The Ledger and the Myth

Some of the oldest writing in any European language is a list of chariot wheels.

The clay tablets come from palace archives at Pylos and Knossos, a little over three thousand years old, written in the script we now call Linear B. There are thousands of them, and they are all like that: oil, wool, rations, javelin shafts, wheels. Inventories. Receipts. The epics already existed at the time, sung from memory by bards who had carried them for generations, and nobody considered any of it worth clay. Writing, in its first European incarnation, was not for songs. It was for the storeroom.

The tablets were never meant to survive. They were unbaked clay, sun-dried and reused, and they outlived their civilization by accident, fired hard in the conflagrations that burned the palaces down around them. Then Greece collapsed into its dark age and forgot writing altogether for something like four hundred years. The songs went on without it. When writing finally returned, adapted from a Phoenician script, one of the first things anyone did with the new alphabet was write down Homer. So the sequence runs: centuries of song kept in memory alone, then an interlude in which the only things written were receipts, then the songs, at last, on paper. The ledger came first.

The Iliad itself mentions writing exactly once. In Book 6, Bellerophon arrives in Lycia carrying a folded tablet, and the tablet bears the instruction to kill the man who brings it. The single appearance of writing in the founding epic of the West is a death warrant, a transaction sent ahead of its object. The bards, as Yann Martel imagines them, looked at writing the way working musicians look at a new gadget: a little gimmick, gone tomorrow. What could be more powerful than the spoken word? They kept singing, and the accountants kept counting, and for a long time nobody confused the two trades.

The novelist and the unread epic

This weekend I sat in the Meighen Forum at the Stratford Festival for an event titled On Myth and Meaning: Ali Hassan, the CBC host, in conversation with Yann Martel, who wrote Life of Pi and has just published his fifth novel, Son of Nobody. The new book is itself an artifact from the boundary I am describing. Each page is split horizontally. Above the line, in free verse, run the fragments of the Psoad, a lost Trojan War epic that Martel has invented, following Psoas of Midea, a goatherd's son, one of the common soldiers the Iliad never lets speak. Below the line run the footnotes of Harlow Donne, the Canadian scholar at Oxford who supposedly discovered the fragments, and the footnotes slide by degrees from scholarship into confession. Martel named him for John Donne: no man is an island. The scholar, cutting himself off from his own life to save a dead poem, becomes one.

Martel told the audience that he read the Iliad for the first time only a few years ago, in Stephen Mitchell's translation, a "lean, mean, breathless" rendering he recommends. He expected something venerable and slightly boring and got, in his description, Game of Thrones. Two things struck him. The first was that no commoner speaks in the entire epic, with one exception: Thersites, in Book 2, who stands up in the assembly to say what the war has cost ordinary men, and is beaten by Odysseus for it while the army laughs. The one common voice in the founding text of the West is a man being struck. The second thing that struck Martel was that the war is a myth. There is a Troy, nine buried layers of it at Hisarlik in western Anatolia, but no attested ten-year war, no Agamemnon, no Achilles, no horse. The books the Greeks used as their school and their dictionary, the texts that taught them how to live, are fictions.

The case for the unverifiable

Martel's claim, delivered without apology, is that the missing facts are the design.

History, he argued, cannot be played with. It can be studied, corrected, deepened, argued over within limits. Of Napoleon we know something like ten million facts, and all those facts pin him down; you cannot retell Napoleon any way you like, because the record will stop you. Of Jesus of Nazareth we know almost nothing, and of the Trojan War nothing at all, and that is precisely why they endure. "There's something about the plasticity of things for which there are no facts," Martel said, "that means we can really make them our own." A myth survives because no one can check it, and because no one can check it, every generation can move into it and rearrange the furniture.

He has a traveler's proof for this. He went to Hisarlik and found, by his account, a little dump of an archaeological site: worn bricks, silted rivers, wheat fields where the bay was. The most evocative object on the site is a huge wooden horse built by the Turkish tourism authority. Jerusalem impressed him, but the shrines did not: you do not want, he said, a sweaty, half-lit cave with a muttering, bored priest to be the place where your myth lives. "It's good that Troy is a dump, because then you can make something up." The relic would falsify the myth. The myth survives by having no relics.

His larger conclusion is that a civilization which congratulates itself on its rationality stands on two fictions, the Homeric epics and the Gospels, and that identity lives in that register, the register that cannot be checked. I heard this in a packed auditorium on a Saturday morning and found myself unable to dismiss it. It is the most serious case I have encountered for the uncheckable.

The fact-free candidate

In the Q&A, someone asked Martel how the facts and unsolvable crises of the time, politics and climate, enter his fiction. His answer wandered, and then it snapped into focus, and it was the moment of the morning.

The interesting thing about Donald Trump, Martel said, is that he is, structurally, the same kind of object as Troy. "He is himself a fact-free tradition." He lies constantly, flagrantly, in ways newspapers correct within the hour, and it costs him nothing, because the corrections are aimed at a register he does not occupy. "Donald Trump has managed to speak at the level of myth, and every other president before him has not managed to do that." The slogan is a founding myth in shape: we used to be great, something was taken from us, we will restore it. You can argue particulars all day against a story like that and never land, because belief was never the mechanism. "You only defeat that kind of story with another story," Martel said, "another mythological story." And those, he noted, cannot be ordered up. For every Jesus there is a graveyard of messiahs nobody remembers.

