Creation from Fire in Snorri’s Edda: The Tenets of a Vernacular Theory of Geothermal Activity in Old Norse Myth

Mathias Nordvig, University of Colorado, Boulder
Abstract: This article argues that Snorri’s version of the creation myth in Snorri’s Edda contains imagery from volcanic activity described in terms of a “vernacular theory of geothermal activity”. The vernacular theory of geothermal activity mythologizes natural events, volcanic and otherwise, and describes them in terms of relatable analogical and metaphorical imagery sampled primarily from the cultural vault of Nordic mythology. There are several natural aspects of Snorri’s creation myth, which seem to not fully make sense if compared with the usual analogies of the natural behavior of ice and fire. These inconsistencies are fully explained if one turns to another analogy, namely the low discharge effusive eruptions, which are so common in Iceland. These processes cannot, however, be described in the vernacular vocabulary without analogies to ice and water, and this is why there is confusion about the details of the process. Lacking such words as “lava” or even “eruption”, the Icelandic creation myth in Snorri’s Edda turns to analogies such as “ice” and “river” to describe these phenomena. The end result is a creation myth based on indigenous theories on volcanic activity, and aspects of Neoplatonism inherited from continental philosophy.


The purpose of this article is to give insight into the ways in which early Icelanders communicated their experience of volcanic eruptions in myths and legends. It is my theory that if the early Icelanders had a vocabulary able to express the experience of volcanic eruptions, it was not particularly detailed. Therefore, early Icelanders may have used mythic language and linguistic tropes to express what they saw in eruptions. This theory relies on observations by scholars and scientists working with different cultures in many parts of the world, and it is a theory that has also been carefully suggested by Oren Falk in connection with the Icelandic material in his article, “The Vanishing Volcanoes” (Falk 2007). Falk has dubbed it the vernacular theory on geothermal activity (see below), and I find this designation clear and broadly applicable. To demonstrate how the vernacular theory of geothermal activity in early Iceland may have functioned, I will analyze the convergence of fire and ice in the creation myth in Snorri’s Edda as a sequence that borrows from the nature of a common Icelandic low discharge volcanic eruption. I will comment on some inconsistencies in the natural image of the creation myth, pointed out by Anne Holtsmark in her 1964 study Studier i Snorris mytologi. I will argue that the inconsistencies that Holtsmark had some difficulties reconciling are in fact explainable if the natural image of the convergence of fire and ice is compared with a lava flow. To strengthen this argument, I make use of the only medieval Icelandic source that describes the details of a volcanic eruption, the poem Hallmundarkviða. I approach the texts as medieval literature, which contains, on the one hand, ancient myth from oral culture, and, on the other hand, contemporary reflections on the part of the authors, based in medieval Icelandic culture and Continental teaching. First, I will contextualize this analysis with a brief overview of volcanism in Iceland and a survey of research into the connection between volcanism and Old Norse myths. I will then expound on the details of Snorri’s creation myth, focusing mainly on the language and descriptions associated with the components of Élivágar and eitr.

Volcanism in Iceland

Icelandic volcano types are highly varied. They range from shield volcanoes such as the infamous Eyjafjallajökull, which erupted in 2010; stratovolcanoes like Hekla; calderas; spatter cones; scoria cones; tuff cones; lava shields; chasms such as Eldgjá; mixed cone rows such as Laki; and, finally, rows of tuff cones and maars (Þorvaldur Þórðarson and Ármann Höskuldsson 2008: 201). Icelandic eruptions are as complex as the various volcano types. They can be explosive or effusive, or a mixture of both (Þorvaldur Þórðarson and Ármann Höskuldsson 2008: 202), and their eruption styles range from Surtseyan to Phreatoplinian in the case of wet eruptions, and in the case of dry eruptions, from Strombolian to Plinian (Þorvaldur Þórðarson and Ármann Höskuldsson 2008: 200). Naturally, the many eruption types have varied effects on the landscape, but the most frequent lava-producing eruptions create the type of rubbly pāhoehoe that is characteristic of the landscape in Iceland (Þorvaldur Þórðarson and Ármann Höskuldsson 2008: 207). Large-scale tephra-producing events are less frequent.
The impact of Icelandic volcanism ranges from local to hemispheric. Minor eruptions in Iceland can affect humans who live in the immediate vicinity of the volcano, but do not necessarily have an effect on the life of residents elsewhere in the region. Major eruptions, on the other hand, have been known to disturb Europe and even Asia. The hazards of Icelandic volcanism include lava flows, ash clouds and ash depositions, debris flows, gas emissions, lightning, and glacier bursts, also known as jökulhlaups. Magma that is exposed to water or ice erupts especially violently as the rapid cooling of the magma generates phreatic activity and large quantities of ash, which is deposited in a wide area. The generated meltwater from subglacial eruptions creates jökulhlaups, which can be highly debris-charged and lahar-like (Tweed 2012: 218).
While the stratovolcano Hekla remains the historically widest known volcano in Iceland—it is one of only a handful mentioned in the medieval Icelandic Annales Regii (pp. 76–155)—the two most severe eruptions of Iceland, in the little over 1100 years that the island has been inhabited, are the fissure eruptions of Eldgjá in 934 CE and Laki in 1783–1784 CE (Zielinski et al. 1995: 129). Eldgjá emitted some 19.6 cubic kilometers of lava and dispersed tephra across roughly 20,000 square kilometers. The eruption may have lasted as long as six years, until 940 CE. Laki emitted 15.1 cubic kilometers of lava and caused around 8,000 square kilometers of tephra fall. It lasted for eight months (Þorvaldur Þórðarson and Ármann Höskuldsson 2008: 212). The Laki fissures occurred in the 28 kilometer-long vent complex stretching from the southwestern tip of Vatnajökull to the icecap covering Katla. The eruption column in Laki was 12–15 kilometers high, and the height of the fire fountains was between 800 and 1400 meters high (Þór Þórðarson and Self 1993: 233). This also seems to have been the case in the Eldgjá eruption, which was even more violent (Zielinski 1995: 132). [1]
Except for a single reference to “an eruption” that displaced the settler Molda-Gnúpr in Landnámabók ch. 329, the Eldgjá event is not mentioned in written sources from Iceland. This means that there are no indications of the impact of the eruption locally, but based on historical sources from Europe, the Middle East (Stothers 1998), and from as far away as China (Fei and Zhou 2006), it can be concluded that the global climatic impact was severe. The 1783–1784 CE Laki eruption was likewise a great menace to the European climate, causing crop failure and famine (Grattan and Pyatt 1993; Grattan 1995), as well as elevated mortality rates as a result of air pollution (Grattan, Durand, and Taylor 2003; Courtillot 2005; Grattan et al. 2005). The famine following Laki caused a mortality rate as high as 25 percent in the Icelandic parishes that were most affected by the eruption (Vasey 1991: 342). Based on that, we may only guess what calamities Eldgjá brought with it in Iceland. Inferring from these events, the settlers coming from Scandinavia must have been terrified by the experience of fire fountains rising higher than the skyline of any modern city, roaring noises, which must have been audible in a large area, and an all-engulfing darkness caused by ashes in the atmosphere, blocking out the sun.
Situated on the mid-Atlantic ridge with approximately thirty singular volcanoes (Þorvaldur Þórðarson and Ármann Höskuldsson 2008: 199–200), Iceland is the most geologically active region in the world. It is estimated that over the last 1100 years, as long as there has been human settlement in Iceland, there have been approximately 250 large and small eruptions (Tweed 2012: 217). From the decade just before the landnám (the settlement period in the ninth century) until the mid-fourteenth century, there were roughly fifty-six verified eruptions. Thirty-nine of these occurred before 1220, the approximate time Snorri Sturluson wrote down his Edda, a prose collection of mythological narratives. Of these, four were effusive events, an additional four were of mixed effusive and explosive nature, and twenty-four were explosive events (Þór Þórðarson and Larsen 2007: 137). In the period from 1179 to ca. 1200, both an explosive and an effusive event occurred in the Katla system (Þór Þórðarson and Larsen 2007: 134), and in the period 1201–1220, two mixed eruptions and one explosive one were recorded to have occurred “somewhere in Iceland” (Þór Þórðarson and Larsen 2007: 137).
Thus nearly every generation of Icelanders, from the time of the landnám to the time when the mythology was recorded in writing, would have had the chance to experience volcanic eruptions. In these eruption types, there is sufficient material to inspire the imagination of Icelanders attempting to give accounts of the subterranean processes particular to the region. With so much geological activity, it is perhaps no wonder that the question of a volcanic imprint on the premodern literature of Iceland has occasionally boggled the minds of some scholars since at least the beginning of the twentieth century.

