Hermann, Pernille, Stephen A. Mitchell, and Jens Peter Schjødt, eds., with Amber J. Rose. 2017. Old Norse Mythology—Comparative Perspectives. Milman Parry Collection of Oral Literature 3. Cambridge, MA: Milman Parry Collection of Oral Literature. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_HermannP_etal_eds.Old_Norse_Mythology.2017.
Framing the Hero: Medium and Metalepsis in Old Norse Heroic Narrative 
Figure 1. Lärbro Stora Hammars I, Gotland, mid-Viking Age, 3.12 x 1.43 m. Image courtesy of Stephen Mitchell.
Figure 2. Detail of sword-pommel from Stora Ihre grave 174, Gotland, second half of the eighth century. Length: 9.5 cm. Image courtesy of the Swedish National Heritage Board.
Figure 3. Detail of Hylestad church portal, Setesdal, first half of the thirteenth century. 2.2 x 0. 52 m. Image courtesy of the Museum of Cultural History, University of Oslo.
Framing in the runestone medium
Figure 4. Krogsta (U 1125), Uppland, sixth century. 1.7 x 0.85 m. Image courtesy of the Swedish National Heritage Board.
Figure 5. Hunnestad 3 (DR 284), Scania, late tenth-early eleventh century. 1.79 x 1.06 m. Image courtesy of Stephen Mitchell.
Figure 6. Jelling 2 (DR 42), southern Denmark, late tenth century. 2.43 x 2.90 m (inscription side). Image courtesy of the National Museum of Denmark.
Figure 7. Stora Runhällen (U 1164), Uppland, eleventh century. 1.67 x 1.6 m. Image courtesy of the Swedish National Heritage Board.
Figure 8. Prästgården (U 855), Uppland, eleventh century. 2.57 x 2.12 m. Image courtesy of Stephen Mitchell.
Figure 9. Ramsund rock engraving (Sö 101), Södermanland, Sweden, eleventh century. 4.7 x 1.8 m. Image courtesy of the Swedish National Heritage Board.
Pictures of Sigurðr
Figure 10. Gök rock engraving (Sö 327), Södermanland, Sweden, eleventh century. 2.5 x 1.65 m. Image courtesy of the Swedish National Heritage Board.
Similar echoes of a traditional association between treasure, the world of gods and heroes and the numinous realms of the afterlife may have persisted even into the Christian commemorative practices at Ramsund and Gök (cf. Zachrisson 1998, who proposes just such a rationale for buried deposits of gold in Viking Age Uppland).
Figure 11. Lellinge Kohave B bracteate (IK 105), Seeland, fifth–sixth century. Diameter: 2.79 cm. Image courtesy of the National Museum of Denmark.
Breaking the frame: metalepsis in textual and visual media
Antique metalepsis acts, then, as Eisen and Möllendorf see it, to enhance the authority, plausibility and effectiveness of textualised traditional materials in the setting of vocality (Schaefer 1992). It supports textual discourse rather than calling it into question, as its modern cousin does. Further:
The consistency with which this distinction is adhered to is striking. Guðrúnarhvǫt is the sole instance of a mixture of the two framing options, as it is introduced by an inquit but does not begin with direct speech from one of the poem’s protagonists. Rather, the poem’s narrator speaks:
trauð mál, talið af trega stórom,
er harðhuguð hvatti at vígi
grimmom orðom Guðrún sono. (Edda p. 264)
(Then I heard quarrelling of the most ill-fated sort,
faltering words uttered out of great grief,
when the fierce-spirited Gudrun whetted for the fight,
with grim words, her sons. (Larrington p. 226))
The opening stanza of Guðrúnarhvǫt, with its emphatic first-person pronoun and multiple verbs of speaking, seems to have triggered the inquit type of framing. This left the compiler no choice but to make Guðrún the subject of the inquit. The medial signature (or, if you prefer, controlling fiction) of the collection is that of authentic, anonymous oral transmission from ancient times. This forbids both the Regius’ narrator from stepping into the limelight as the poem’s originating instance (þá kvað ec “then I spoke” is thus not possible),  and the use of the kind of stereotyped introductory phrases ubiquitous in the skaldic corpus (sem skáldit kvað, “as the skald said”). The framing prose is, then, an important site of Codex Regius’ “medium theory”.  The frame is also the place where one of the main stumbling-blocks to the collection’s “medieval cyclic impulse” (Clover 1982: 59) is negotiated, namely the existence of conflicting variants of heroic narratives such as the death of Sigurðr.  Variants become problematic—or even visible as such—as a consequence of the “remediation” (Grusin and Bolter 1999) of oral poetry in a medium, the written codex, which adds a lasting, material dimension to the oral poem’s temporal, performance-bound mode of existence. It is this process that yields an emphatic sense of sjá kviða (this poem), and the possibility (or necessity) of evaluative commentary (cf. note 22). Where previous researchers have seen in the inclusion of multiple versions of events witnesses to the compiler’s completist or antiquarian tendencies, they become on a narratological reading reckonings with the poem’s historicity. A good example is the text headed “Frá dauða Sigurðar”:
As Andreas Heusler wrote, “Die Heldenfabeln geben sich als Geschichte” (the heroic narratives purport to be history) (1941: 162; my emphasis). Historicity is becoming hard for the Regius narrator to maintain in the face of the variability of the tradition, even though, paradoxically enough, a sense that the heroic poems were historical documents, transmitting genealogically relevant information, was probably a significant motivation for the preservation of the variants in the first place (Rowe 2006). And it is a small step from a conception of the traditional material as a body of meaningful variation, to an understanding of it as fictional, in the modern sense of “a representation portraying an imaginary/invented universe or world” (Schaeffer 2013), whose events unfold in a manner determined by literary rather than referential patterns and constraints. Once the convention that the narrator is reporting events that have happened has been broken with, the heroic material can be developed in ways that exploit narrative contingency.
Figure 12. Kirk Andreas cross fragments: left, Thorwald’s cross (inventory nr. 128) and, right, Sigurd’s cross (inventory nr. 121), Isle of Man, tenth century. Length: 35 cm (128), 68 cm (121). Images courtesy of Manx National Heritage.