Snorri Wasn't Wrong, He Was Early

On why a thirteenth-century Icelander had better priors about religion than eight centuries of scholarship that came after him.

Around 1220, in Iceland, a politically ambitious historian named Snorri Sturluson sat down to write a handbook of Norse mythology for the next generation of skalds. The oral tradition was collapsing, Christianity had won, and if the old poetry was going to survive at all, it needed a manual. Snorri wrote one. He called it the Prose Edda, and he opened it with a prologue that has embarrassed his readers for eight hundred years.

The prologue claims that Odin was a Trojan prince.

Not a god. Not an archetype. A man. Specifically, a descendant of Priam who fled the ruins of Troy after the Achaean sack, migrated north through Saxony into Scandinavia with a retinue of warriors and wise men, established dynasties along the way, and was eventually remembered as divine by populations who no longer understood where he had come from. The word Æsir, Snorri claims, is etymologically related to Asia. The whole Norse pantheon, in his telling, is a displaced Anatolian warrior caste mythologized over centuries by Scandinavian peasantry who had lost the thread.

The consensus view, for most of the last four centuries of serious scholarship, has been that Snorri was either lying, fabricating, or doing a kind of medieval Christian apologetics: making his ancestral religion legible inside a Virgil-Homer prestige frame so the Church would not burn the manuscripts. Every medieval European people claimed Trojan descent. The Franks invented Francio, the Britons invented Brutus, the Romans had Aeneas and made a whole epic of it. Snorri giving the Norse a Trojan founder was, in this reading, just another instance of the same prestige move. Harmless, genre-bound, not meant to be taken as history.

Price the alternatives honestly. What does modern academic mythography offer in place of Snorri's claim? The dominant framework, in various forms, holds that gods like Odin are purely constructed, that Norse theology is a literary-religious system developed by its practitioners for sociological, aesthetic, or functional reasons, with no particular person at any root. The Odin of the Eddas is a crystallization of Proto-Germanic religious intuition, shaped by priestly classes, refined in poetic transmission, never reducible to any historical figure because no historical figure is there to reduce to.

This is the respectable position. It is also, if you think about it for longer than a minute, extraordinarily strange.

We have zero examples of religions forming this way in historical time. Not one. Every religion we can actually watch coming into existence has a person at its root. Christianity has Jesus. Islam has Muhammad. Buddhism has the Buddha. Mormonism has Joseph Smith, whose life is documented in newspaper archives. The cult of deified Roman emperors has specific emperors, named, with birthdays we can look up. The cult of Imhotep, which lasted two thousand years in Egypt, began with an actual architect who served Pharaoh Djoser and got gradually upgraded, first to demigod, then to god of medicine, over about a millennium of accumulating reputation. The Hawaiian aliʻi who became akua, the Shinto kami who began as remarkable ancestors, the hundreds of Christian saints who ate local pagan spirits across Europe and absorbed their functions: every case we can inspect shows the same arrow. Embodied person, then memory, then mythologization, then god.

The reverse arrow, abstract theology crystallizing without any person to anchor it, is not an arrow we have ever actually observed. It is a theoretical category without empirical instances. A peculiar foundation for a whole field.

Snorri, writing in 1220, had better priors than the field that spent the next eight centuries dismissing him. He knew that religions form around persons because every religion he had direct knowledge of had formed around a person. Christianity was four miles from his front door. He was a Christian himself. The whole European intellectual landscape he worked inside was built on the premise that a real man in Judea had done specific things at a specific time, and that the theology came after, not before. Why would he assume the Norse case worked differently? The parsimonious move for a medieval historian looking at the mythology of his ancestors and asking where it came from was to assume it worked the way the only other case he knew worked. Person, then story, then god.

His mistake was the Troy part.

What he got wrong, narrowly, matters. Norse religion is not descended from Mediterranean religion through a single migration of a named warrior. The relationship between Norse and Greek pantheons is real, but it runs much deeper than the Bronze Age. Both are branches of Proto-Indo-European religion, a common ancestor roughly five thousand years back, somewhere in the Pontic-Caspian steppe or Anatolia depending on which reconstruction you accept. The sky-father name Dyēus ph₂tḗr erodes through phonological change into Zeus, Jupiter, Tiw, Tyr: same god, same function, same etymology, articulated differently by cousin languages that diverged millennia before Troy existed. The pantheons look parallel because they are parallel. Siblings with a shared ancestor. Not independently invented, not convergently evolved, not migrated from one to the other. Inherited.