Research survey

In 1905, Bertha S. Phillpotts published the article “Surt” in Arkiv för nordisk filologi, arguing that the fire-giant Surtr was indeed a volcano-giant, and that his mythical role as the antagonist of the Æsir, who sets the world on fire, is reminiscent of observations of volcanism in Iceland. This argument served the greater purpose of arguing that the eddic poem Vǫluspá is originally an Icelandic poem. Sigurður Nordal welcomed this argument in his analysis of the poem in Vǫluspá from 1927. The interpretation has been widely accepted, though not universally (Falk 2007: 7). Nevertheless, Rudolf Simek reiterates the notion in the Dictionary of Northern Mythology (Simek 2007: 303–04).
Dorothy B. Vitaliano also investigates the question of volcanism in premodern Icelandic narratives in her book on geomythology, entitled Legends of the Earth (1973). Her focus is on a mix of medieval narratives, primarily the Kristni saga account of the eruption that created Kristnitökuhraun, and folk legends, such as one about the cause of eruptions in Katla volcano. Vitaliano’s investigations are unfortunately marred by a general lack of theoretical knowledge about literary genres, myth as a genre, and a critical grasp of premodern Icelandic literature.
More recently, Oren Falk has revisited the question in the article “The Vanishing Volcanoes: Fragments of Fourteenth-Century Icelandic Folklore” (2007). Falk provides an encompassing survey of the different narratives of Old Norse-Icelandic literature that can be associated with volcanism. He concludes his article by proposing an interpretation of certain passages in the saga literature, arguing that, for instance, images of dragons lying on gold hoards are expressions of a “vernacular geothermal theory” in a genre that, according to Falk, dares not speak of volcanoes (Falk 2007: 10–12).
In his article “Perception of Volcanic Eruptions in Iceland”, geologist Þorvaldur Þórðarson has made a case for a special pragmatic conception of volcanoes in premodern Iceland. He argues that pre-Christian Icelanders were less prone to react with awe and supernatural interpretations of the volcanic phenomena than their later Christian descendants were. This argument is based on his readings of Vitaliano’s interpretation of the story about the Kristnitökuhraun eruption compared with Herbert the monk’s fantastic descriptions of volcanic eruptions in Iceland in Liber Miraculorum from ca. 1170 CE (Þorvaldur Þórðarson 2010).
Archaeologists Bo Gräslund and Neil Price, in the article “Twilight of the Gods? The ‘Dust Veil Event’ of AD 536 in Critical Perspective”, discuss the theme of a natural disaster, possibly of volcanic origin, affecting the mythology of the Scandinavians in the Migration Age (Gräslund and Price 2012). Already in 2009, Gräslund suggested that the Fimbulvetr and related eschatological images in Snorra Edda have affinity to a volcanic catastrophe (Gräslund 2009: 318–26).
The most recent publication combining Old Norse mythology with volcanic eruptions is geologist Árni Hjartarson’s thorough comparison of the images of Hallmundarkviða with volcanic eruptions in his article “Hallmundarkviða, eldforn lýsing á eldgósi” (2014). Aside from pointing out that myths have been used to convey experiences of volcanic activity in a European context as far back as Greek mythology (Árni Hjartarson 2014: 28), he meticulously explains the connection between the poem, eruptions, and certain mythical figures in the narrative.
In this context, it should be mentioned that the idea that literary and oral traditions can shed light on natural phenomena and geological events has been gaining increasing acceptance across a wide spectrum of scholarly disciplines. The archaeologist-folklorist pair Elizabeth W. Barber and Paul T. Barber have provided an extensive interpretive method for understanding how oral societies relate superior geological events in their book When They Severed Earth from Sky (2004). Likewise, the anthology Myth and Geology, edited by L. Piccardi and W.B. Masse, reveals the potential of myths to yield data of value to the geological sciences. Scholars continue to find examples of how myths and legends narrate experiences of geological events. The most well-known example may be the myth (as much as 7,675 years old) of the Klamath tribe in northwestern North America about the eruption of Mount Mazama (Zdanowich, Zielinski, and Germani 1999). Most recently, a team of Australian researchers in geology and linguistics has discovered that indigenous Australians have retained up to 10,000-year-old stories about sea-level change in southern Australia (Reid, Nunn, and Sharpe 2014). My approach to the idea of a vernacular theory on geothermal activity in early Iceland can be expressed with the following quotation by M. Juneja and G.J. Schenk:
Nature is commonly perceived as an entity distinct from humans, possibly inhabited by the divine forces, endowed with degrees of power, a potentially dangerous—or alternatively sublime—force, to be domesticated, appeased, controlled by technology, in general a field of flux marked by a shifting relationship between humans, the environment and god(s). Local cosmologies provide an entry point into this relationship which they articulate through specific cultural media—texts, images and objects. (Juneja and Schenk 2014: 10)
This suggests that we can expect that the early settlers of Iceland, within the parameters of their worldview, would have attempted to exert some form of cultural control over the geo-activity that they faced. Old Norse mythology is a local cosmology in Iceland of the kind mentioned by Juneja and Schenk. It can therefore function as an entry point to the relationship between early Icelanders and the volcanic activity of the region. In light of this, and that millennia-old stories about geological events seem to occur worldwide (Vitaliano 1973; Barber and Barber 2004; Piccardi and Masse 2007; Cashman and Cronin 2008; Hamacher and Norris 2011), it should not be controversial to assume that the premodern Icelandic literary material may contain more data on volcanism than is commonly recognized.

