THE CODEBREAKERS

by DAVID KAHN

  Work Note 10 resumed the discussion of the enclitic “and.” Work Note 11 showed that two alternating phrases represented male and female genders of “servant.” Work Note 12 classified the sign-groups into what Ventris thought were personal names, names of institutions or places, names of trades and titles applied to men and women, and a general vocabulary. Work Notes 13 and 14 showed that men’s names were declined in at least six different declensions. The lack of a regular sign for a nominative ending in -s militated against Greek or any related Indo-European language. Work Notes 15 and 17 expanded the grid and proposed a few tentative phonetic assignments. By September 28, 1951, Ventris, in Athens, drafted a grid in his beautiful hand that inserted 50 signs in its 85 cells.

  During the winter of 1951-52 Ventris made advances in elucidating a variety of minor points. In February, 1952, appeared Scripta Minoa II, edited by the elderly Sir John Myres, Evans’ old colleague, and presenting at last the tablets that Evans had found half a century before. Work Note 19 of March 20 gave considerable space to possible parallels in Etruscan for the inflectional activity of a particular sign.

  Ventris then took up a puzzling spelling variation that had grown more evident with the publication of Scripta Minoa II. To improve the symmetry of a corner of the grid in the light of this feature, he returned to the value jo for a sign for which he had, in Work Note 9, summarily rejected that value. This would make the many men’s names end (in the genitive) in -jo or -jojo (the j is a semiconsonant, like y). Ventris found precedents for this in derivative Greek, Etruscan, and Lycian names. The grid automatically compelled every sign in that column to share the vowel -o. Ventris thought that the throne sign was i and—even though the grid did not require him to do so—gave the same vowel to the signs that he had placed in that column. This was bolstered by the near identity of the sign for Cypriote ti and a sign in the throne column. He stuck with a for the double ax and the signs of that column, and with the help of this made his third consonant assumption, n, because of the similarity of the Cypriote na and a sign in the -a column.

  With these in mind, he looked again at some of the very words that Miss Kober had used in her original analysis. He thought that certain of these might be names of places. Their longer forms had added on the symbols for jo and ja to form masculine and feminine adjectives, just as “France” expands to “français” and “française.” In the words’ short forms, all their vowels were known. In the first word, for example, the grid showed that all three vowels were the same; the jo assumption made them o. The consonants were not known, but here again the grid showed that, in the first name and the second, the last consonants were the same. Moreover, the second consonant of the first word was the same as the third consonant of the second—even though they were mated to different vowels. The partial decipherments, in which the unknown consonants are represented by the numbers of the row of the grid in which the Minoan sign stood, were:

  60-80-130 a-7i-8i-130

  Among the place-names likely to occur in tablets found at Knossos would be that of its harbor town, Amnisos, which would have to be spelled with an extra vowel between the m and the n to conform to the consonant-vowel nature of the script. As Ventris later wrote, “It did not require very great imagination to realize” that the second of the two could be A-mi-ni-so. If so, the imperatives of the grid demanded that the first name become 6o-no-so, which, again on the basis of an inserted vowel, could be Ko-no-so, or Knossos itself. This looked good. Perhaps the scribes dropped the final -s. This might explain the puzzling failure of the sign that so resembled the Cypriote se to appear at the end of words as often as it would have in Greek. Though the inserted vowels differed in the two cases, they both followed the rule of anticipation: they were the same as the vowel of the following syllable.

  By a chain reaction, the grid now determined part of the sounds of dozens of other signs and the entire phonetic value of several others, just as it had indicated the consonant of no in Ko-no-so. For example, the sign like a tall beer glass stood at the intersection of the row for k and the column for i. It had to be ki. That was the beauty of the grid system. It forced its decipherment out of itself.

