An English architect cracked Linear B by treating it like a crossword

In May 1952, in a modest London office, Michael Ventris, a thirty-year-old architect with an unrelenting curiosity for ancient scripts, composed the twentieth in a series of unofficial 'Work Notes'. These notes were circulated among a select group of about twenty scholars engrossed in deciphering the Mycenaean writing system known as Linear B. Work Note 20 was unlike its predecessors. Within it lay a grid—a matrix mapping the recurring patterns of signs from Linear B inscriptions. Ventris, who had been meticulously studying these patterns for six years, described his approach as 'cryptographic', eschewing any preconceived notions about the language encoded by the script. Through trial and error, he assigned particular sound values to signs, noting that the most frequent contexts were represented by open syllables, ending in vowels, while specific columns and rows shared consonants and vowels, respectively. The breakthrough came when familiar names emerged: 'Po-se-da-o-ne' as Poseidon, and 'A-mi-ni-so' as Amnisos, a port near Knossos. The revelation was monumental—Linear B was Greek.

What Linear B was

Linear B, a syllabic script where each sign represents a syllable, typically consonant-plus-vowel, was employed by the Mycenaean palaces of mainland Greece and Crete from approximately 1450 to 1200 BCE. The script owes its preservation to a fortuitous accident: clay tablets used for palace administration were inadvertently baked when the palaces succumbed to fires during the Late Bronze Age catastrophes. Sir Arthur Evans, who unearthed Knossos on Crete beginning in 1900, unearthed thousands of such tablets, naming the scripts Linear A and Linear B. Evans held the belief that Linear B was a Minoan script and not Greek, a view that influenced scholarly interpretations for decades. Until his death in 1941, Evans and others worked under the assumption that the language behind Linear B was a Minoan linguistic isolate. This assumption contributed to their inability to decipher the script.

Ventris's background

Michael Ventris was born in 1922 in Wheathampstead, Hertfordshire, into a family that moved frequently due to his father’s illness. This transient upbringing did not deter his prodigious linguistic abilities. By the age of six, he had delved into a German book on Egyptian hieroglyphs, facilitated by the multilingual environment of his home—his mother was Polish-German. At fourteen, Ventris attended a lecture by Sir Arthur Evans at Burlington House, where he first encountered Linear B. The encounter was decisive; he resolved to decipher the script. His education at Stowe School was followed by an unenthusiastic stint at the Architectural Association School in London. Ventris served as a navigator in the RAF during World War II, applying his fluency in German to intelligence work. Post-war, his architectural career never eclipsed his passion for Linear B. He cultivated a network of correspondents among academic scholars, despite lacking formal qualifications in classics or linguistics. His expertise lay in cryptography, his visual acuity for pattern recognition, and his ability to engage with the scholarly community on their own terms.

The grid method

Ventris’s primary innovation was the creation of what he termed the 'grid'. By compiling the corpus of Linear B inscriptions, which by the mid-20th century included about 3,000 tablets from Knossos and later, thanks to Carl Blegen's 1939 excavations, additional tablets from Pylos, Ventris analysed every recurring sign pattern. He examined the positions of signs, their alternation in inflectional endings, and their co-occurrence within the texts. This analysis was purely formal, devoid of assumptions about the language underlying the script. Through this method, he grouped signs into a structural grid where columns shared consonantal features and rows shared vocalic ones. Although this grid initially lacked specific sound assignments, it represented a phonological framework of the script. Ventris initially hypothesized that the language of Linear B might relate to non-Greek tongues, perhaps Etruscan. The breakthrough came when he assigned consonant and vowel values to the grid, leading to meaningful transliterations.

Why Greek

The definitive breakthrough came with a set of tablets from Pylos, which contained potential place names grouped with numeral counts. Ventris, employing his trial values, transliterated one frequent token as 'ko-no-so', corresponding to Knossos. Other tokens yielded 'a-mi-ni-so' for Amnisos and 'tu-ri-so' for Tylissos, all recognizable as Cretan localities. This critical finding pointed decisively towards Greek being the language of the tablets. The language was an archaic form of Greek, predating the Homeric dialect by centuries. It featured case endings and verb forms anticipated by Indo-European linguists in their reconstructions of proto-Greek. The transliterated texts revealed inventories, lists of workers, ration records, and religious offerings to gods like Zeus and Poseidon, confirming the administrative nature of the language used in the Late Bronze Age Greek world.

Chadwick and the publication

On 1 July 1952, Ventris announced his decipherment in a BBC radio broadcast. This broadcast caught the attention of John Chadwick, a classical philologist at Cambridge, who reached out to offer his expertise in early Greek dialects. Their collaboration resulted in the 1953 publication, 'Evidence for Greek Dialect in the Mycenaean Archives' in the Journal of Hellenic Studies, which built scholarly momentum for the decipherment’s acceptance. The pair expanded on this work, culminating in the 1956 publication, 'Documents in Mycenaean Greek', which provided comprehensive evidence for the decipherment. Although initial reception was cautious due to the entrenched belief in Linear B's Minoan identity, by 1954, Ventris and Chadwick's findings had secured widespread acceptance. Linear B was definitively recognised as a record of the Mycenaean Greek administrative system.

What the tablets say

The Linear B tablets offer a glimpse into the bureaucratic machinery of Mycenaean palaces. These records are devoid of literary content; they focus strictly on administrative data. They enumerate inventories of wool, oil, perfume, and grain, describe numbers of craftsmen like bronze workers and weavers, detail agricultural allocations, record ration distributions, and document tributes from various towns. The tablets also list offerings to gods—Zeus, Hera, Poseidon, among others—highlighting the presence of familiar deities from the later Greek pantheon. Some personal names echo those found in the Homeric epics, such as Achilleus and Theseus. Despite the absence of narrative texts, the language of the tablets enriches our understanding of the Mycenaean world, anchoring it firmly within the continuum of Greek linguistic history.

The legacy of Michael Ventris's decipherment is profound, yet his own life was tragically brief. On 6 September 1956, Ventris died in a car crash on the Great North Road at the age of thirty-four. He had been experiencing a prolonged period of depression, and while his death has been interpreted by some as a possible suicide, no official determination was made. Ventris lived to see his work validated and published, notably in 'Documents in Mycenaean Greek', two months prior to his death. Afterward, John Chadwick carried the torch of Mycenaean studies, contributing for another five decades. Meanwhile, Linear A—the script's earlier cousin—remains undeciphered. The techniques Ventris honed for Linear B do not easily apply to Linear A, as its underlying language is likely neither Greek nor related to any known language family. The grid waits for another mind capable of unravelling its mysteries, reminiscent of Ventris's diligent, pattern-seeking approach.