Method of approach

The early Scandinavian settlers of Iceland presumably did not have much vocabulary with which they could express their experiences of volcanic events; they migrated to the highly volcanically active island of Iceland from the geologically stable Scandinavian region. They must immediately have been forced to find words for the phenomena. It is possible that in order to convey such experiences, and share them with posterity, myths and mythic language, functioning as master narratives, were employed to construct a meaningful and coherent account of eruptions. In When They Severed Earth from Sky, Barber and Barber describe a common human communicative strategy associated with geological and volcanic events. Oral narrative processing often takes the form of analogical tales, which describe the event in anthropomorphic terms: the mountain is a living entity or the dwelling of a being, and the eruption is caused by a supernatural being, a god, a demiurge, or a demon. The various details of eruptions are described in terms of comprehensible known objects. Lava, for instance, can be a “river of fire” (Barber and Barber 2004: 43). In more recent colloquialisms, the sound of a volcanic explosion has been described as “artillery”, the shockwave of an explosion as a “blizzard” (Barber and Barber 2004: 72–75). Tephra (rock fragments) can be described as “dry snow” (Beaudoin and Oetelaar 2006a, 2006b). Indeed, the word lava originates in the Latin lavare (to wash), and magma is originally Greek for “ointment”. That early Icelanders had trouble finding exact words to describe the various aspects of volcanism is attested in the use of interchangeable words and phrases for volcanic events in the annals and in Landnámabók, for instance jarðeldr, lit. “earth-fire” and eldzuppkváma, lit. “fire-up-throw”. The widely used Icelandic word for lava, hraun, is imprecise: it originally meant “little island”, “rubble”, or “stone ground” (de Vries 1961: 252)—a meaning that is still intact in other Scandinavian languages.
This is the essence of the so-called vernacular geothermal theory proposed by Falk. He writes: “[but] neither does it seem far-fetched […] to see the glowing dragons and their hoard as a stylised depiction of a volcanic crevice, and the devastating poisonous fire and blood as euphemisms for lava” (Falk 2007: 11). Upon presenting even more examples, Falk suggests that reformulating volcanic events in terms of a gold treasure that may burn, swallow, or in other ways kill you, relates to a certain hostility in both Icelandic law and the sagas to buried treasure (Falk 2007: 12). A taboo seems to have shrouded geothermal activity as much as the activity of burying treasure, but, Falk writes:
This folk tradition never quite got off the ground as a saga narrative motif. Whether as superstition or as proverb, such folklore would certainly have served useful social functions. Besides helping to explain to Icelanders the natural environment they saw all around them, a literal belief of this sort would have helped dissuade both treasure-hunters and would-be hoarders […] As metaphor, meanwhile, the equation of molten rock with burning gold would have had an even wider potential amplitude, whose full range, now lost to us, we can only imagine. (Falk 2007: 12)
Barber and Barber’s theory of linguistic modulation of geological phenomena in When They Severed Earth from Sky is very similar to that of Falk. They call it the Analogy Principle: “If any entities or phenomena bear some resemblance, in any aspect, they must be related” (Barber and Barber 2004: 34–40). In the case above, a glowing dragon, its gold hoard, blood, and poisonous fires all bear some resemblance to the products of an eruption. This principle of analogy is, in fact, expressed quite explicitly in context of Old Norse mythology. According to Snorri’s Edda, when the god Loki is bound in the underground after having caused the death of Baldr, it is said that he causes earthquakes. As Loki jerks his bonds whenever his wife Sigyn does not protect his face from the poison dripping from the venomous snake above him, he makes the ground shake. Hár, who is telling this story to King Gylfi, says “Þat kallið þér landskjálpta” (Gylfaginning p. 49) (That is what you call an earthquake). [2] This may be an example of a vernacular theory of earthquakes particular to Iceland, seeing as the medieval Norwegian Konungs Skuggsjá explains earthquakes as the result of winds entering caves in the ground and making the earth shake. This is a notion that has its origins in traditional classical literature, and can be found in, among other texts, Ovid’s Metamorphoses (Nordvig 2013: 166). The two explanations are quite different from one another, and it seems that the one including Loki may be of Scandinavian origin, whereas the other one is not. In the following, I will employ this interpretative strategy in connection with the creation myth in Snorri’s Edda, and address some inconsistencies in the myth, which Holtsmark found puzzling in her 1964 analysis.