  Ventris looked at the two-sign word that Miss Kober had determined stood for “total.” The first sign stood in the same row as ti and in the o column; it had to be to. The second sign in the masculine form was so, and the second sign in the feminine form stood in the same row but in the a column: sa. Thus the two words were to-so and to-sa. They strongly resembled archaic forms of the Greek tossos and tossa, “so much,” or tossoi and tossai, “so many.” Greek? Everybody, including Ventris, thought that the Linear B language had to resemble some Aegean tongue that reflected the cultural domination of the Cretans, whose ethnic origin was widely regarded as non-Greek. The occurrence of an isolated word in a different language would not shake this view. It would not mean that Linear B was Greek any more than the presence of “habeas corpus” in a Supreme Court decision would mean that it was written in Latin.

  Borrowed words usually indicate a need in the borrowing language. Words for common, everyday things, on the other hand, are usually filled from the native stock. Consequently Ventris may have been a bit surprised when the first syllable of the word that, in masculine and feminine forms, had been identified by ideograms as “boys” and “girls” deciphered as ko, the beginning of the classical Greek “kouros” (“boy”) and “korē” (“girl”). Even this would not be conclusive: the English words for such homely concepts as “uncle” and “sky” have been imposed by invaders and do not come from Anglo-Saxon.

  Then Ventris recalled that philologists had reconstructed the primitive Greek forms of “koros” and “kore” as “korwos” and “korwa.” He thought that these primitive forms might be rendered in Linear B as ko-wo and ko-wa, and he drew from the back of his mind the suspicion that the barred A sign represented the pure vowel e. This would make -e-wo part of a common declension, and he remembered another reconstruction: -ewos, the primitive Greek genitive of the many words ending in -eus, such as the names “Odysseus,” “Peleus,” “Idomeneus.” He made further assumptions—perhaps just in the spirit of seeing where they would lead. These produced a whole phrase of what appeared to be mutilated archaic Greek from a table with a chariot ideogram; it translated as “fitted with reins.”

  Greek had now thrust itself upon him in vocabulary, syntax, and meaning. Always the assumptions, rigidly controlled by the grid, mutually interlocked. Could it be that the Linear B tablets were—contrary to every tenet of orthodox Bronze Age archeology—written in Greek?

  Ventris was not convinced. In Work Note 20 of June 1, 1952, which set out these results, he wrote: “If pursued, I suspect that this line of decipherment would sooner or later come to an impasse, or dissipate itself in absurdities.” He scrupulously pointed out that the button sign would not fit the archaic Greek word that scholars had reconstructed for an enclitic “and,” which was te. He called the Work Note “a frivolous digression,” and regarded the appearance of the Hellenic language as “the Greek chimera.”

  But while the Work Note still was in the mail, Ventris discovered that the chimera was astonishingly real. He had pursued the line of decipherment and found that the Greek solution could not be denied. His logic had conquered his preconceptions. He had recovered archaic forms of four well-known Greek words (for “shepherd,” “potter,” “goldsmith,” and “bronzesmith”) and translated eight phrases. On a B.B.C. broadcast over which he had previously been invited to give a talk on the general problem of the scripts, he said, “Once I made this assumption [that the tablets were written in Greek], most of the peculiarities of the language and spelling which had puzzled me seemed to find a logical explanation.” In June of 1952, Ventris felt that he had deciphered Linear B. Work Note 20 was the last.

  Michael Ventris’ grid of Linear B signs

  One of the most interested listeners to the broadcast was a young Cambridge philologist specializing in Greek, John
Chadwick. At the time, the Ventris theory was just the latest in a long line of supposed “solutions,” every one of which had failed. But Chadwick, who had himself failed to read the tablets on the assumption that they were Greek, was interested. He obtained Ventris’ Work Notes from Sir John Myres and went home to test the solution for himself. Within four days he was convinced. He had deciphered 23 plausible Greek words, some of which had not then been read by Ventris. On July 9, Chadwick wrote to the architect, congratulating him on the solution. They formed a close association and together wrote a report of the decipherment under a title which they had carefully chosen to avoid extravagant claims: “Evidence for Greek Dialect in the Mycenaean Archives.”

