Vol. 18 (2009) pp.12-22.

Title: Snorri’s Trollwives

Author: William Sayers
Statement of responsibility:
Marked up by
Martin Holmes
Statement of responsibility:
John Tucker University of Victoria
Statement of responsibility:
Book Review Editor/Rédactrice des comptes rendus
Helga Thorson University of Victoria

Marked up to be included in the Scandinavian-Canadian Studies Journal
Source(s): Sayers, William. 2007-2009. Snorri’s Trollwives. Scandinavian-Canadian Studies Journal / Études scandinaves au Canada 18: 12-22.
Text classification:
  • Snorri Sturluson
  • Sturluson, Snorri
  • trolls
  • names
  • JT: started markup 3rd March 2009
  • MDH: entered minor proofing corrections 9th June 2009
  • MDH: entered editor's proofing corrections 2nd October 2009
  • MDH: entered author's proofing corrections 10th October 2009
  • MDH: entered editor's proofing corrections 22nd October 2009

Snorri’s Trollwives

William Sayers

ABSTRACT: The 60 names for female trolls associated with Snorri Sturluson’s Skaldskaparmál are constructed according to specific phono-semantic criteria. Some are represented elsewhere in the literary record, e.g., Grýla; others appear more arbitrary constructs, no less typical. The names suggest conflict, the noise of weapons, darkness, disturbed emotional states, and point to the warrior’s susceptibility to panic and fear. The trollwives are thus potential judges of male courage and competence.
RÉSUMÉ: Les 60 noms de trolls femelles qui se retrouvent dans la tradition manuscrite du Skaldskaparmál de Snorri Sturluson se sont construits sur des critères phono-sémantiques spécifiques. Certains se retrouvent ailleurs dans le corps littéraire, par ex. Grýla; d’autres semblent des assemblages plus arbitraires, toutefois également typiques. Les noms suggèrent le conflict, le vacarme des armes, l’obscurité, des états psychiques troublés. De même, ils reflètent la susceptibilité du guerrier à la panique et à la terreur. Les femmes-trolls sont donc de futurs juges du courage et de la compétence mâles.
At the conclusion of Skaldskaparmál Snorri Sturluson has a versified list of the names of female trolls. It might be thought to lie somewhere between the catalogue of the Æsir and the list of parts of the ship that may be used metonymically to represent the ship as a whole, that is, they may qualify as the “real” names of supernatural beings or, more abstractly, stand as “typical” names for such monstrous supernatural females. A recent article by Thomas D. Hill, “Perchta the Belly Slitter and Án hrísmagi: Laxdæla saga cap. 48-49,” raises a number of pertinent questions that can rewardingly be posed of the list as a whole.
In this saga Án dreams of a horrible woman with a huge knife and trough, who eviscerates him and replaces his entrails with brushwood, hence the nickname hrísmagi [Brushwood-belly] when his dream is not taken seriously. Publics familiar with saga conventions will have known that ominous dreams are just that—omens. Hill pursues analogues to the dream scene, focusing on a figure from German tradition, Perchta or Berhta. The cognate of such a name in Old Norse would be based on bjartr “bright, shining” (and in Old English on beorht), although it should be conceded at the outset that Snorri’s list, as we have it, does not contain it.
The list by Snorri or incorporated in his work, reproduced here in an appendix, comes after comparable lists of the names of legendary sea-kings, the names of—or for—giants, and is followed by a brief list of bynames for Þórr and then the names of the Æsir. These lists are an important part of the skaldic tool kit and are introduced by Snorri’s comments on word-play—homonymity—and the substitution of metonyms or homologues for more common words in poetry. Thus, what the pun achieves on the level of sound, the substitute word does on the semantic level, with mechanics rather similar to the metaphor. In practice, any of a woman’s domestic tasks could be used to characterize her, the part for the whole, while a reference to a giant or giants, or their attributes, environments, such as mountains, could be erected on any giant name, in a substitution not of a part for the whole but of the specific for the general. These practices and Snorri’s lists are then an enormous resource for a poetics in which alliteration, assonance, internal rhyme and the like and, as importantly, allusion, indirection, metaphor are key elements.