The nature image in Snorri’s creation myth

Snorri’s creation myth describes how the cosmos was created from fire and ice. There was a cold world called Niflheimr with a well in the middle, which was named Hvergelmir. From this well, eleven ice-cold rivers flowed. In the southern part of the cosmos, there was another world, which was hot and fiery. It was called Muspell or Muspellzheimr, and it was guarded by a jǫtunn “giant” who brandished a flaming sword. The convergence of the cold from the icy rivers from Niflheimr and the heat and sparks from Muspellzheimr created a milder climate in the center, a region called Ginnungagap. When the water and the sparks met there, the world was eventually created (Gylfaginning pp. 9–10).
Several scholars have argued that the creation myth in Snorri’s Edda relies fully or in part on classical Latin learning, and that it is inspired either by the notion of the four elements (Holtsmark 1964: 29–30) or by Neoplatonic dualism (Dronke and Dronke 1977: 172; Faulkes 1983: 288; von See 1988: 53; Guðrún Nordal 2001: 273–77). While this may be the case, there are parts of the descriptions of the natural image of the convergence of heat and cold that are not entirely harmonic or, rather, seem to contest the normal rules of nature. It was Holtsmark who initially took notice of them. They occur in relation to the description of the rivers flowing from Hvergelmir, which reads as follows:
Þá mælir Hár: “Ár þær er kallaðar eru Élivágar, þá er þær váru svá langt komnar frá uppsprettunni at eitrkvikja sú er þar fylgði harðnaði svá sem sindr þat er renn ór eldinum, þá varð þat íss, ok þá er sá íss gaf staðar ok rann eigi, þá héldi yfir þannig úr þat er af stóð eitrinu ok fraus at hrími, ok jók hrímit hvert yfir annat allt í Ginnungagap.” (Gylfaginning ch. 5)
(These rivers that are called Élivágar, when they had come far enough away from the source, so that the poisonous flow that followed hardened like the cinders that run out of the fire, then it turned to ice, and when this ice stopped and did not run, then the vapor that stood off the poison froze to rime on top of it in the same direction, and this rime increased in layers all over Ginnungagap.)
In the following, I will concentrate on interpreting two important elements in this description of water and ice in the creation myth: eitr and the Élivágar. Initially, I will treat the term eitr, because this term has most potential to illuminate the vernacular theory of geothermal activity.


