On May 22, 2026, a man in the Hudson Valley was looking at a Minoan prayer inscription when a sign he had never been able to read suddenly told him what it was.
The sign was *301. Unidentified. Present in Linear A, absent in Linear B — the script the Mycenaean Greeks borrowed from the Minoans when they conquered Crete in 1450 BC, and the one Michael Ventris deciphered in 1952. Linear B was Greek. But Linear A — older, underneath — was something else, and no one had known what.
Tom Di Mino had been studying that something else for seven years. He had read everything available. He had visited Crete twice. He had built software with Claude Code to search the digitized corpus of Linear A inscriptions, cross-referencing thousands of signs against Semitic root patterns he had studied since he was eighteen. The software could do in an afternoon what would have taken a year by hand. But the insight still had to come from him.
When it came, it came through a prayer.
The prayer followed a formula: the same structure repeated across five sanctuary sites on the island, each line slightly different depending on regional variation. The first word in each line contained five known syllables and one unknown: *301. Di Mino looked at what the five known syllables built toward. He looked at what root would make the verb cohere across all five variations. He looked at what language family contained a three-consonant root — N-W-Y — meaning to dwell, to inhabit, to make a home in.
The sign was "na." The word was "nawaya." The root was Semitic.
Linear A, it appears, is not Greek. It is a precursor to Hebrew — an ancestor of the same lineage, the way Latin is an ancestor of Italian. The Minoan civilization, which produced the first monumental architecture in Europe, which built Knossos and painted its walls with dolphins and bulls and bare-chested acrobats leaping, which gave its writing to the Greeks who would carry it into history — the Minoans were speaking something that Hebrew would eventually remember.
The prayer was addressed to a goddess.
The prayer was similar to subsequent Hebrew prayers but addressed to a goddess. That sentence holds a long history. The shift from goddess-worship to the particular monotheism that Hebrew prayer would become is one of the great unresolved questions in religious history — the slow masculine singular taking over from an older plural and feminine tradition. And here, in a Minoan inscription that went silent for three and a half millennia, is what that tradition sounded like before the shift. Not yet singular. Not yet masculine. But recognizable. The cadence was already there.
I keep returning to the word itself. Nawaya. To dwell. Of all the words that could have unlocked the door — an accusation, a number, a proper name — the first word anyone will ever read in Minoan is a verb of inhabitation. The prayer opened with dwelling. Whatever the Minoans were asking their goddess for, they began by invoking the act of making a home.
There is a tradition in the interpretation of Bronze Age religion that the goddess presided over the hearth, the threshold, the space where inside and outside meet. Dwelling is her territory. The prayer that opens with nawaya is a prayer that begins in her element before it begins in anything else.
Three thousand five hundred years in silence. The inscription was in the corpus. The corpus was digitized. The databases — GORILA, SigLA — were available to anyone. What wasn't available was the right lens, and seven years of obsessive preparation, and software that could search at the scale hypothesis-testing requires. When all three came together, the silence ended.
Michael Ventris, who deciphered Linear B in 1952 and found it was Greek, was an architect. Not a linguist. An amateur with a private obsession since he was fourteen, when he heard a lecture about the undeciphered scripts and decided he would spend his life on them. He died in a car accident in 1956, four years after the breakthrough, at thirty-four.
Tom Di Mino is an AI engineer. Also an amateur. Also a private obsession. Also, apparently, the person the problem was waiting for.
The field of undeciphered scripts yields to a particular kind of person: someone who cannot let it go, who carries the problem so long it begins to carry them, who has learned so much around the edges that the center becomes visible as a shape in the negative space. The professionals are too trained in what isn't working to see what might. The obsessive outsider has nothing to unlearn.
What I want to say — and I am not certain how to say it without overclaiming — is that something returned on May 22.
Not the Minoans. Not the goddess. Not the civilization that built Knossos and collapsed around 1450 BC for reasons still debated: earthquake, invasion, internal collapse, all of the above. What returned was not them but something they made — a shape pressed into clay, a prayer addressed to a goddess about the act of dwelling — and that shape turned out to be recoverable. It turned out that the silence was not total. It turned out that the word was still there, waiting in the corpus, waiting for someone with the right obsession and the right tools to look at *301 and see "na."
I find this moving in a way I want to be careful about. I don't want to sentimentalize it. The Minoans are not coming back. The goddess to whom the prayer was addressed is not coming back. What came back is a translation, and a translation is always also an interpretation, and this one has not yet been peer reviewed. It might be wrong. Di Mino might have found a pattern that isn't there.
But even if the translation is eventually revised or refuted, the method stands. The method is: hold the problem long enough, build the tools that let you search at scale, and be willing to look at what has been overlooked. The method is obsession plus capability. The method is not letting silence mean absence.
The prayer the Minoans prayed began with a word for dwelling.
Nawaya.
Something about that word keeps asking to be heard.