We may first consider some formal elements of Snorri’s list of names for female trolls. We note the rough stanzaic form, the prominent alliterative effects in the early part of the list, less strictly maintained in the latter part, the alternation between monosyllables and what at first glance appear compounds, and the tendency to place words of more than two syllables at the end of short runs or at the end of the stanza. Like the individual stanza with its multisyllabic final item, the list is given a certain integrity and quasi-authenticity by the introduction “Skal ek tröllkvinna telja heiti” [I shall list the names of troll-women] and the closure effect of the words “Viljum nefna Rýgi síðarst ok Rifingöflu” [We wish to name last Rygi and Rifingafla]. An alphabetized list of the names yields the following chief frequencies: 13 names starting with G (often followed by N or R), 12 with H, nine with vowels, after which we drop to sets of four and two. No names begin with D, N, P (not too surprising), or T.
The catalogue was a popular medieval subgenre, often lightly narrativized as people are seen engaged in a variety of actions, mariners readying a ship, for example. In medieval English literature Widsith and The Knight’s Tale offer parallels from the geographical and ethnic spheres. Snorri’s list is more austere, and simple juxtaposition or phonetic resemblance is not enough to create a larger context. Chief among the questions that may be raised is whether these names are deeply rooted elements of early Norse belief in the supernatural or whether some, perhaps a majority, may not be what I have called, in the Irish context, “supernatural pseudonyms,” onomastic neologisms—Snorri’s riff on a few widely known names and their characteristic phonological and semantic elements (Sayers 1994).
Of the 60 names, just under half are attested from some other context. This may be, variously, 1) as a name for a weapon (a kind of personal appropriation of the otherwise maleficent supernatural), 2) in another catalogue (possibly in variant form), as a byname for another supernatural being, divinity, or “beast of battle,” or for a cultural phenomenon such as “war,” 3) as the name of a supernatural being killed by a god or hero without additional story-telling particulars, or, in contrast to these passing references, 4) as a sharply contoured figure in a fully realized narrative context. Five names—among which Hrimgerðr, for example—qualify for inclusion in this last category which is, of course, an arbitrary one, established for the present inquiry. Thus, there seems good reason to believe that Snorri or another did not cut the list of names for troll-women out of whole cloth and that both many of the names and, more tentatively, the underlying principles of name formation are authentic parts of the tradition. But just what is authentic here? Any storyteller can coin a plausible name with the right “ring” to it, and this can then enter the community’s greater cultural goods. Think of Tolkien and The Lord of the Rings, particularly of Elvish names with Celtic precedents.
Since none of the stanzas is denser than the others as concerns names attested elsewhere, the first stanza has been selected for closer scrutiny because of its necessary representativity as an introduction to the list as a whole. These first names both test the “public’s” prior knowledge and establish precedents for names to follow. The discussion begins with simple etymologizing, if this is a term applicable to what may be an invented name. Gríðr might be thought to stand in ironic contrast to grið “home, domicile” with its differing vocalism, but there are no other words on this apparent root. With similar initial consonants are greypr “fierce, fearful,” gretta “to frown,” gramr, “angry.” Yet in the following such distant echoes will not be pursued. Nor will there be, in most cases, an attempt to trace these names back to putative Indo-European roots or Germanic and other cognates. The discussion will remain more or less on the level of the medieval public for a poem in which the name might figure. Gnissa prompts thoughts of gnísta “gnash teeth, snarl, biting of frost” and gneista “spark.” Grýla is left for more extensive discussion below. Brýra is close—but perhaps not close enough—to brýna “whet” of weapons and people. Glumra is more satisfactorily matched with glumr “noise, rattle, clatter” and glymr “clatter, clash.” To anticipate one conclusion of this study, this fits the larger picture of female trolls as embodying contention, conflict, the noise of weapons, harsh weather, and—its interiorization—aroused or disturbed emotional states. Geitla suggests only geit “she-goat,” perhaps an ironic inversion. Gríma prompts an association with grimmr “fierce” but the word, as a noun, also meant “mask, cowl, the beak of a ship,” even “night” itself. Bakrauf meant “anus,” back-hole. Guma could seem the converse of gumi “man” but there is also the verb gumsa “to mock,” as well as gumpr “bottom, fundament,” perhaps a tie to the preceding name. This would be a semantic link, as distinct from the phonic or alliterative links. Gestilja could be related to gestr “guest” but in the sense of an unwelcome one. Lastly, Grottintanna suggests overgrown or grinding teeth. In summary, these first names are not simple changes rung on well-known words with negative connotations. It rather seems that it is their “phono-semantic” constituents—all the GRs—that create the associations, but this is a very fuzzy area of speculation. There is both alliteration, Gríðr and Grýla, and a sort of assonance, Grýla and Brýa, yet no lexical and metrical effects comparable to those of formal skaldic verse. Perhaps the rudimentary poetic features were intended only to aid memorization or make for an impressive litany-like effect.