  It gave a decidedly confusing explanation of the decipherment. And it did not gather up all the loose ends. Some signs remained unknown; some translation difficulties arose. Yet its conclusions were cogent. In the first place, the deciphered words made sense. The language was Greek, truncated and primitive as compared to the polished classical tongue, but Greek. Its roughness could be attributed to the fact that the language of the tablets was a thousand years older than that of Plato, “a difference in date,” they noted, “as great as between Beowulf and Shakespeare.” Besides, many of the archaic forms agreed with predicted ones. In the second place, the deciphered texts reflected what the tablets appeared to be talking about. Where the grid produced the word for “sword,” a pictogram of a sword stood nearby. The two young authors submitted their article to the editors of the Journal of Hellenic Studies, who, recognizing its importance, made room for it in the 1953 volume despite the backlog of World War II articles that was still crowding their pages.

  The “tripod” tablet that clinched the decipherment of Linear B

  While waiting for it to appear, Ventris heard favorably from many experts who knew of his work. Professor Sittig, for example, who had been working on his own method of decipherment, now abandoned it and wrote to Ventris on May 22, 1953: “Your demonstrations are cryptologically the most interesting I have yet heard of, and are really fascinating.” Of course, not everybody climbed on the bandwagon. But in that same month Ventris received a letter from Blegen that settled the matter:

  Since my return to Greece I have spent much of my time working on the tablets from Pylos, getting them properly ready to be photographed. I have tried your experimental syllabary on some of them.

  Enclosed for your information is a copy of P641, which you may find interesting. It evidently deals with pots, some on three legs, some with four handles, some with three, and others without handles. The first word by your system seems to be ti-ri-po-de and it recurs twice as ti-ri-po (singular?). The four-handled pot is preceded by qe-to-ro-we, the three-handled by ti-ri-o-we or ti-ri-jo-we, the handleless pot by a-no-we. All this seems too good to be true. Is coincidence excluded?

  It was. The obvious relation of ti-ri- to tri-, of a- to the prefix meaning “nothing” or “without,” of -po-de to Greek root -pod- meaning “foot,” and of -o-we to the Greek -oues for “ear” or “handle” could not be denied. The language of Linear B was indubitably Greek. The tripod tablet results brought immediate agreement in principle by the vast majority of scholars competent in the field—including many who had themselves been defeated by the solution, such as Georgiev, Sundwall, and Bennett.

  But some disagreed. The same journal that published “Evidence” provided space for a rebuttal by A. J. Beattie, professor of Greek at the University of Edinburgh. He did not understand the grid—a misapprehension that was due not to any obtuseness of his but to the obscurity of the Ventris-Chadwick account. “Let us suppose that he [Ventris] used all the texts available to him,” Beattie wrote, “and that he counted every single sign in initial, medial and final positions, and so obtained three figures and an overall total for each sign, as well as an assortment of information about alternatives or concomitants. Are we then to suppose that these figures fell naturally into groups, so that the signs to which they referred could be disposed lengthwise and crosswise in such a way that they would ultimately be found to correspond to series of the type, i, pi, ti, ki, etc. and pa, pe, pi, etc.? This is evidently what Mr. Ventris means us to believe.” This same lack of understanding misinformed Beattie’s more valid linguistic criticisms: “What Mr. Ventris has given us by his transcription is not in fact the Greek language but a language of his own making. It is a strange language, which looks like Greek because he has been careful to provide it with a selection of Greek suffixes…. And by devising spelling-conventions of primitive simplicity, he has ensured that the syllables preceding the suffix of each word may occasionally be intelligible as Greek word-stems.” Even Beattie admitted that the tripod tablet yielded some Greek, however, and in the end was reduced to impugning, not the results, but the data itself: “We should in any case suspect the validity of a list that has no one-handled or two-handled pots but knows only those with three or four handles or none at all.”

  Unlike the attacks upon the HroznÝ and Georgiev “solutions” of Linear B, the critics’ objections failed to convince. The solution rapidly won acceptance, and classical circles began to use its results. The most important result was, of course, the very fact that the language was Greek. Greek was spoken at the seat of former Minoan power in Knossos because Greeks ruled there. This vindicated Wace’s rebel view that the mainland dominated Crete during the questioned years of 1400 to 1125 B.C. and thereby revised the Late Bronze Age history of the Aegean.

