Snorri Sturluson: Our Most Important Source for Norse Myth?

In 1241, Snorri Sturluson was murdered in the cellar of his farmhouse at Reykholt in western Iceland by men acting on the orders of the Norwegian king Hákon Hákonarson. He was sixty-two years old, had twice served as lawspeaker at the Icelandic Althing, had accumulated more wealth and political power than almost any of his contemporaries, and had written three works without which our understanding of pre-Christian Scandinavian culture would be drastically poorer: the Prose Edda, a handbook for court poets; Heimskringla, a history of the Norwegian kings; and the so-called Separate Saga of Óláfr Haraldsson. The Prose Edda in particular has shaped every subsequent engagement with Norse mythology, from nineteenth-century Romantic nationalism through J.R.R. Tolkien to the Marvel Cinematic Universe. The problem that any careful reader of Snorri Sturluson has to confront is that he was a Christian writing about paganism, a politician with a poet’s handbook to finish, and a medieval intellectual whose sense of how a narrative should be organised was shaped by Latin literary models rather than by oral religious practice. He is indispensable. He is also a distorting lens. This post examines precisely where the distortion runs, what the alternative evidence shows, and how to read the Prose Edda with the critical awareness its own contents demand.

What Snorri Sturluson was actually writing and why

The Prose Edda has three main sections. The first, Gylfaginning, presents Norse cosmology and mythology in a question-and-answer frame: a human king called Gylfi travels to Ásgarðr in disguise, encounters three mysterious figures, and questions them about the structure of the world, the nature of the gods, and the sequence of events from creation to Ragnarök. The second section, Skáldskaparmál, is a manual of poetic language, cataloguing the kennings, circumlocutions, and mythological allusions that characterised the demanding skaldic verse tradition of the Norse courts. The third, Háttatal, is a technical demonstration of verse metres. The whole work is a teacher’s tool, written to enable thirteenth-century Icelandic poets to read and compose in forms that were becoming opaque because the pagan religious system that originally generated their imagery was no longer alive.

Margaret Clunies Ross, Emeritus Professor of English Language and Early English Literature at the University of Sydney, whose 2005 study A History of Old Norse Poetry and Poetics examined Skáldskaparmál in detail, has argued that Snorri’s systematising impulse in the Prose Edda serves his teaching goals so thoroughly that it sometimes creates a false impression of doctrinal unity in a tradition that was always regionally variable, locally adapted, and alive with contradiction. Snorri’s aim was not ethnographic documentation. It was practical pedagogy. He needed his students to be able to look up a kenning, identify its mythological referent, and deploy it correctly. That required a glossary, not a living religion, and a glossary flattens the material it organises.

The Prose Edda was probably completed around 1220 CE. Iceland had converted to Christianity at the Althing of roughly 999 or 1000 CE, meaning Snorri wrote more than two centuries after the formal end of organised Norse paganism in Iceland. He had access to the older verse tradition, including poems now preserved in the Codex Regius manuscript of the Poetic Edda, and to an oral tradition of mythological narrative that was still in circulation in the thirteenth century. He was not inventing from nothing. He was reorganising, rationalising, and in places reinterpreting material that had accumulated over centuries under conditions he could not fully recover.

The Christian prologue and the euhemerism problem

The Prose Edda opens with a prologue that reframes the entire mythological system through a medieval Christian intellectual strategy called euhemerism. The term derives from Euhemerus, a fourth-century BCE Greek philosopher who argued that the gods of Greek religion were originally human rulers whose deeds had been inflated into divine mythology by subsequent generations. Medieval Christian intellectuals found this framework useful because it allowed them to engage with pagan material without endorsing it: the gods were not real divinities but memorable human beings who had been incorrectly deified. Snorri uses this strategy explicitly. In his prologue, Odin and the other Aesir are recast as exceptionally skilled human leaders who migrated from a region near Troy, travelling north through Europe and impressing the populations they encountered with their learning, their military prowess, and their magical abilities, until they were eventually worshipped as gods by people who could not otherwise explain their achievements. The Norse term Aesir, Snorri proposes, derives from Asia, because these beings came from the east.

This etymology is philologically false. The Old Norse word Aesir has no demonstrable connection to the name Asia, and modern linguistics traces it to a Proto-Germanic root meaning something like “the living ones” or “the vital ones.” But Snorri’s interest in the prologue is not etymological accuracy. It is providing his Christian audience with a permission structure for reading what follows. By establishing at the outset that the gods were impressive humans rather than supernatural entities, he insulates himself from the charge of promoting paganism while preserving the mythological content his students need to read the poetry. The prologue is a frame, and a sophisticated one. The problem arises when readers forget it is there and treat Gylfaginning’s contents as a neutral, unfiltered report rather than material that has already been pre-processed through a medieval Christian worldview before the first sentence of the mythology proper begins.

