Ch. 9. Metrical encoding

Version 1.1 (5 May 2004)

9.1 Introduction
9.2 Stanzas
9.3 Markup of the metrics in lines

Back to list of contents

9.1 Introduction

There are good reasons to establish a system for the encoding of metrics in the medieval manuscripts even if this structure is not generally represented graphically in the manuscripts. For many users of the established text the stanzas are of great interest and it is therefore practical to mark them in a way that makes it possible to find and delimit them from the surrounding text. A more detailed encoding of the stanzas can open up for new ways of research on different metrical variants, concerning e.g. alliteration, internal rhyme and stress.

The same rules that apply for prose are relevant also for the encoding of stanzas. In addition to these rules we suggest codes that fascilitate the search for and identification of stanzas. We also give guidelines as to the encoding of the metrics. It should be pointed out, however, that these codes would have to be given a more detailed form if the stanzas should be analysed in a special database.

On the primary level we recommend that the stanzas are encoded with the following elements and attributes:

9.2 Stanzas

A stanza is marked with the element <lg>, linegroup. It should be noted that the TEI Guidelines do not accept an element <lg> within an element <p>. The element <p> will therefore always have to be closed before each encoded linegroup and reopened after </lg>.

The attributes n and id are used to identify the stanza. The first attribute defines the stanza within the manuscript, while the second attribute indicates the relation of the stanza to the medieval poetic corpus.

The information given in the attribute n is divided into two sub-groups with a dot.

The first field in the attribute n indicates a manuscript siglum. We suggest that the sigla from Finnur Jónsson Den norsk-islandske skjaldedigtning (Skj) are used. Manuscripts that are not given a siglum in Skj are marked with the relevant catalogue number. For example the Codex Regius of the Elder Edda should be marked as RE and the Codex Wormianus as W. In the second field the place of the stanza within the manuscript is given. The linegroup for a stanza should be given as follows:

<lg n="W.1">
   <l> ... </l>
   <l> ... </l>
</lg>

The attribute id identifies the stanza in relation to the total corpus of Old Norse-Icelandic poetry. It has three sub-groups divided with a dot that can be presented as follows:

In the first field, information is given as to genre, i.e. Eddic or Skaldic, the later referred to by an attribute for the known skald constituted by the two first letters of the name and the first letter of the patronymicon as shown above. For skalds that have an attributive name, i.e. skaldaspillir the first letter in this name is given instead of the first letter of the patronymicon. The Eddic poems are indicated with an E. Anonymous Skaldic stanzas are marked NN and century according to Skj, e.g. NN13 if the stanza is dated to the thirteenth century.

In the second field information about the individual poem is given. The Eddic stanzas are marked with sigla (see below). The Skaldic stanzas are marked with the first word as it is given in Skj B, but without accent marks. The special signs used in Old Norse-Icelandic texts are changed (ð>d, þ>t, æ>ae, ö>oe and o caudata>o). In cases where the same skald or a number of anonymous skalds from the same century has the same first word, a second word is included without space.

In the third field the number of the stanza is given in relation to Skj for Skaldic stanzas and Bugge 1867 for Eddic stanzas.

For a Skaldic stanza attributed to the skald Bragi Boddason this should be given as follows:

<lg id="BrB.vilid.13">
   <l> ... </l>
   <l> ... </l>
</lg>

And the anonymous stanza number four from Eiríksmál should be encoded as follows:

<lg id="NN10.hvat.4">
   <l> ... </l>
   <l> ... </l>
</lg>

9.2.1 Sigla for Eddic poems

We suggest that the Eddic poems should be marked as in the following list. (For technical reasons "o caudata" is rendered as "ö" in this list).

Thereby stanza 13 from Völuspá should be rendered as follows:

<lg id="E.vsp.13">
   <l> ... </l>
   <l> ... </l>
</lg>

9.2.2 Metrical categories

With the attribute type it is possible to characterise the metrics closer. The stanza can be related to the general groups according to the following list:

For the stanza from Eiríksmál above the metrical group should be rendered as follows:

<lg id="NN10.hvat.4" type="dk">
   <l> ... </l>
   <l> ... </l>
</lg>

9.3 Markup of the metrics in lines

Every individual line of a stanza should be marked with the element <l>. This encoding is based on metrics and is a complement to the element <lb/> for graphically marked line breaks in the manuscript text. If there is doubt about a line break in the stanza we recommend that the lines are analysed in accordance with Skj and Bugge 1867 for Skaldic and Eddic poems respectively.

The attribute n in the <l> element identifies the line number of the stanza, e.g. n="1", n="2" and so on. This encoding will be generated automatically.

It is also possible to include information on number of syllables, alliteration and internal rhymes in the markup. In the primary encoding of manuscript text this is of course not obligatory.

The attribute ass is used to indicate internal rhyme in the stanzaic text. It can be given for each individual line of the stanza as in the following example:

<lg>
   <l ass="00S0S0">Vask með gram þeims gumnum</l>
   <l ass="A000A0">goll bauð dróttinhollum</l>
</lg>

In the first line we have an example of internal half rhyme (skothending) marked with an S. In the second line the internal rhyme is full (aðalhending) which is marked with an A. Syllables that are not included in the internal rhyme are marked with 0.

The attribute all is used to indicate alliteration in the stanzaic text. It can also be given for each individual line of the stanza as in the following example:

<lg>
   <l all="00L0L0">Vask með gram þeims gumnum</l>
   <l all="L00000">goll bauð dróttinhollum</l>
</lg>

The alliterating syllables are here marked with an L, while syllables that are not included in the alliteration are marked with 0. When internal rhymes and alliterations are marked it is not necessary to mark the text for the number of syllables of the line. This system does not include the markup of syllable stress.

The attribute met can be used to include information on the metrical form of the stanzaic line built on Sievers's categorization (Sievers 1893). In the primary encoding of manuscript text this is of course not obligatory. It is, however, possible to include this markup when it is relevant for a specific study of metrics. We suggest that the groups A-E used by Sievers should be applied in a first encoding, perhaps with the addition of sub-groups (see e.g. Gade 1995: 39). This should be given as in the following example:

<lg>
   <l met="B"> ... </l>
</lg>

In the following linegroup we have excluded elements such as <w>, <facs> and <dipl> to make it easier to see the structure of the encoding for a half stanza. We have also silently expanded abbreviations not to obscure the structure of the stanza. (The orthograhy in this example is slightly regularised.)

<lg n="W.00" id="EiG.Setbergs.1" type="dk">
   <l n="1" ass="S000S0" all="L000L0" met="A2k">Setbergs kveða sitia</l>
   <l n="2" ass="A000A0" all="L00000" met="A">suðr at urðar brunni</l>
   <l n="3" ass="00S00S0" all="00L00L0" met="A2k">sua hefr rammr konungr remðan</l>
   <l n="4" ass="0A00A0" all="L00000" met="D4">roms banda sík löndum</l>
</lg>

The encoding of metrical type is a time-consuming and demanding task that requires scholars with insigths into the metrics of skaldic and Eddic poetry. It is, however, possible to limit the markup to the main groups described above, and sub-groups can be marked preliminary with an *. The type A2k could e.g. be marked A*. Thereby the uncertainty of the more detailed analysis is marked.

Top of page