Holtsmark has some difficulties reconciling what happens with the eitrdropar “drops of poison” and the described process of smelting, quickening, and freezing. She writes:
Naturbilledet er ikke slående: elver som fryser til is, frostrøk over isen som rimer, det går an; men rim over hele Ginnungagap som senere blir beskrevet som hlætt sem lopt vindlaust er merkelig; det spørs om isen og eitrdropar fra Élivágar ikke tar omveien om rimet bare for å forklare navnet hrímþursar. (Holtsmark 1964: 30–31)
(The image of nature is not obvious: rivers freezing over, rime over the ice, that is reasonable; but rime over the whole of Ginnungagap, which is later described as hlætt sem lopt vindlaust [mild as a windless sky] is strange; one wonders if not the ice and eitrdropar from Élivágar take a detour over the rime only to explain the name hrímþursar [frost-giant].)
Holtsmark proceeds to attempt an explanation based on examples of the use of eitr in other instances. Eitr does not only mean “poison” but may also denote pus from a wound. It is also used in a word for a special type of glaciers in Norway and Iceland, which are called eitrár. The use of eitr can then be explained as a function that refers to a frozen waterfall coming over a cliff (Holtsmark 1964: 31). Holtsmark goes on to discuss the term sindr “slag”. Similarly, while she accepts the image of sindr as reasonable to anyone who has seen the process of iron ore smelting, she goes on to wonder:
Men der kommer jo motsetningen inn igjen, Élivágar er kalde, og en blåstermile er varm, og sindr størkner når det kommer i kaldere luft, mens eiter-elvene skulle fryse til is når de kom fra selve kulden ut i det varmere strøk, i Ginnungagap. (Holtsmark 1964: 31)
(But here comes the contradiction again, Élivágar are cold and the wind gust is warm, and the sindr quicken when it comes into colder air, while the eitr-rivers are supposed to freeze when they come from the cold itself into the warmer region of Ginnungagap.)
The image projected is that of the rivers flowing from a cold source and freezing when they have flowed far enough from their well. From this flow of eitr, vapor rises and turns to rime that builds layer upon layer over Ginnungagap. This rime persists in Ginnungagap even though the region is mild as a windless sky. The flow of eitr is compared, not to water freezing over, but to the cinders flowing from a smelter. As Holtsmark points out, this does not add up—at least not if the terminology relates to water and ice. However, if we consider lava and ash ejected from an eruption, the nature image may become more palatable.
A lava flow from an effusive eruption behaves like water turning to ice. To understand this image, it is useful to consult a scientific description of the behavior of certain eruption types. Geologists Þór Þórðarson and Larsen explain the formation of lava shields in Iceland:
Lava shields are the principal representatives of low-discharge (≤ 300 m3/s) flood lava eruptions. These eruptions produce vast pahoehoe flow fields (up to 20 km3) that are fed by a lava lake residing in the summit crater. The lava cone of each shield is essentially constructed by fountain-fed flows and overspills from the lake, whereas the surrounding lava apron is produced by tube-fed pahoehoe where insulated transport and flow inflation enables great flow length […] The high-discharge (> 1000 m3/s) flood lava events, such as the 1783–1784 A.D. Laki and 934–938 A.D. Eldgjá fissure eruptions, represent some of the greatest spectacles of Icelandic volcanism. (Þór Þórðarson and Larsen 2007: 131)
A volcano’s lava flows from a lake in its newly formed crater and advances upon an area where it quickens. This results in the creation of a hraun, a lava field. The rubbly pāhoehoe caused by pulsating discharges characterizes the most common basalt flow types of Iceland (Þór Þórðarson and Larsen 2007: 131).
Strip the geologists’ description of its scientific discourse and it may be reduced to the following: a liquid substance flows from a reservoir; it quickens and creates a hraun. Moreover, the liquid itself is poisonous (eitr), it flows (at renna) from its source (uppspretta), and it becomes hard like the slags (sindr) from the smelter (eldr). Vapor (úr) rises from the flow, it lands—it freezes (at frjósa)—on top, becoming rime (hrím) that builds in layers.
The term eitr is ambiguous. It refers both to poison and to ice-cold rivers. However, based on what Holtsmark remarks above, it does not seem to refer to the quality of being cold in relation to the eitrár, but rather to the color yellow and a thick flowing substance. Holtsmark herself mentions that the common reference between frozen ice and pus is the color (Holtsmark 1964: 31). Eitr is poisonous and yellow—just like lava. That volcanoes and related phenomena can be poisonous (from gasses) was known to medieval Icelanders. This is attested in Konungs Skuggsjá (ch. 17–21). As noted above, the emission of hydrogen fluoride was the main culprit in the Laki eruption in 1783–1784, and it has also been observed in relation to eruptions in Hekla (Haraldur Sigurðsson et al., 1985: 1003). Saxo, giving a detailed description of the wonders of Iceland, mentions deadly wells there: “Illic etiam fama est pestilentis undę laticem scaturire, quo quis gustato perinde ac ueneno prosternitur” (Gesta Danorum Praefatio, 2,7) (It is also told that there are springs up there with water that is so dangerous that if you taste it you die instantly, as if it were poison). If the eitr is in fact lava, there should be nothing contradictory in the notion of it flowing from a well and freezing up when it flows too far, even though, as Holtsmark notes, it flows into a warmer climate. Neither is there any conflict between the image of the eitr–rivers and the image of the slags flowing from a smelter; they are analogous.
The vapor that freezes and turns to rime did not as such cause Holtsmark any distress, but there is one problem with this image: the rime builds layer upon layer as if it were ice. Rime is an accumulation of already frozen moisture particles in the air, and does not generally build in layers. Ice may do so, and when one interprets this image as a process of an ice formation, this inconsistency is easily overlooked. I will presumptuously suggest that if the author were looking for a frozen watery substance to build in layers across Ginnungagap, the image of snow would have been more suitable. Where rime would start depositing on the surface and keep accreting in a leeward direction, snow builds layers upon layers across any exposed surface. However, snow melts in the warmth, and Holtsmark noted the inconsistency between Ginnungagap being mild as a windless sky and the rime building layers there.
Relying on Árni Hjartarson’s verdict on the contents of Hallmundarkviða, aspects of the poem can be useful in terms of certain words, phrasings, and analogies pertaining to volcanism. As Hallmundarkviða describes an eruption, it provides an early vocabulary that can clarify the meaning of some of the imagery expounded in the creation myth. Árni Magnússon copied Hallmundarkviða in 1686 from the Vatnshyrna manuscript, which perished in the Copenhagen fire in 1728. It is preserved in Bergbúa Þáttr, and consists of twelve stanzas in drottkvætt meter. The language has been dated to either the twelfth or the thirteenth century (Bergbúa Þáttr pp. cciii–ccv). If nothing else, this puts the poem in the same period as Snorra Edda.
By consulting the language of Hallmundarkviða, the image of the rime is no longer an unresolvable inconsistency. Hrím has a secondary meaning, namely “soot”, and the jǫtunn-name Hrímnir can mean “sooty” or “the one that causes soot” (i.e., fire) (Finnur Jónsson 1966: 284–85). Hrímnir appears in Hallmundarkviða in connection with the eruption. The ashes are called mjǫll “newly fallen snow” (Hallmundarkviða st. 11). This confirms the prevalence of an analogy between the ejecta from a volcanic eruption and frozen precipitation. Unlike rime, soot and ash build up, layer upon layer, when ejected in an eruption and unlike snow (and rime), soot and ash persist even though it enters a milder climate: it cools down rather than heating up. In “The Day the Dry Snow Fell: The Record of a 7627-year-old Disaster”, archaeologists Alwynne Beaudoin and Gerald Oetelaar provide some interesting points of comparison with the image of rime, vapor, and eitr (Beaudoin and Oetelaar 2006a). The First Nation peoples of the Alberta and Saskatchewan provinces in Canada relate a tale of the 7,627-year-old cataclysmic Mazama eruption in modern-day Oregon. The Mazama event is remembered as “the day the dry snow fell”; the legend recounts that it fell for days from a darkened sky. It built up in layers, and when you walked in it, dust (ON úr) would rise up and choke you (eitr), and rain would turn the dust into a thick and slimy substance that eventually dried to a crust (sindr, aurr, íss) (Beaudoin and Oetelaar 2006a: 41–43).
This comparison may explain the apparent inconsistencies that Holtsmark noted in the ice and rime, which did not behave as these elements are supposed to. The creation myth in Gylfaginning draws upon images of ice and water because there is nothing in the vocabulary to provide an independent description of a volcanic eruption. In the same way that tephra became “dry snow” to the First Nation peoples in Alberta and Saskatchewan, the ashes of an Icelandic eruption became hrím and úr in Snorri’s Edda and mjǫll in Hallmundarkviða. Such a description is provided as a consequence of the principle of analogy in the vernacular theory of geothermal activity.


