Network analysis of the manuscript context of Old Icelandic literature

paper, specified "short paper"
  1. 1. Katarzyna Anna Kapitan

    University of Copenhagen

Work text
This plain text was ingested for the purpose of full-text search, not to preserve original formatting or readability. For the most complete copy, refer to the original conference program.


This paper explores the possibilities of applying computer-assisted methods to the field of Nordic manuscript studies, with a special emphasis on a network analysis—in a broad sense—of manuscript context. A case study of one Icelandic legendary saga's manuscript tradition is used to test the hypothesis that the manuscript context can carry information about ethnic genre associations of the text (on ethnic genre in Old Norse literature see: Harris, 1975; on legendary sagas as a genre see: Quinn, 2006).

Research Questions

Hromundar saga Gripssonar traditionally belongs in the corpus of legendary sagas (fornaldarsogur); it was included in the second volume of Rafn's (1829) Fornaldarsogur Nordrlanda, and in Bjorner's (1737) Nordiska kampa dater. The saga as it is known today, however, is a post-medieval re-working of a metrical version of the story known as rlmur (Brown, 1946), and is probably not much older than seventeenth century. Therefore it does not necessarily fit well with the other texts included the corpus of legendary sagas, as they usually date from the fourteenth and fifteenth centuries (Driscoll, 2005:207). This makes Hromundar saga Gripssonar an interesting case study for investigations of the text's genre affiliation in the extant manuscripts preserving the saga. Does it appear frequently in manuscripts with the older legendary sagas or with younger rimur-based narratives? To answer this question, I first examine the manuscript context of legendary sagas as a corpus, based on collaborative research with Rowbotham and Wills (Kapitan et al., 2017); second, I examine the position of Hromundar saga Gripssonar within the corpus and its relationships with other texts.

Much discussion in the field of Old Norse studies centers on whether the legendary sagas deserve to be considered a separate literary genre, or should instead be analyzed as chivalric literature (Quinn, 2006). One of the main reasons for these considerations seems to be the fact that the term fornaldarsogur is not attested in the medieval texts; it was introduced in the early nineteenth century by C.C. Rafn, who published a collection of texts under the title Fornaldarsogur Nordrlanda (Rafn, 1829). Rafn's selection of texts and the definition of fornaldarsogur as a corpus of texts dealing with events taking place in Scandinavia before the settlement of Iceland, however, was not detached from pervious scholarship of early eighteenth century (Lavender, 2015). The current discussion on the legendary sagas as a corpus (or a genre) is polarized around contradicting opinions. Some scholars suggest that the legendary sagas had to be considered a separate category in pre-modern period, because they are frequently bound together in the manuscripts (Gudmundsdottir, 2001:cxlvii; Mitchell, 1991:21) while others, using the same argument, emphasize strong connections between the legendary sagas and the chivalric sagas (Driscoll, 2005:193). Additional problems arise when classifying the generic hybrids (Rowe 1993; 2004) appearing within the corpus, or distinguishing between the legendary sagas (fornaldarsogur) and the late legendary sagas (fornaldarsogur sldari tlma; Driscoll 2005). Even though scholars eagerly turn towards the manuscript context to support their claims regarding the genre classification, no comprehensive overview of the legendary sagas' codicological context has yet been presented. This gave rise to the project Stories for all times conducted at the University of Copenhagen, which created a complete catalogue of manuscripts preserving legendary sagas. The catalogue contains 818 TEI-conformant XML-based manuscript descriptions with over 8000 items, 1764 of which are classified as legendary sagas and 920 as chivalric sagas. This amount of data is much too large to be analyzed manually, therefore it is necessary to apply computer-assisted analysis in order to draw some general conclusions regarding this corpus and the relationships between these texts.


The first part of my paper, which aims to establish the position of Hromundar saga Gripssonar within the wider context of the manuscript, draws on the network analysis of the corpus, conducted in collaboration with Rowbotham and Wills (Kapitan et al., 2017). There, the codicological context of a text was considered as a system of relationships between texts, and following Hall's (2013) approach in his network of chivalric sagas, texts were represented as nodes, manuscripts as edges, both visualized with the free visualization software Gephi. The second part is based on database queries aimed at obtaining detailed information about particular manuscripts and their contents. The main focus of the analysis was to examine the manuscripts preserving the complete texts of Hromundar saga Gripssonar in Icelandic, therefore the manuscripts containing excerpts and translations were ignored. The distribution of texts appearing frequently alongside Hromundar saga Gripssonar by century has been obtained using XPath queries of the online catalogue Stories for all times. Main Findings

As a result of this research, the hypothesis can be confirmed: generally, texts belonging to one genre appear most frequently in manuscripts with other texts belonging to the same genre. However, an interesting transmission history of Hromundar saga Gripssonar suggests a close association of this saga with the late legendary sagas, and in particular Bragda-Olvis saga. Both Bragda-Olvis saga and Hromundar saga Gripssonar are post-medieval reworkings of older metric versions of the stories (rimur), and for both texts the manuscript AM 601 b 4to (Arni Magnuisson Institute, Reykjavik) was suggested as the witness carrying the best text of the saga (Andrews, 1911; Brown, 1946; Hooper, 1934; Hooper, 1932). Even though Hromundar saga Gripssonar appears most frequently with texts classified as late legendary sagas in pre-1800 manuscripts, after 1800 the texts classified as (traditional) fornaldarsogur start to dominate. The late Bragda-Olvis saga dominates the pre-1800 setting, but the distribution changes in the nineteenth century when Porsteins saga Vikingssonar, Starkadar saga gamla, Fridpjofs saga ins fr&kna, and Halfs saga Halfsreka appear more frequently (as presented on figure below). Starkadar saga gamla is a late-eighteenth century saga written by Snorri Bjornson (1710-1803), utilizing traditional legendary motifs of Saxo's Gesta Danorum and legendary sagas (Driscoll, 2009:209; Simek and Hermann Palsson, 1987:331), so its co-occurrence with other legendary sagas starting from the eighteenth century onwards is not surprising. The three remaining texts that started to appear more frequently alongside Hromundar saga Gripssonar in nineteenth-century manuscripts were all published in the same volume of Rafn's Fornaldarsogur Nordrlanda, in which Hromundar saga Gripssonar was published (volume II); likewise Fridtjofs saga ins fr^kna, and Halfs saga Halfsreka appeared in Bjorner's edition from 1737 together with Hromundar saga Gripssonar. This shows how printed editions influenced the saga's transmission in the manuscript form. A text, which once showed strong connections to another rimurbased narrative, became detached from its previous setting and gained new, print-influenced context after becoming part of printed editions.

Figure 1. Texts appearing frequently with Hromundar saga Gripssonar in manuscripts by century


The topic of this paper fits in the advertised panel "Quantitative stylistics and philology, including big data and text mining studies," as it employs database quarrying and network analysis of significant amount of data.


The author of this paper owes a great debt to Tarrin Wills for merging the data from various databases (,,, to obtain the input data for the network analysis, and to Tim Rowbotham, who worked on standardizing the uniform titles and genre classifications in the XML-files in the Stories for all times catalogue.


Gudmundsdottir, A. (2001). Ulfhams Saga. (Stofnun Arna

Magnussonar A Islandi 53). Reykjavik: Stofnun Arna

Magnussonar a Islandi.

Andrews, A. L. (1911). Studies in the fornaldarsogur

Nordurlanda. Modern Philology, 8: 527-44.

Bjorner, E. J. (1737). Nordiska kampa dater i en sagoflock samlade om forna kongar och hjaltar. Volumen historicum, continens variorum in orbe hyperboreo antiquo regum, heroum et pugilum res praeclare et mirabiliter gestas. Accessit, praeter conspectum genealogicum Svethicorum regum et reginarum accuratissimum etiam praefatio. Stockholmiae: typis J.L., Horrn. (accessed 28 January 2016).

Brown, U. (1946). The saga of Hromund Gripsson and Porgilssaga. Saga-Book, 13: 51-77.

Driscoll, M. J. (2005). Late prose fiction (lygisogur). A Companion to Old Norse-Icelandic Literature and Culture. Oxford: Blackwell Publishing Ltd, pp. 190-204.

Driscoll, M. J. (2009). Editing the Fornaldarsogur NorSurlanda. A Austrvega, Saga and East Scandinavia, Preprints of the 14th International Saga Conference Uppsala 9th - 15th August 2009. Gavle: University of Gavle, pp. 207-12.

Gephi. (n.d.) Gephi: The open graph viz platform:

Hall, A. and Parsons, K. (2013). Making stemmas with small samples, and digital approaches to publishing them: testing the stemma of KonraSs saga keisarasonar. Digital Medievalist, 9.

Handrit. (n.d.) Online catalogue

Harris, J. (1975). Genre in the saga literature: A Squib. Scandinavian Studies, 47(4): 427-36.

Hooper, A. G. (1932). “BragSa-Qlvis saga” now first edited. Leeds Studies in English, 1: 42-54.

Hooper, A. G. (1934). Hromundar saga Gripssonar and the Griplur. Leeds Studies in English, 3: 51-56.

Kapitan, K. A., Rowbotham, T. and Wills, T. (2017). Visualising genre relationships in Icelandic manuscripts. Conference Abstracts. Gothenburg: The University of Gothenburg, pp. 59-62.

Lavender, P. (2015). The Secret Prehistory of the Fornaldarsogur. The Journal of English and Germanic Philology, 114(4): 526-51.

Mitchell, S. (1991). Heroic Sagas and Ballads. Ithaca and London: Cornell University Press.

Nordisk Forskningsinstitut (n.d). The “Stories for all time” Project.

Quinn, J. (2006). Interrogating Genre in the Fornaldarsogur: Round-Table Discussion’. Viking and Medieval

Scandinavia, 2: 276-96.

Rafn, C. C. (1829). Fornaldarsögur Nordrlanda. Kaupmannahofn.

Rowe, E. A. (1993). Generic Hybrids: Norwegian ‘Family’ Sagas and Icelandic 'Mythic-Heroic' Sagas. Scandinavian Studies, 65(4): 539-54.

Rowe, E. A. (2004). 'Porsteins ^attr uxafots, Helga ^attr Porissonar,’ and the Conversion '^iettir'. Scandinavian Studies, 76(4): 459-74.

Simek, R. and Palsson, H. (1987). Lexikon Der Altnordischen Literatur, Die Mittelalterliche Literatur Norwegens Und Islands. (Kroners Taschenausgabe 490). Stuttgart: Kroner.

Skaldic Project Academy Body. (n.d.) Skaldic project:

University of Copenhagen (n.d.) Ordbog over det norr0ne prosasprog Registre:

If this content appears in violation of your intellectual property rights, or you see errors or omissions, please reach out to Scott B. Weingart to discuss removing or amending the materials.

Conference Info


ADHO - 2017

Hosted at McGill University, Université de Montréal

Montréal, Canada

Aug. 8, 2017 - Aug. 11, 2017

438 works by 962 authors indexed

Series: ADHO (12)

Organizers: ADHO