Of these first 11 names (Grýla again excepted) only Gríðr and Gríma are found outside the þulur or lists, but the latter is fairly frequently met in the fornaldarsögur. The evidence reviewed to date is not encouraging as concerns answers to the fundamental question set out above. Hrimgerðr and Grýla will, however, weigh rather heavier in the balance than the other names.
Hrimgerðr figures in one of the Eddic poems about Helgi Hjörvarðsson. This is the Helgi on whom names will not stick, so that his true warrior identity can only be bestowed by a supernatural being, in this case a valkyrie who also shares attributes of a fylgja. Atli is another hero in these poems, and he has a slanging match with the daughter, Hrimgerðr, of a giant, whom Helgi and Atli have just killed. Identities, names, are a first concern. Then the imagery of reciprocal insult turns sexual and equine, with mare, stallion, gelding, and a heart in a horse’s arse, that is, cowardice, all mentioned. Patricia Terry has called this exchange a flyting but it may be contended that the true flyting can occur only between con-specific beings, human peers, as between Guðmundr and Sinfjotli in another Helgi poem. More on this below but first a consideration of the name Grýla.
The Grýla figure has been extensively studied by Terry Gunnell, from medieval references down to the folklore of the Faroes, Orkeny, and Shetland, where mummers dressed in straw outfits and called grøleks still go trick or treating. What is most telling, from the present perspective, is that in the putatively realistic story-telling environment of Sturlunga saga, Grýla figures in direct speech and in a simile that the public within the saga and without would be expected to understand. In Íslendinga saga the attacker on a farm-house likens himself to Grýla coming down into the fields with fifteen tails on her back. How consciously is the name Grýla positioned in Snorri’s list, not first but the alliterating head word in the second verse? Or is it just coincidence that the name about which we have most information should have a position of prominence in the þula?
Grýla, as well as now being a more fully sketched figure, also supports some interesting etymological sleuthing. The name is associated with English gruesome and can be traced to the Indo-European root *ghreu-d- “to affect forcibly a mental or emotional state.” Its Old English reflex is gryre, used 14 times in the Beowulf poem (nine of these in compounds) but always (with one exception) in reference to non-conspecific adversaries, that is, the unhuman beings faced by the human hero. To the set of troll names with initial GR- should naturally be added Grendel. Modern scholarship has advanced a large number of etymologies for Grendel but this perspective—in consideration of the subtle, subliminal effects of words from specific phono-semantic clusters—may be reversed in order to point up in how many different ways and on how many different levels critics find the name Grendel appropriate and satisfying. Here it is appropriate to recall a few other troll-women for their affinities with Grendel’s mother and her weaponry. From Snorri’s list we note Harðgreip, Járnglumra, Járnsaxa, Járnviðja, and other names suggestive of a veiled, hidden face or appearance. Also worth recalling is Grettir’s encounter with the troll woman and her single-edged blade in the Sandhaugar episode.
The non-conspecifity of the supernatural opponent results in a potential for a terror, gryre in Old English, that is more fundamental than the fear of a human opponent with some possible martial advantage. Comparable Irish evidence for this bowel-loosening terror prompts the descriptor “eviscerating fear.” Points and means of attack are not heroic blows to the head, chest or arms but rather demeaning injuries to the lower body and to male sexuality—Dumézil’s third function. Supernatural female opponents with domestic utensils instead of conventional weapons, knife and trough in Án’s case, pose several problems for the hero. First is fear of the alien species, second is fear of the alien gender—the threat of both sexual and martial deficiency—third might be the no-win situation, in which the hero is shamed if defeated by a female with a kitchen knife as used to butcher animals but scarcely gains in honor if he is victorious. Recalling Grýla’s descendants among the Shetland mummers, the warrior is the potential victim of the trick but stands to gain little by a treat. Allied scenes are Þórr’s encounters with giantesses (in which the treat may be taken first), the scenes with Atli, Helgi, and Hrimgerðr, the washerwoman who eggs on Hrafnkel. In fact, all incitation or whetting by women in the sagas may be put under this rubric and likened to Otherworld testing scenes, of which Irish tradition offers a good example in Cú Chulainn’s relations with the Morrigú. It is striking how often women’s work and words are the backdrop to violent male action, e.g., Guðrún doing laundry when Bolli kills Kjartan, the very event that Án’s dream was meant to warn against (Sayers 1990).