Élivágar is a term for a body of ice-cold waters that occurs in a couple of medieval narratives, namely the eddic poem Hymiskviða and the prose story Þorsteins þáttr bœjarmagns. In Hymiskviða, it seems to be synonymous with the sea encircling the world (Finnur Jónsson 1966: 117) and in Þorsteins þáttr bœjarmagns it is an ice-cold river that separates Miðgarðr (the human world) from Jǫtunheimar (the land of giants). Élivágar is also mentioned in Vafþrúðnismál, another eddic poem, and Hallmundarkviða, but whether it is a term for a body of ice-cold water is not clear. We have already seen how the eitr in Snorri’s creation myth may connote lava, rather than water so cold that it is poisonous. In Snorri’s creation myth, it seems natural to assume that both the eitr and its source, the Élivágar, are extremely cold bodies of liquid. However, the association with extreme cold is caused by a series of juxtapositions that can be attributed to the author. According to Snorra Edda, Hvergelmir is situated in Niflheimr, and this is the source of the rivers. In the eddic poem Grímnismál, water flows from the hart Eikþyrnir’s horns into the well of Hvergelmir, and this is the source of all the water in the world (Grímnismál st. 26). The river names that are recounted in Snorri’s creation myth are borrowed from Grímnismál (st. 27–28), but it seems that the connection to Élivágar is only established through the narrative of Gylfaginning. There is no precedence for this connection elsewhere (Simek 2007: 73). One of the rivers flows past the Hel-gates, the deepest part of the underworld. At this point in Snorri’s narrative, it is not mentioned that Niflheimr is cold; in fact the text immediately proceeds to describe the fiery realm of Muspell and its guardian Surtr. After this description, and a citation from Vǫluspá inn skamma, the Élivágar are described as flowing from some place (Gylfaginnning ch. 5). It is not stated directly that the Élivágar come from Hvergelmir, but the phrasing “Ár þær er kallaðar eru Élivágar” (Gylfaginning ch. 5) (These rivers, which are called Élivágar) naturally leads to this assumption on part of the reader. This is most likely the intention of the text, but it is important to note that Snorri has not yet defined Niflheimr as cold, the description of the Élivágar is not precisely pinned to Niflheimr, and most importantly, the Élivágar do not seem to be an original name for the rivers that flow from Hvergelmir. They have individual names in Grímnismál, whereas Vafþrúðnismál tells a different story about Élivágar. It may therefore be possible that the direct association of Élivágar with a cold body of water does not underlie Snorri’s creation myth.
Vafþrúðnismál may facilitate an alternative understanding of Élivágar. It is one of the most prominent sources for Gylfaginning (Gylfaginning p. xxv) and also most probably contains the oldest attestation of the term “Élivágar.” It does not give any clear reason to assume that the Élivágar are cold. In Vafþrúðnismál, the god Óðinn asks the wise giant Vafþrúðnir where the primordial jǫtunn Aurgelmir came from. Vafþrúðnir answers: “Ór Élivágom stucco eitrdropar, svá óx unz varð ór iǫtunn” (Vafþrúðnismál st. 30) (From Élivágar sprang poison drops, so they grew until there came a jǫtunn from them). This stanza is also included in Gylfaginning, but it is interpolated after Snorri announces that Niflheimr is cold. In this manner, the idea that the Élivágar are cold is established. Undoubtedly, the reference to Aurgelmir’s offspring as hrímþursar in stanza 33 plays a role here, but as I noted above, hrím may well refer to soot, not rime or ice. If we stay with the volcanic image of the flowing eitr from Gylfaginning, the image in stanza 30 of Vafþrúðnismál fits well. The eitr in this instance springs from its source, the Élivágar, and build up until they become a figure, Aurgelmir.
Snorri claims that Aurgelmir and Ymir are the same, i.e. the primordial giant. According to Grímnismál (st. 40–41) and Vafþrúðnismál (st. 21), the human world is made from the body of Ymir. Gylfaginning also relates this creation myth but insists (in discord with Vafþrúðnismál) that Aurgelmir and Ymir are the same being. It is possible that this connection between the two figures is established through the relationship of extreme cold. Vafþrúðnismál relates that Ymir was ice-cold: “ins hrímkalda iǫtun[s]” (Vafþrúðnismál st. 21) (the rime-cold giant) (emphasis added). There are strong indications that this association of the two figures is a conscious construction, designed to polarize the jǫtunn race into hot and cold in the creation myth (Clunies Ross 1983: 51). This could, in accordance with Klaus von See’s theory (1988: 52–55), be an attempt to reconcile elements of a seemingly pagan complex of mythos with medieval Neoplatonic theories.
Underneath this attempt, a widespread terminology that can be associated with volcanism can be detected: the first part of the name Aurgelmir is probably aurr- “wet sand, gravel” (Simek 2007: 24). Vafþrúðnismál says that Aurgelmir fathered Þrúðgelmir, who fathered Bergelmir (st. 29). Þrúð- seems to mean “power” (Simek 2007: 329), and the first part of Bergelmir, ber(g)-, may mean “bear” (Finnur Jónsson 1966: 44) or “mountain” (Simek 2007: 34). The last part of these names, -gelmir, means “roarer” (Simek 2007: 24). This word is also associated with wells and rivers such as Hvergelmir (hverr- meaning “kettle-roarer” [Finnur Jónsson 1966: 300]) and Vaðgelmir (Finnur Jónsson 1966: 587). If -gelmir refers to the roaring, churning commotion of the rivers and wells of the underworld, the image of roaring, churning, violent gravel and sand is highly interesting, as it is so readily applicable to a jökulhlaup, a lahar, or other volcanic ejecta. The powerful roars of Þrúðgelmir and Bergelmir, whether the latter means “bear-roarer” or “mountain-roarer”, is similarly easy to apply to volcanism, but the association of these names to volcanism does not rest on this assumption alone.
The jǫtunn Aurnir is a figure in Hallmundarkviða. He appears in stanza 9, where he receives an iron-braced stone boat from the bergbúi “mountain-dweller” who is relating the course of the eruption. In the last stanza, the bergbúi says that Aurnir’s well is not yet dry. In other words, the jǫtunn presides over a well, which seems to be the source of eruptions. Aurnir’s name is probably derived from the same aurr- as Aurgelmir (Simek 2007: 252). [3] This notion coincides with stanza 7, where the bergbúi describes the eruption of lava with the term aurr “clay”: “aurr tekr upp at fœrask undarligr ór grundu” (Hallmundarkviða st. 7) (a strange clay begins to flow from the ground). This word was already known to the tradition from its association with Aurgelmir as the man who was built from the eitr of the Élivágar. The term Élivágar also occurs in Hallmundarkviða. In stanza 7, the underground commotion is described as a fight in Élivágar (bág, í Élivága). In this way, there is a direct association between the Élivágar and volcanism. In Vafþrúðnismál the Élivágar produce a figure whose name is closely associated with a term for lava, and in Hallmundarkviða they are the site of a battle in the course of an eruption. Presumably, they are to be understood as a euphemism for underground lava—just like Aurnir’s well. Both Aurgelmir and Aurnir may then be personifications of lava, and the material in jökulhlaups, on the same terms as Hrímnir, can be a personification of soot and ash, and Surtr can be a personification of volcanic fire.