So Snorri was wrong about Troy. He was right that there was a kinship between Norse and Mediterranean religion. He was right that it ran through persons and migration at some level. He just misidentified the time-depth and the mechanism. The Indo-European expansion rather than the fall of Troy. A migration of a whole people over a millennium rather than a single prince and his retinue over a generation. Five thousand years rather than one thousand eight hundred.

A distinction worth drawing. There are two axes of accuracy here. The first is narrow factual accuracy: did the specific claim, as stated, correspond to historical events exactly as described? On this axis Snorri scores badly. No Trojan prince migrated to Uppsala. Maybe three percent, being generous.

The second axis is ontological category fit: does the claim point at the right type of explanation, even if the specifics are garbled? On this axis Snorri scores somewhere around eighty-five percent, and modern academic mythography scores much lower. The academic framework locates itself in a category of explanation where religions form by abstraction from cultural need, which has no observable examples in the historical record. Snorri locates himself in a category where religions form around embodied persons and generate mythology over time, which has every observable example in the historical record.

When you have to choose between a compression that has the right category and the wrong specifics, and a compression that has the wrong category and no specifics at all, the first is closer to truth. Not because truth is unknowable and therefore any plausibility will do, but because the category is where most of the epistemic work lives. Get the category right and the specifics can be corrected. Get the category wrong and the specifics will always be wrong even when they look right.

There is a deeper thing underneath this, which is why the embodied-god story keeps winning across cultures that had no contact with each other. Gods who were once people are sticky. Their biographies carry narrative density that pure abstractions cannot match. A god who hung himself on a tree for nine nights to seize runic knowledge is memorable in a way that a "god of wisdom" as a bare category is not. A god whose eye is permanently in a well because he traded it for a drink is a god you can tell stories about, and stories get retold. This is how cultural transmission actually works under conditions of oral literacy, which is the condition most human cultures have operated under for most of human history. What survives is what can be told. What can be told is what happened to someone. Abstractions evaporate across three generations. Biographies last three thousand years.

The academic framework, then, is not merely empirically unusual. It is mechanistically implausible. It posits a transmission process that cannot actually transmit the thing being transmitted. Whereas Snorri's framework, embodied origin and gradual mythologization and accumulated biographical density, is how the transmission observably works, every time we can watch it work.

Was Snorri right about Odin being a Trojan prince? No.

Was Snorri right that Odin was once a person, that the person came from somewhere south, that he arrived into a religious landscape that already had a slot shaped for him, that his biography fused with a pre-existing archetype, and that what we now call the Norse pantheon is the end result of this process operating over many centuries? Much more likely than not. The mechanism is correct. The arrow is correct. The time-depth and the specific geography are garbled, because Snorri did not have access to comparative Indo-European linguistics or Bronze Age archaeology or the Ynglinga saga genealogies parsed against radiocarbon dating. He had oral tradition, Christian prestige frames, and his own instincts about how religion forms. He used what he had. The instincts were sound.

What is sobering is that eight hundred years of subsequent scholarship, with vastly better tools, mostly moved away from his instincts rather than toward them. The field professionalized into skepticism. The skepticism became orthodoxy. The orthodoxy forgot to check whether the phenomenon it was skeptical of, religion forming around real people, was actually a rare or debatable occurrence, or whether it was just how religion has always formed, every time, without exception, in every case available for inspection.

Snorri, reading the bones of his own religion, reached the conclusion that someone had to have been the bone. He was right about the existence of a bone. He was wrong about whose it was and where it came from. Under those conditions, and in that epistemic environment, that is about as close to truth as a historian working in 1220 was going to get.

The field owes him an apology it will probably never offer, because the apology would require the field to notice that its own foundational premise is the less parsimonious one, and fields do not cheerfully notice such things about themselves.

The reader can notice. That is the point of writing this down.