections on the ends of Minoan nouns. (He thought they indicated gender, or perhaps case, much as suffixes on German and Latin nouns do.) But inflections are wily things, and it is all too easy for an investigator to spot them where they don't actually exist. Consider the following four English words, written in the Dancing Men font to make them more cipherlike:

A decipherer will immediately spot the three-character sequence
at the end of each word. It certainly looks like a suffix, perhaps
-ing
. “Eureka! Inflection!” he cries. But is it really?