In Snorri’s version of the creation myth, we find a possible vernacular theory of geothermal activity embedded in the image he paints of nature. The nature-image makes use of references to ice and water, but, as Holtsmark has pointed out, the image is not convincing when treated on the parameters offered by the natural behavior of ice and water. Instead, it makes sense when it is treated with consideration of the observable aspects of the commonly occurring Icelandic low discharge effusive eruptions, where lava gushes out from its source and flows like a raging river until it hardens and adds new layers to the landmass. The term eitr refers to the qualities of the lava as a glowing hot, yellow-red substance that is poisonous (or harmful) in every way. This substance flows from the caldera as a river and becomes “ice”. From the flow there are fumes—vapors—which rise and crystallize like snow and ice, and fall down again to build layers of this strange “ice” in the primordial void. These processes cannot be described in the vernacular vocabulary without analogies to ice and water, and this is why there is confusion about the details of the process. The creature that is produced from this flow of lava is, in the original story in Vafþrúðnismál, Aurgelmir “clay-roarer”. In Vafþrúðnismál, the primeval jǫtunn was originally created from lava, referred to as eitr, but a conscious effort on behalf of the author of Gylfaginning reconfigures the information from the poetry, and rearranges the semantics of the nature-image in the creation myth. Aurgelmir is professed to be the same as Ymir, and inspired by Neoplatonic philosophy, Snorri infers a dichotomy between fire and ice. The creation myth in Gylfaginning looks to be an attempt to align a vernacular theory of geothermal activity—and the creation of the world—with the teachings of continental philosophy of the twelfth and thirteenth centuries.
The tendency to formulate mythic stories and legends with a basis in the theme of geology has been observed among many peoples who live in close proximity to volcanoes. These stories are more or less transparent to individuals who are outsiders to the culture in question. One of the factors that can obscure a narrative layer centered on geological aspects in a myth or legend is vernacular vocabulary. Such vocabulary may be very culture-specific. In the case of Old Norse myths, if one operates under the assumption that these narratives reflect, at least in part, aspects of the worldview of pre-Christian Scandinavians, there is a strong possibility that these myths can be considered entry points into the relationship between early Icelanders and their environment.

Works cited

Primary sources

Annales Regii
Islandske Annaler indtil 1578. Ed. Gustav Storm. Christiania: 1888.
Bergbúa Þáttr
Íslenzk fornrit, 13. Ed. Þórhallur Vilmundarson and Bjarni Vilhjálmsson. Reykjavík: 1991. Pp. 441–50.
Gesta Danorum
Saxo Grammaticus. Gesta Danorum. Ed. Karsten Friis-Jensen. Transl. Peter Zeeberg. 2 vols. Copenhagen: 2005.
Grímnismál: see Poetic Edda
Snorri Sturluson. Edda: Prologue and Gylfaginning. 2nd ed. Ed. Anthony Faulkes. London: 2011.
Hallmundarkviða: see Bergbúa Þáttr
Hymiskviða: see Poetic Edda
Kristni Saga
Kristnisaga, Þáttr Þorvalds ens víðförla, Þáttr Ísleifs biskups Gizurarsonar, Hungrvaka. Ed. Bernhard Kahle. Altnordische Saga-Bibliothek, 11. Halle an der Saale: 1905.
Konungs Skuggsjá
Konungs Skuggsiá. Ed. Ludvig Holm-Olsen. Oslo: 1945.
Íslenzk fornrit, 1. Ed. Jakob Benediktsson. Reykjavík: 1986. Pp. 29-397.
Liber Miraculorum
Herbert von Clairvaux und sein Liber Miraculorum. Die Kurzversion eines anonymen bayrischen redaktors. Ed. Gabriela Kompatscher Gufler. Lateinische Sprache und Literatur des Mittelalters, 39. Berlin: 2005.
Metamorphoses, P. Ovidi Nasonis. Ed. Richard J. Tarrant. New York: 2004.
Poetic Edda
Edda. Die Lieder des Codex Regius nebst verwandten Denkmälern. 3rd ed. Ed. Gustav Neckel and Hans Kuhn. Heidelberg: 1962.
Vafþrúðnismál: see Poetic Edda
Vǫluspá: see Poetic Edda