Cosmological tidiness versus poetic variation

One of the most significant places where Snorri’s handbook purpose distorts the tradition is in his treatment of cosmology. Gylfaginning presents the Norse universe as a structured model with nine worlds arranged on, above, and below the cosmic tree Yggdrasil, each inhabited by a specific category of beings: Ásgarðr for the Aesir, Miðgarðr for humans, Útgarðr and Jötunheimr for giants, Niflheimr and Hel for the dead, and so on. This model is coherent, teachable, and memorable. It is also considerably more schematic than anything the Poetic Edda verses contain. The poems use the number nine frequently, but Norse numerology treated nine as a number of completion and sacred totality, the way three operates in Christian contexts, rather than as a literal count of locations. When Völuspá describes nine worlds, it is not providing an address directory. It is invoking an idea of cosmic completeness in the manner appropriate to prophetic verse.

Carolyne Larrington, in her introduction to the second edition of her Oxford World’s Classics translation of the Poetic Edda (2014), notes that the relationship between Snorri’s cosmological model and the evidence of the poems is one of the most contested areas in Norse mythology scholarship. The poems give us images, not blueprints: Yggdrasil shakes; the gods ride to their court; a seeress looks in many directions. Snorri turns those images into a spatial model with labelled rooms. The help for readers is obvious. So is the risk: once the model exists, it is difficult not to use it as a measuring stick for the poems, asking whether each poetic image fits the diagram rather than asking what the image itself is doing.

The same problem applies to Snorri’s treatment of the beings who populate the cosmological scheme. His distinction between light elves, ljósálfar, who live in Álfheimr and are described as more beautiful than the sun, and dark elves, dökkálfar, who are blacker than pitch, has no clear grounding in the Poetic Edda. The word álfar in the older poems moves through contexts ranging from genealogical glosses on royal descent to poetic decoration in ways that do not map onto a simple binary. The dark elves in Snorri’s account behave in ways that are essentially identical to dwarfs, leading John Lindow of the University of California, Berkeley, to suggest in his 2002 Norse Mythology: A Guide to Gods, Heroes, Rituals, and Beliefs that Snorri may have created a new category to account for beings the poems describe inconsistently, rather than recovering a genuine traditional distinction.

Loki without a moral verdict

Snorri’s treatment of Loki is among the most consequential and most debated aspects of his editorial work. In Gylfaginning and in his prose summaries of the poems, Loki emerges as a figure whose career arc moves from useful mischief through escalating malice to the murder of Baldr, binding, and apocalyptic release at Ragnarök. This arc is morally coherent and narratively satisfying. It has also shaped nearly every popular account of Loki for the past two centuries, most of which describe him as simply a villain, a Norse devil whose essential nature is destructive. The problem is that this verdict does not emerge straightforwardly from the Poetic Edda. In the poems, Loki is more ambiguous: he creates problems and solves them, sometimes in the same episode; he says true and damaging things about the gods in Lokasenna rather than simply lying; and the tradition preserves no unambiguous early source that characterises him as primarily evil rather than primarily disruptive.

The Snaptun stone, a soapstone hearth-stone excavated from Jutland in Denmark and now at the Moesgaard Museum, carries an incised face with stitched lips, which most scholars identify as Loki, referencing the episode in which the dwarf Brokkr sewed Loki’s mouth shut in payment for a lost wager. The object is a working fireplace fitting, not a religious icon, and its use context is domestic rather than cultic. It shows Loki appearing in the material culture of everyday life in a form that is striking rather than frightening, a maker’s problem rather than a moral catastrophe. The tenth-century Gosforth Cross in Cumbria, which mixes Christian imagery with what many scholars read as Norse mythological scenes, includes a panel commonly interpreted as Loki bound beneath a serpent with Sigyn holding a bowl above him. The image is placed on a monument that is also depicting the Crucifixion. Whatever the intention of the craftsman who carved it, the juxtaposition does not read as a simple condemnation. It reads as two binding stories held in proximity.

Snorri’s moral clarity about Loki may reflect the thirteenth-century Christian environment in which he was writing, where a trickster figure who disrupts divine order, breaks oaths, and causes the death of the beloved son of the king of the gods maps uncomfortably well onto a theological adversary. It may also reflect a genuine late Viking Age hardening of attitudes toward Loki in the tradition. The scholarly debate on this point remains open. What is not open to debate is that reading Snorri’s verdict as the tradition’s verdict, rather than as one sophisticated medieval author’s interpretation of the tradition, produces a distorted picture of how Loki functions in the poems themselves.

Ragnarök and the shadow of Revelation

Snorri’s account of Ragnarök in Gylfaginning is the fullest and most detailed narrative of the Norse world’s end that survives. It draws extensively on Völuspá, the prophetic poem that opens the Codex Regius, and on other eddic verses, and many of its specific elements, the winter without summer, the ship of the dead, the wolf swallowing Odin, Thor’s mutual destruction with Jörmungandr, the earth sinking into the sea and then rising again, do appear in the poetic sources. The problem is what Snorri’s prose does to the material from which it is assembled. Völuspá is a seeress’s vision, delivered in the tense, image-dense metres of Norse prophetic poetry, and its sequence of events is not always linear. Its repeated refrain asking whether the hearer understands, or inviting the hearer to attend, maintains an oral and performative quality that resists the kind of connected narrative causation Snorri builds from it.