The present discussion can only allude to one of the great medieval debates about language: whether names are arbitrarily composed and assigned to things or whether they do indeed reflect, however obscurely, the essential truth of the object designated. Here one might quote, almost at random, Snorri’s near-contemporary Thomas Aquinas’s commentary on Aristotle’s Peri hermeneias:
Some, however, have said that although nouns do not signify naturally in so far as their signification is not from nature, as Aristotle says here, still they do signify naturally in so far as their signification accords with the natures of the things, as Plato said. Nor is it any impediment that one thing is signified by many names, for there can be many copies [similitudines] of one thing, and similarly many diverse names can be imposed on one thing in accordance with diverse properties.
As Isidore notes, names are essential for understanding. Knowledge of numerous names for troll-women had clear advantages for the medieval Norse poet. Such knowledge, since possession of a name means power, could also have had a prophylactic function for the wider public. Possession of a name was like having the answer to a riddle before it was posed. Even a negative litany might have its purposes and knowing the names of trolls could have been one way to keep them at bay. This said and to answer the earlier question as to the authenticity of the list, the attempt to establish a prosopography of early medieval troll-women seems a wrongly designed project. All the names are authentic because, whatever their historical depth or frequency, all are realized on a narrowly defined set of semantic concepts and on the preferential use of certain phonic effects. It would be of interest to conduct some onomastic trigonometry, to look at sets with names of the same sex but different function, such as the valkyries, and sets dealing with the other sex but at comparable distance from humanity, such as giants and dwarves. A quick mental review of the names of the male Æsir immediately shows how fundamentally they differ from the typical constituents of troll-woman names: Óðinn, Baldr, Bragi, Forseti, Heimdall, Höðr, Loki, Njörðr, Þórr, Týr, Ullr, Váli or Áli, Viðarr, Yngvi-Freyr. Not an initial GR- or GN- among them. But this is true of what we consider principal names only. The many bynames and epithets that we also find in Snorri could emphasize a variety of facets of the divine “personality,” including more ominous aspects, and then call on other phonetic resources.
German Perchta, as noted, has been interpreted as meaning “bright” and the name is now associated with spring festivals, the return of light. Toward the close of his article, Hill writes: “Whether there are narratives in Icelandic or other relevant Scandinavian folk traditions about a witch or troll woman of that name [i.e., Perchta] is a question which I cannot answer. But, if I am unable to answer this questions, at least my interpretation of Án’s dream suggests promising new lines of inquiry” (523). The þula reviewed above has not yielded a comparable name among the troll-women, although the reversal of brightness is evident in Gríma as “night.” More importantly, perhaps, Grýla and her congeners, like the Perchta of Germanic tradition, must be recognized as a judgmental figure, just as women, real and supernatural, so often are in Old Norse-Icelandic literature. Does the hero have intestinal fortitude, have guts? Who is qualified, or perhaps existentially called on, to judge? The troll-woman can bestow ris eller rosor, as the Swedes put it, brushwood or roses, or on another level, inspire bottomless terror or heightened courage. The number of incidents like that of Án and his kona ... óþekkilig, “horrible woman,” is relatively limited in our extant texts, but women of that ilk and their names are myriad—one for every hero.


  1. Snorri Sturluson 1998, Skáldskaparmál, in Edda, ed. Faulkes, I.112-13. Snorri’s term is trollkvinna. I have used “trollwife” in my title to evoke its ambiguous resonances (human or non-human, spouse, sexual partner or opponent?), but in the following use both “female trolls” and “troll-women.” This note originated in a paper presented at Norsestock: Cornell Annual Conference on Medieval Icelandic Studies, Ithaca, May, 2008, and, as part of my rhetorical strategy, I have retained its brevity and pace. I am, however, grateful for the astute observations of an editorial reader of the written version.