Secondary sources

Árni Hjartarson. 2014. “Hallmundarkviða, eldforn lysing á eldgósi.” Náttúrufræðingurinn 84 (1–2): 27–37.
Barber, Elizabeth W., and Paul T. Barber. 2004. When They Severed Earth from Sky: How the Human Mind Shapes Myth. Princeton.
Beaudoin, Alwynne B., and Gerald A. Oetelaar. 2006a. “The Day the Dry Snow Fell: The Record of a 7627-year-old Disaster.” In Alberta Formed Alberta Transformed. Ed. Michael Payne, Donald Grant Wetherell, and Catherine Anne Kavanaugh. Edmonton. Pp. 36–53.
———. 2006b. “Darkened Skies and Sparkling Grasses: The Potential Impact of the Mazama Ash Fall on the Northwestern Plains.” Plains Anthropologist 50 (195): 285–305.
Cashman, Kathy, and Sean Cronin. 2008. “Welcoming a Monster to the World: Myths, Oral Tradition, and Modern Societal Response to Volcanic Disasters.” Journal of Volcanology and Geothermal Activity 176: 407–18.
Clunies Ross, Margaret. 1983. “Snorri Sturluson’s Use of the Norse Origin-legend of the Sons of Fornjótr in his Edda.” Arkiv för nordisk filologi 98: 47–66.
Courtillot, Vincent. 2005. “New Evidence for Massive Pollution and Mortality in Europe in 1783–1784 may have Bearing on Global Change and Mass Extinctions.” C. R. Geoscience 337: 635–37.
Dronke, Ursula, and Peter Dronke. 1977. “The Prologue of the Prose Edda: Explorations of a Latin Background.” In Sjötiu Ritgerðir: Festskrift til Jakob Benediktsson. Ed. E. G. Pétursson and J. Kristjánsson. Reykjavík. Pp. 153–76.
Falk, Oren. 2007. “The Vanishing Volcanoes: Fragments of Fourteenth-century Icelandic Folklore.” Folklore 118 (1): 1–22.
Faulkes, Anthony. 1983. “Pagan Sympathy: Attitudes to Heathendom in the Prologue to Snorra Edda.” In Edda: A Collection of Essays. Ed. R. J. Glendinning and H. Bessason. Manitoba. Pp. 283–314.
Fei, Jie, and Jie Zhou. 2006. “The Possible Climatic Impact in China of Iceland’s Eldgjá Eruption Inferred from Historical Sources.” Climatic Change 76: 443–57.
Finnur Jónsson, ed. 1966. Lexicon Poeticum Antiquae Linguae Septentrionalis: Ordbog over det Norsk-Islandske Skjaldesprog. 2nd ed. Copenhagen.
Grattan, John. 1995. “An Amazing and Portentous Summer: Environmental and Social Responses in Britain to the 1783 Eruption of an Iceland Volcano.” The Geographical Journal 161: 125–36.
Grattan, John, and F. Brian Pyatt. 1993. “Acid Damage to Vegetation Following the Laki Fissure Eruption in 1783—an Historical Review.” The Science of the Total Environment 151: 241–47.
Grattan, John, M. Durand, and S. Taylor. 2003. “Illness and Elevated Human Mortality in Europe Coincident with the Laki Fissure Eruption.” In Volcanic Degassing. Ed. Clive Oppenheimer, David M. Pyle, and Jenni Barclay. Geological Society of London Special Publication, 213. London. Pp. 401–14.
Grattan, John, Roland Rabartin, Stephen Self, and Þorvaldur Þórðarson. 2005. “Volcanic Air Pollution and Mortality in France 1783–1784.” Comptes Rendus—Geoscience 337: 641–51.
Gräslund, Bo. 2009. “Ekki nýtr solar. När himlen färgades röd av gudarnas blod.” In vol. 1 of Á austrvega: Saga and East Scandinavia. Preprint papers of the 14th International Saga Conference Uppsala, 9th-15th August 2009. Ed. Agneta Ney, Henrik Williams, and Fredrik Charpentier Ljungqvist, et al. Gävle. Pp. 318–26.
Gräslund, Bo, and Neil Price. 2012. “Twilight of the Gods? The ‘Dust Veil Event’ of AD 536 in Critical Perspective.” Antiquity 86: 428–43.
Guðrún Nordal. 2001. Tools of Literacy. Toronto.
Hamacher, Duane W., and Ray P. Norris. 2009. “Australian Aboriginal Geomythology: Eyewithess Accounts of Cosmic Impacts?” Archaeoastronomy 22 (2009): 62–95.
Haraldur Sigurðsson et al., eds. 1985. Encyclopedia of Volcanoes. New York.
Holtsmark, Anne. 1964. Studier i Snorres Mytologi. Oslo.
Juneja, Monica, and Gerrit J. Schenk. 2014. “Viewing Disasters: Myth, History, Iconography, and Media Across Europe and Asia.” In Disaster as Image: Iconographies and Media Strategies across Europe and Asia. Ed. Monica Juneja and Gerrit J. Schenk. Regensburg. Pp. 7–40.
Nordvig, A. Mathias V. 2013. Of Fire and Water: The Old Norse Mythical Worldview in an Eco-Mythological Perspective. PhD diss., Aarhus University.
Phillpotts, Bertha S. 1905. “Surt.” Arkiv för nordisk filologi 21: 14–30.
Piccardi, Luigi and W. Bruce Masse, eds. 2007. Myth and Geology. Geology Society Special Publications, 273. London.
Reid, Nick, Patrick Nunn, and Margaret Sharpe. 2014. “Indigenous Australian Stories and Sea-level Change.” In Indigenous Languages and their Value to the Community: Proceedings of the 18th Foundation for Endangered Languages Conference. Ed. Patrick Heinrich and Nicholas Ostler. Bath. Pp. 82–87.
von See, Klaus. 1988. Mythos und Theologie im skandinavischen Hochmittelalter. Heidelberg.
Sigurður Nordal. 1927. Völuspá. Copenhagen.
Simek, Rudolf. 2007. Dictionary of Northern Mythology. Suffolk.
Stothers, Richard B. 1998. “Far Reach of the Tenth Century Eldgjá Eruption, Iceland.” Climatic Change 39: 715–26.
Tweed, Fiona S. 2012. “‘Now that the Dust has Settled ...’ the Impacts of Icelandic Volcanic Eruptions.” Geology Today 28 (6): 217–23.
Vasey, Daniel E. 1991. “Population, Agriculture, and Famine: Iceland, 1784–1785.” Human Ecology 19 (3): 323–50.
Vitaliano, Dorothy B. 1973. Legends of the Earth. Bloomington.
de Vries, Jan. 1961. Altnordische etymologisches Wörterbuch. Leiden.
Zdanowicz, C. M., G. A. Zielinski, and M. S. Germani. 1999. “Mount Mazama Eruption: Calendrical Age Verified and Atmospheric Impact Assessed.” Geology 27: 621–24.
Zielinski, Gregory et al. 1995. “Evidence of the Eldgjá (Iceland) Eruption in the GISP2 Greenland Ice Core: Relationship to Eruption Processes and Climatic Conditions in the Tenth Century.” The Holocene 5 (2): 129–40.
Þór Þórðarson. 2010. “Perception of Volcanic Eruptions in Iceland.” In Landscapes and Societies: Selected Case. Ed. I. P. Martini and Ward Chesworth. New York. Pp. 285–96.
Þór Þórðarson and G. Larsen. 2007. “Volcanism in Iceland in Historical Time: Volcano Types, Eruption Styles and Eruptive History.” Journal of Geodynamics 43: 118–52.
Þór Þórðarson and S. Self. 1993. “The Laki (Skaftár Fires) and Grimsvötn Eruptions in 1783–1785.” Bulletin of Volcanology 55: 233–63.
Þorvaldur Þórðarson and Ármann Höskuldsson. 2008. “Postglacial Volcanism in Iceland.” Jökull 58: 197–228.


[ back ] 1. Strangely, Annales Regii does not recount the Eldgjá eruption (see Annales Regii p. 103).
[ back ] 2. Unless otherwise stated, all translations are my own.
[ back ] 3. In fact, there are several jǫtunn names derived from aurr- and berg- (Finnur Jónsson 1966: 24, 44-45).