  But what did the tablets say?

  Sample texts read like this: “Koldos the shepherd holds a lease from the village: 48 litres of wheat.” “At Pylos: five sons of the Ti-nwa-sian weavers (sons of rowers at A-pu-ne-we), two boys.” “Four (or more) slaves of Koradol-los in charge of seed-corn.” “One pair of wheels bound with bronze, unfit for service.” The tripod tablet, with pictograms in italic capitals: “Two tripods: Aigeus the Cretan brings them: 2 TRIPODS. One tripod: it is not sound as regards one foot: 1 TRIPOD. One tripod: the Cretan brings it; charred around the legs…1 TRIPOD. Wine-jars: 3 JARS. One larger cup with 4 handles: 1 CUP. Two larger cups with 3 handles: 2 CUPS. One smaller cup with 4 handles: 1 CUP. One smaller cup with 3 handles: 1 CUP. One smaller cup with no handle: 1 CUP.” A votive tablet reporting, “To all the gods, one amphora of honey: 1 AMPHORA. To the Mistress of the Labyrinth, one amphora of honey: 1 AMPHORA” reads in the syllabic script pa-si-te-o-i me-ri AMPHORA 1/da-pu-ri-to-jo po-ti-ni-ja me-ri AMPHORA 1.”

  None of the tablets contains any literary work, nor any diplomatic instructions, personal letters, religious texts, historical writings, nor anything, in fact, beside these minutely detailed bureaucratic records of petty commercial transactions. Professor Denys Page describes the impression they make as a whole:

  These palace archives are the records of a comprehensive and pervasive bureaucracy, administering for hundreds of years a most elaborately organized society…. It is as if everything done by everybody was open to official inquiry and subject to official orders. We possess a part only of the archives for a single year at Pylos: they record thousands of transactions in hundreds of places…. But even more astonishing and significant is the omniscience, the insatiable thirst for intimate detail. Sheep may be counted up to a glittering total of twenty-five thousand; but there is still a purpose to be served by recording the fact that one animal was contributed by Komawens and another by E-te-wa-no. Restless officialdom notes the presence in Pe-se-ro’s house of one woman and two children; the employment of two nurses, one girl, and one boy, in a Cretan village; the fattening of an insignificant number of hogs in nine places; the existence somewhere of a single pair of brassbound chariot wheels and labelled “useless”…. One would suppose that not a seed could be sown, not a gram of bronze worked, not a cloth woven, not a goat reared or a hog fattened, without the filling of a form in the Royal Palace; such is the impression made by only part of the files for a single year.

  But was this all? Was this piddling minutiae to be the only result
of a brilliant achievement? Was there to be no Epic of Gilgamesh, no Code of Hammurabi, no pharaonic boasts of kings conquered and cities sacked—only the Bronze Age equivalent of some incomplete county clerk records, some transactions of a farm cooperative, and some small donations to a parish church? Only that. But although no poetry has been discovered, the information contained in the tablets has, by inference, illuminated some of the greatest of Western man’s poetry—the Iliad and the Odyssey.

  The tablets do so because they are 400 years older than the time of Homer and contemporary with the events of which he sings. Linear B represents the written form of the language spoken by the almost legendary figures of the Trojan War. The clay tablets found at Mycenae and Pylos were written at almost the very moment that Agamemnon and Nestor, the kings of those cities, were waging war on the Trojan plain. Soon thereafter, waves of impoverished and barbaric Dorians from the north, brandishing their iron weapons, overran the Bronze Age civilization of heroic Greece. They extinguished the light of learning and the love of art, and for four centuries Greece dwelled in the illiterate darkness of its early Iron Age. In the eighth century B.C. the Hellenes began to emerge from this eclipse. They started to write with the Phoenician alphabet—the precursor of the present Latin alphabet. At the same time, a blind genius molded the stories of gods and men that had been transmitted orally from the past into a great unified theme and won undying fame for his name—Homer. The names and noble deeds and language of his epics in turn helped fix the ideals of the classical Greek civilization that flowered a few centuries later, and so helped mold the West.