The Riddle of the Labyrinth (9 page)

Determinatives can be found, to a greater or lesser degree, in other writing systems. As Ventris's biographer Andrew Robinson points out in his book
The Story of Writing
, when capital letters are used to signal proper names in English and other languages, they are functioning as determinatives of a sort. Whether you were conscious of it or not, capitals most likely helped you pick out the proper names in the Classical Greek “tablet” earlier in this chapter about sending gold to Homer.

And it was the presence of determinatives in Linear B—or something that looked a great deal like them—that would bring Arthur Evans a world of trouble as he began to sift the strange Bronze Age symbols.

THE SCRIBES AT KNOSSOS never imagined they were writing for posterity. They were simply keeping the records of their own community, in their own language, as any chroniclers would do. But with the passage of millennia, those records had become, quite literally, cryptic.

A decipherer approaches an ancient script much as a cryptanalyst does a secret code. In some respects, the decipherer's task is easier: Unlike codes, real-life writing systems are rarely meant to conceal or deceive. In other ways his task is harder—often much harder. Unlike the cryptanalyst, the decipherer may not know what language is being encoded. When you do the cryptogram in your Sunday paper, you are secure from the start that the solution will be in English. The decipherer of a forgotten script may have no comparable assurance.

But even when the language of a script is unknown, the script itself is filled with internal clues, if only one knows where to look. Confronted with a piece of unknown writing, the decipherer must begin by subjecting it to a series of diagnostic tests. Each test is a small forensic exercise designed to coax the script, bit by bit, to yield up its identity.

The first test is so obvious it is sometimes overlooked. The decipherer has to establish at the start whether the tangle of symbols before him actually is writing. This is not always as straightforward as it seems. Suppose you unearth another tablet, one that looks like this:

Does this tablet, with its rows of primitive symbols, contain writing? No: It is a Modernist painting,
Pastorale (Rhythms)
, done in 1927 by the artist Paul Klee.

What about these whimsical sculptures? Could they conceivably be writing?

Yes: The photograph shows a fragment of an ancient Mayan tablet. The carvings are Mayan hieroglyphics, a writing system deciphered only in the late twentieth century.

With some symbol systems, the answer is far less clear. Scholars have practically come to blows over Rongorongo, the hieroglyphic script of Easter Island. Carved on wooden tablets, Rongorongo contains hundreds of logograms. In their vibrant stylization, they evoke the stick figures of Sir Arthur Conan Doyle's “Dancing Men” cipher as reimagined by the pop artist Keith Haring:

Though Rongorongo has resisted decipherment for well over a century, some modern scholars insist it is a true writing system. Others call it a kind of proto-writing, used as a memory aid for traditional rituals performed aloud. To still others, it is visual art and nothing more.

Fortunately for Arthur Evans, Linear B passed this first test handily. Stored in archival boxes and carefully labeled, the tablets were unmistakably documentary records. It was plain to him from the moment he unearthed them that the symbols they contained were writing, meant to be read.

Once the decipherer is sure he has writing, he must establish precisely what kind of writing system he has. In principle, this can be done with a very simple test: Count the number of different characters the writing system contains, and the total will tell you what kind of system it is. If the number is very large—say, in the thousands—you are looking at a logographic system, like Chinese, in which each word of the language is written with a separate symbol. If the number is between about 80 and 200, then you have a syllabic system. The Cherokee syllabary, invented by Chief Sequoyah in 1819 and containing 85 characters, is just such a system:

If the character count is only in the dozens, then you are looking at an alphabet. The alphabets of the world range in size from about a dozen letters, used by the Rotokas alphabet of the Solomon Islands, through the thirty-three Cyrillic letters used in Russian, to the more than seventy characters of the Khmer alphabet of Cambodia.

With a known script, counting characters is easy. With an unknown script, it can be a nightmare that drags on for decades. Consider the following scenario: One day, an alien lands in Times Square. He is a very bright alien, but he knows no Earth language, nor has he encountered writing of any kind. Assaulted by a blizzard of text from the billboards and newsstands and neon signs around him, he tries to make sense of the Roman alphabet by scrutinizing all its available forms. He sees it on video screens and the printed page, on vertical signs and horizontal ones, in an array of fonts, colors, and sizes. It will take our alien years of minute comparison before he can be positive, say, that
and
are mere variations of the same letter, while
and
are all different letters entirely. And so on, for every symbol he sees. There is difference, and there is meaningful difference. You cannot count what you cannot distinguish.

The decipherers of Linear B found themselves in similar straits. As David Kahn wrote evocatively in
The Codebreakers
:

After decades of digging at Knossos, Evans had thousands of tablets at his disposal. Their combined texts ran to tens of thousands of signs. He noticed immediately that many signs looked similar though not quite identical, as in these pairs: