Hrólfs saga kraka: An electronic edition of AM 285 4to

Although relatively young, AM 285 4to is one of the earliest surviving manuscripts of Hrólfs saga kraka. Its 35 paper leaves were written in a form of hybrida script by Brynjólfur Jónsson from Efstaland (†1664), a scribe employed by Þorlákur Skúlason (1597-1656), Bishop of Hólar. According to the colophon, the text of Hrólfs saga was completed on 31 January 1654. Typical of 17th-century Icelandic manuscripts, AM 285 4to is a modest production with no decoration.

The saga text was formerly part of a larger codex acquired by Árni Magnússon in 1702. He subsequently divided it into at least four separate parts, three of which survive and are now called AM 342 4to, AM 285 4to and AM 587c 4to. According to a note by Árni in AM 342 4to, the last part of the codex once contained Sálus saga ok Nikanórs and possibly some lygisögur. The beginning and end of Hrólfs saga kraka ended up in the other manuscripts during the partitioning process, so Árni added two leaves (fols. 1 and 35) to AM 285 4to on which he supplied the missing text. The original first and last lines of the saga are still preserved, albeit heavily blacked out, in AM 342 4to and 587c 4to.

Hrólfs saga kraka has been one of the more popular fornaldarsögur and is preserved in its entirety in at least 40 manuscripts and fragmentarily in a further 11. The saga is thought to have been composed some time during the 14th or 15th century, although the oldest extant manuscripts stem from the 17th century, including AM 285 4to and AM 9 fol. The version of Hrólfs saga kraka preserved in the former has been described as more literary than that in the latter owing to its hypotactic syntax, polished language and archaic expressions. Even though AM 285 4to is the exemplar from which most subsequent Icelandic manuscripts derive, it has hitherto only been used as the base text for one edition: Desmond Slay, Hrólfs saga kraka, Editiones Arnamagnæanæ, Series B 1 (København, 1960). All other full text editions are based on AM 9 fol. (see the complete list in our project bibliography here) with the exception of Eric Julius Biörner’s edition in Nordiska kämpadater (Stockholm, 1737), which was based on manuscripts available in Stockholm at the time (Papp. fol. nr 1, Papp. fol. nr 25, Papp. 4to nr 13, Papp. 4to nr 17 and Papp. 4to nr 38).

The digital markup of the AM 285 4to text is done in XML according to the TEI P5 standard (see the “Stories for all time” Guidelines). There are two levels of transcription: diplomatic and normalised. The diplomatic level, marked with the element <orig>, preserves the original spelling, including orthography and special characters. The normalised level, marked as <reg>, has standardised spelling, and the text has been divided into chapters. Users have the ability to switch between transcription levels through the web viewer ("Choose rendering mode"). Other elements of the text, such as abbreviations, word division, punctuation and line breaks, are marked up separately and can also be displayed according to user preference (press buttom "Options"). By default the text is presented at the diplomatic level with expanded abbreviations, normalised punctuation and paratextual features shown. A series of download options permit users to obtain individual pages in HTML or the complete text in XML or PDF format.

Thanks to the flexibility of the electronic markup, the editors of AM 285 4to were able to retain a significant amount of information from the manuscript. Nevertheless, some editorial decisions had to be made. For this edition, variant letter forms are only retained where different allographs appear side by side. This is the case for long and round s as well as ordinary and insular f. No distinction is made between ordinary and round r (“r-rotunda”) or uncial forms and their ordinary variants, since the manuscript does not show any variation. Small capitals are retained, and ligatures with an independent phonemic value are preserved. Moreover, accent marks (“acute”) and umlaut markers are shown, except where the acute only serves the purpose of distinguishing the letters i and u from other minims.

Editorial emendations are marked and can be distinguished by the following symbols:

[ ] Supplied because of damage

< > Supplied because of omission

* Corrected by the editor

⸡ ⸠ Deemed superfluous by the editor


Changes or additions present in the manuscript itself are indicated using the following signs:

⸢ ⸣ Addition in the margins

⸌ ⸍ Addition above the line

⸍ ⸌ Addition below the line

\ / Addition within the line

ɨʉ Deletion


In order to provide multiple text and layout options, the markup incorporates a number of TEI elements, most of which are children of the element <choice>, which is used to group alternative encodings of the same point in a text, such as (1), abbreviations <am> and their expansions <ex> or (2) apparent errors <sic> and their editorial emendations <corr> (3).



          <orig>þ<choice> <am>̅</am> <ex>eir</ex> </choice>ra</orig>


     <sic>þä</sic> <corr>þeim</corr>

The <g> element (“glyph”) is used for variant letter forms. In the diplomatic display of the edition, a separate character declaration permits the user to choose between glyphs and standard letter forms. Glyphs are defined using different mappings for the variant letter form and the normalised display (4). In the transcription itself, the glyph is referred to by the <g> element, leaving the display choice to the user interface (5).

<glyph xml:id="slong">
     <glyphName>LATIN SMALL LETTER LONG S</glyphName>
     <mapping type="dipl">ſ</mapping>
     <mapping type="codepoint">U+017F</mapping>
     <mapping type="norm">s</mapping>

          <orig>þe<g ref="#slong"/>su</orig>

Original capitalisation has been retained and can be shown if desired. Words in the text which are capitalised, often seemingly at random, have been marked with the attribute @rend=”cap” (6), and can be so displayed. Where the manuscript uses lowercase letters for the first words of sentences, for example, these can be capitalised during processing, as can proper names, which are tagged using the <name> element and the appropriate attribute @type (7). Word division and rendition of initial characters are handled the same way. The <hi> element is used for highlighting various features, such as the first word of a chapter division (8), and the <seg> element is used for orthographic compounds, where two words are written as one (9). Compound words written separately are encoded by putting whitespace between the elements.

<w rend="cap">kraka</w>

<name type="person">

<hi rend="1">Epter</hi>

<seg type="ortho" rend="compound">

Paratextual features such as catchwords, hyphens and line-fillers are encoded in a variety of ways. For catchwords the markup uses the “forme work” element . Distinctions between the various types of catchword are achieved by specifying the @type attribute as ”catch”, ”pseudocatch” or ”last-line-catch” (10).

<fw type="catch">

The XML document containing the edition of the text also has a physical description of the manuscript within the header. This ensures that information regarding the physical properties and the content of the manuscript are kept in one document. The manuscript description will also serve as a catalogue entry for the manuscript and will soon be available on the Stories for All Time website. Currently, the XML source code is only available in page sections and can be chosen instead of the diplomatic or normalised view.

The electronic edition of AM 285 4to is a collaborative effort by the research group “Stories for all time: The Icelandic fornaldarsögur” at the Arnamagnæan Institute in Copenhagen. Contributors to this edition have been: Matthew Driscoll, Reynir Þór Eggertsson, Charlotte Johansen, Eric Haswell, Philip Lavender, Tereza Lansing and Beeke Stegmann.

Beeke Stegmann, 2015

(Please cite this edition as: Stories for all time: The Icelandic fornaldarsögur, Ed. Hrólfs saga kraka: An electronic edition from AM 285 4to (København: 2013) [http://].)