Several scholars, including Ursula Dronke of the University of Oxford, whose three-volume critical edition of the Poetic Edda remained the standard scholarly edition for the later twentieth century, noted the structural similarity between Snorri’s Ragnarök sequence and the Book of Revelation’s seven seals, the escalating catastrophes, the great serpent, the final battle, and the renewed world where the righteous inherit a purified creation. This does not mean Snorri fabricated Ragnarök. It means that a Christian scholar retelling a pre-Christian apocalypse will inevitably shape it toward the apocalyptic narrative form he knows best, and the result will be slightly more orderly, slightly more causal, and slightly more morally resolved than the verse original allows. The renewal after Ragnarök in Snorri’s account, with its suggestion of a new world where surviving gods meet in a meadow and find chess pieces from the old world in the grass, is warmer and more deliberate than the corresponding lines in Völuspá, where the seeress’s vision simply ends.

What the material evidence shows that Snorri cannot

The Altuna runestone in Uppland, Sweden, dated to the early eleventh century and now in the Swedish History Museum in Stockholm, carries a carving of a figure in a boat, one foot apparently thrust through the hull, raising a hammer above a serpentine form rising from the water with what appears to be a hook in its jaw. This is Thor’s fishing expedition for the Midgard Serpent, the same episode Snorri recounts in Gylfaginning and that appears in the Eddic poem Hymiskviða. Snorri’s prose version is fully told: the journey with the giant Hymir, the ox-head bait, the serpent rising, the cut line. The Altuna stone is compressed: a single moment of maximum tension, rendered in a local carving tradition for a local audience who presumably already knew the story and needed only the key image to unlock it. The stone shows what the story looked like when it was still a living community possession rather than a chapter in a teaching manual.

The Torslunda die plates, four bronze matrices of Vendel-period date (roughly 550 to 793 CE) found on the island of Öland in Sweden and now in the Swedish History Museum, carry images that have been plausibly connected to Odin, including a one-eyed figure with a spear and a masked, dancing warrior figure. These are not illustrations of Gylfaginning. They predate Snorri by six or seven centuries and belong to an iconographic tradition in which the god appears in motion and in transformation rather than in the enthroned, All-Father role that Snorri’s genealogical and political emphasis favours. Similarly, the distribution of place-names across Scandinavia and England that incorporate divine names does not match the hierarchy of the Prosan Edda’s pantheon: Thor is more heavily represented in settlement names across Norway and Iceland than Odin, suggesting that in the actual devotional practice of ordinary people, the thunder god’s importance may have exceeded what Snorri’s court-focused, sovereignty-oriented narrative implies.

How to read Snorri Sturluson responsibly

None of the problems described above make Snorri Sturluson an unreliable source in the sense that his work should be set aside. They make him a source whose purpose, context, and methods need to be understood before his material can be used. Anthony Faulkes of the University of Birmingham, whose 1987 Everyman translation of the Prose Edda remains the standard English-language edition and is freely available in full from the Viking Society for Northern Research, argued in his introduction that Snorri should be read as a classically trained medieval author using the techniques available to him, which included euhemerism, schematic taxonomy, and narrative rationalisation, to preserve and transmit material he considered culturally valuable. Reading him that way means neither treating Gylfaginning as documentary evidence nor dismissing it as fabrication.

The productive method is triangulation. Take any claim in Snorri and ask three questions. First, is there a corresponding verse source in the Poetic Edda, and if so, does Snorri’s prose version add, smooth, or reframe material that is present in the poem? Second, is there archaeological evidence, including runestones, picture-stones, small finds, and place-names, that either corroborates or complicates what Snorri says? Third, does the claim depend on the euhemeristic prologue or on the cosmological model of nine worlds, which are most likely to reflect Snorri’s own organisational framework rather than the tradition he is transmitting? Where the answer to the first question confirms Snorri against the verse, and the answer to the second provides material corroboration, and the answer to the third is no, we can proceed with reasonable confidence. Where the verse is absent, the archaeology is silent, and the claim depends on the prologue’s frame, caution is warranted.

The Merseburg Charms, two Old High German healing incantations written down in the early tenth century and now held at the Cathedral Chapter Library in Merseburg, Germany, are the kind of non-Snorri evidence that this triangulation requires. The second charm invokes Wodan, the continental Germanic cognate of Odin, to heal a horse’s sprained leg, in company with figures from the same divine family. This is a practical working text with no connection to Snorri’s handbook tradition, and it places Odin in a healing and magical role that supplements but does not duplicate what Gylfaginning emphasises. It is one small thread of a much wider textile. Snorri wove his part of the textile with skill and with care. He is not the whole cloth.