  2. For a fuller discussion of such bynames, epithets, metonyms, etc. see Sayers 1998.
  3. Hill 2007; Laxdæla saga 1934, Ch. 48-49.
  4. Skáldskaparmál, I.109. On word play and the much discussed term ofljóst “obvious, too clear,” see Sayers 2006.
  5. Important recent works on Old Norse-Icelandic poetics are Clunies Ross 2005, Nordal 2001, and Skaldsagas 2001, ed. Poole.
  6. Catalogues qualify as the object of study of nanophilology, devoted to shorter genres. See Howe’s early recognition (1985) of a subgenre important in the Middle Ages.
  7. This conclusion is based on Anthony Faulkes’s notes in his “Index of Names,” in Edda, II.443-528.
  8. Relevant lexicographical reference works here are An Icelandic-English Dictionary 1957 and Norrøn Ordbook 1990.
  9. These conclusions are confirmed by an examination of the remaining stanzas, the coincident detail of which is here suppressed for reasons of space. The present study is admittedly synchronic, based on a modern edition rather than a manuscript. For the important stratigraphical dimension of individual þulur, see studies by Elena A. Gurevich, e.g., the representative “Skaldische Synonymik” (1992).
  10. “Helgaqviða Hjörvarðzsonar,” in Edda 1962, ed. Kuhn, 140-49.
  11. Terry 1990, 113; for Guðmundr and Sinfjotli, see “Helgaqviða Hundingsbana in fyrri,” Edda 1962, ed. Kuhn, 130-39.
  12. Gunnell 1995, 82, 112, et passim; a mumming tradition is also associated with the Perchta figure discussed by Hill.
  13. Íslendinga saga, in Sturlunga saga 1981, Ch. 7; discussed in Gunnell, 161-62.
  14. Indogermanisches etymologisches Wörterbuch 1959, I.461.
  15. Fuller discussion with bibliography of secondary literature in Sayers 2003 and 2007b; subsequent further attention to the issues of onomastics and origins in Cardew 2005 and Slade 2007.
  16. Grettissaga 1936, Ch. 65.
  17. On the last-named, see Sayers 2007a.
  18. Thomas Aquinas 1964, as cited in translation in Fyler 2007, 87.
  19. In this respect, Hale 1983 may still be consulted with profit for the methodology employed.


Skal ek tröllkvinna
telja heiti:*
Gríðr ok Gnissa
Grýla Brýja
Glumra Geitla
Gríma ok Bakrauf
Guma Gestilja
Gjálp Hyrrokkin
Gneip ok Gnepja
Geysa Hála
Hörn ok Hrúga
Harðgreip Forað
Hrygða Hveðra
ok Hölgabrúð  
Hrímgerðr Hæra
Herkja Fála
Imð Járnsaxa
Íma Fjölvör
Mörn Íviðja
Ámgerðr Simul
Sívör Skríkja
ok Járnglumra
Ímgerðr Áma
ok Járnviðja
Margerðr Atla
Leikn Munnharpa
ok Munnriða  
Leirvör Ljóta
ok Loðinfingra
Kráka Varðrún
ok Kjallandi
Vígglöð Purbörð
Viljum nefna
Rýgi síðarst
ok Rifingöflu.  
* Names in bold are discussed in the essay.


  • Cardew, Philip. 2005. Grendel: Bordering the Human. The Shadow-Walkers: Jacob Grimm’s Mythology of the Monstrous. Ed. Tom Shippey. Tempe, AZ, and Turnhout: Arizona Center for Medieval and Renaissance Studies and Brepols. 189-205.
  • Clunies Ross, Margaret. 2005. A History of Old Norse Poetry and Poetics. Woodbridge: Boydell and Brewer.
  • Edda: die Lieder des Codex Regius. 1962. Ed. Hans Kuhn. 4th edition. Heidelberg: Carl Winter.
  • Fyler, John M. 2007. Language, and the Declining World in Chaucer, Dante, and Jean de Meun. New York and Cambridge: Cambridge University Press.
  • Grettis saga Ásmundarsonar. 1936. Ed. Guðni Jónsson. Íslenzk fornrit 7. Reykjavíkv: Hið íslenzka fornritafélag.