This is a novelist reading politics through his own material, on a stage in Canada, and it deserves the discount that implies. But I will say for it: it is the first account I have heard of the defining confusion of the last decade, the failure of refutation, that explains the mechanics instead of deploring the morals. A claim that crosses out of the checkable register and into the myth register becomes audit-proof. Refutation is a ledger tool. It does not work on songs. Where his answer thins is the remedy, and he knows it: the counter-myth cannot be commissioned, so the prescription is to wait for one, or become one.

The opposite crossing

This blog has spent the year documenting the crossing in the other direction: claims born in the myth register, dragged onto the ledger.

Two cases, both fresh. Bittensor's Delegates Charter carries a sentence I once quoted with respect: the purpose of power is to give it away. That sentence is a myth in Martel's exact sense, beautiful, unfalsifiable, available to everyone. Then someone read the chain and found the sudo key live while the Senate pallets the documentation described were gone, and the founder himself published the concession that the protocol is "currently not decentralized" at its core; I wrote that record up in Human, All Too Human. And this week a seventy-day training competition's diary, a public random genesis with a heroic first-week climb, met a weight-level audit and turned out to descend, parameter by parameter, from an Alibaba datacenter model; that record is in Crowned on the Second Day. In each case the story ruled exactly as long as nothing could check it, and collapsed within hours of the check being published.

Martel's framework predicts the shape of these episodes. A myth can govern only where there is no ledger. What this decade keeps adding is ledgers in places that never had them. Weights can now be aligned against their ancestors tensor by tensor. Chains can be read by anyone with a laptop. A benchmark suite can be rerun under a pinned evaluation pipeline by anyone with an afternoon and a GPU. Even the classics arrive with an unaudited lineage, as Martel's own example showed: he described asking Emily Wilson why her Iliad translation places Achilles' and Patroclus' beds together in the middle of the tent, where every earlier translation separates them, and her answer came down to one Greek preposition. A single translator's choice, invisible unless you check the original, changes a relationship three thousand years old.

The democratic novel

Martel called Life of Pi a democratic novel, and the description turned out to be a piece of the same framework. There is one set of facts: a ship sinks, one survivor reaches Mexico after 227 days. Over those identical facts the book lays two irreconcilable stories, and Martel was careful, he said, to make neither impossible. The reader votes. The vote reveals the reader.

The audit this week left exactly such a fork, and I said so at the time. The inheritance is proven; the intent is not. The weights cannot show who knew. Where the ledger ends, the story you choose to tell about that gap is a confession, and everyone involved is telling one now.

The myth the ledger needs

Honesty requires the other half of this, and I want to give it room, because the temptation in essays like this one is to anoint the checkable side and go home. The industry I work in runs on myth too, and not accidentally. Bitcoin's founder published a nine-page text and a ledger, and then vanished, leaving no biography, no body of facts, nothing to falsify, which is the exact myth structure Martel describes, and its plasticity is the reason the origin story still coordinates millions of strangers: everyone can move into it and rearrange the furniture. The ghost who built private money, Nicolas van Saberhagen, whose trail I tried to follow, is the same structure twice over. Whitepapers are scripture in form if not in intent; a pope who grasps decentralization better than most protocols is, at some level, a joke about our canon. And the fiction this blog keeps warning about, the market's story that a token is a claim on a model or a team, priced as if proximity were ownership, is a myth in the strict sense: believed beyond the facts, durable because nothing in the price mechanism forces the check. So the two registers stack into layers, and I suspect every durable institution needs both. The question is never whether a project has a myth, because they all do. The question is whether the myth submits to the ledger or replaces it. Bitcoin's trick, if you want to call it that, was a myth whose content is the ledger: the story says check the receipts, and the receipts check out, and the founder's absence means the hero of the story can never be caught lying. Whether the verification side of the world can grow a counter-myth deliberately, the way Martel says politics requires, I do not know. Myths that work cannot be commissioned, that is his point and I believe it, and "we have the receipts" has never once been a song anyone sang. I keep turning this over and I have not landed.

What each register is for

Here is the division of labor I carried out of the theater.

Myth holds what measurement kills: belonging, consolation, identity, the reasons people get up in the morning. The ledger must hold whatever power touches: anything that allocates, rewards, excludes, crowns, censors. The two failure modes are mirror images, and both are on display this decade. Power stored in the myth register is unaccountable by design; that was the sudo key wearing the Charter, and the crown sitting on imported weights. Meaning stored in the ledger register dies of exposure; that was Martel's half-lit cave in Jerusalem, the place he did not want his myth to live. The signature error of the age is the crossing. Facts get sold as myths so they cannot be checked. Myths get policed as facts until they cannot console. A workable civilization keeps both technologies and remembers which is which. The audit age, the thing this blog has been documenting one episode at a time, is the ledger reasserting itself in rooms where songs had been running the accounts.

The tablets at Pylos survived because the palace burned around them, baked hard by a fire the scribes never planned for. The epics survived the same centuries by a different trick: they were not written down at all, so there was nothing to burn, only something to sing. The accountants' records endured because they were written. The bards' endured because they were not.

---

Source note: Yann Martel's remarks are from "On Myth and Meaning" at the Meighen Forum, Stratford Festival, August 22, 2026, hosted by Ali Hassan; the author attended. Quotations are transcribed from a recording of the event and lightly corrected where machine transcription garbled the sense (the Iliad's commoner is Thersites). Book details are from Son of Nobody (Knopf, 2026). The Linear B tablets, the Bellerophon passage (Iliad 6), Thersites (Iliad 2), the layers at Hisarlik, and the survival of the tablets by palace fire are standard scholarship. The Emily Wilson translation anecdote is Martel's account, told on stage. The blog posts linked above are this site's own record.