  • Gunnell, Terry. 1995. The Origins of Drama in Scandinavia. Woodbridge: Boydell and Brewer.
  • Gurevich, Elena A. 1992. Skaldische Synonymik und ihre Interpretation in der frühen isländischen gelehrten Poetik (Über eine Systematisierungs­möglichkeit der heiti in den Þulur. Sowjetische Skandinavistik. Eine Anthologie. Ed. Lilja Popowa. Frankfurt: Peter Lang. 15-30.
  • Hale, Christopher. 1983. The River Names in Grímnismál. Edda: A Collection of Essays. Eds. Robert J. Glendinning and Haraldur Bessason. Winnipeg: University of Manitoba Press. 165-86.
  • Hill, Thomas D. 2007. Perchta the Belly-Slitter and Án hrísmagi: Laxdæla saga cap. 48-49. Journal of English and Germanic Philology 106: 516-23.
  • Howe, Nicholas. 1985. The Old English Catalogue Poems. Copenhagen: Rosenkilde and Bagger.
  • An Icelandic English Dictionary. 1969. Eds. Richard Cleasby Gudbrand Vigfusson and William A. Craigie. 2nd ed. London: Clarendon Press.
  • Indogermanisches etymologisches Wörterbuch. 1959-69. Ed. Julius Pokorny. 2 vols. Bern: A. Francke.
  • Íslendinga saga. 1981. Ed. Guðni Jónsson. Vol. 2 in Sturlunga saga. 3 vols. Reykjavík: Íslendingasagaútgáfan, Haukadalsútgáfan.
  • Laxdæla saga. 1934. Ed. Einar Ól. Sveinsson. Íslenzk fornrit 5. Reykjavík: Hið íslenzka fornritafélag.
  • Nordal, Guðrún. 2001. Tools of Literacy: The Role of Skaldic Verse in Icelandic Textual Culture of the Twelfth and Thirteenth Centuries. Toronto and Buffalo: University of Toronto Press.
  • Norrøn Ordbook. 1993. Eds. Leiv Heggestad Finn Hødnebø and Erik Simensen. Oslo: Det Norske Samlaget.
  • Poems of the Elder Edda. 1990. Trans. Patricia Ann Terry. Rev. ed. Philadelphia: University of Pennsylvania Press.
  • Sayers, William. 1990. Women’s Work and Words: Setting the Stage for Strife in Medieval Irish and Icelandic Narrative. Mankind Quarterly 31: 59-86.
  • ⸻. 1994. Supernatural Pseudonyms. Emania 12: 49-60.
  • ⸻. 1998. The Ship heiti in Snorri’s Skáldskaparmál. Scripta Islandica 49: 45-86.
  • ⸻. 2003. Grendel’s Mother, Icelandic Grýla, and Irish Nechta Scéne: Eviscerating Fear. Proceedings of the Harvard Celtic Colloquium 16 and 17 (1996-97). Ed. John T. Koch. Andover, MA, and Aberystwyth: Celtic Studies Publications. 256-68.
  • ⸻. 2006. Onomastic Paronomasia in Old Norse-Icelandic: Technique, Context, and Parallels. TijdSchrift voor Skandinavistiek 27: 91-127.
  • ⸻. 2007a. Ethics or Pragmatics—Fate or Chance—Heathen, Christian, or Godless World? (Hrafnkels saga). Scandinavian Studies 79: 385-404.
  • ⸻. 2007b. Grendel’s Mother (Beowulf) and the Celtic Goddess of Territorial Sovereignty. Journal of Indo-European Studies 35: 31-52.
  • Skaldsagas: Text, Vocation, and Desire in the Icelandic Sagas of Poets. 2001. Ed. Russell Poole. Berlin: de Gruyter.
  • Slade, Benjamin. 2007. Untydras Ealle: Grendel, Cain, and Vrtra. Indo-European Sruti and Christian Smrti in Beowulf. Geardagum: Essays on Old and Middle English 27: 1-32.
  • Sturluson, Snorri. 1998. Skáldskaparmál. Ed. Anthony Faulkes. 2 vols. London: Viking Society for Northern Research.
  • Thomas Aquinas. 1964. In Aristotelis Libros Peri Hermeneias et Posteriorum Analyticorum Expositio. Ed. Raimondo M. Spiazzi. 2nd ed. Turin: Marietti.