Le Médiéviste et l’ordinateur
Le Médiéviste et l’ordinateurHistoire médiévale, informatique et nouvelles technologies
n° 41 (Hiver 2002) : L’apport cognitif

Text encoding of manuscripts :
Danish prayer books from the 16th century

Anne Mette Hansen
Det Arnamagnaeanske Institut,
University of Copenhagen


[Consulter également le résumé en français rédigé par Élisabeth Lalou]

Electronic text encoding lays the groundwork for a new way of scholarly editing. Text encoding means to preserve and augment information of artefacts, it opens the eyes to the variance and the visuality of manuscripts and other primary sources, and it questions the traditional concept of text and the critical edition.

Danish medieval prayer books have been edited from a textual point of view in historical-critical editions [1] which take no account of the nature of the prayer texts and the social function of the books. An adequate edition is an edition which focuses on the material form in which the texts are found and presents the prayer books as the artefacts they are. A modern edition is an edition which can be used in different ways by an interdisciplinary audience. A new electronic edition combines such a complex edition and a textual and linguistic corpus [2]. The prayer books are encoded with the Extensible Markup Language (XML)[3]. See also Agnès Tutin, Chantal Wionet, Nathalie Lanckriet : « SGML pour l’informatisation des dictionnaires anciens : l’expérience du Dictionnaire Universel de Furetière revu par Basnage (1702) », in Le Mediéviste et l’Ordinateur. Le texte mediéval sur Internet-2 : Mettre des textes sur Internet, n° 32, 1999. based on the Text Encoding Initiative’s recommendations for the encoding of textual material [4], and the MASTER [5] standard for manuscript description.

The text encoding is a new philological way of encoding which integrates different philological disciplines, literary criticism (the texts are verbal utterances in context), rhetoric (textual structure, figurative language and metrics) linguistics (morphology and syntax), textual criticism (emendations and normalization), palaeography (scribal analysis), codicology (layout and physical structure) and visual arts. The present paper presents the type of transcription and the textual and codicological encoding on the source level, but first it gives an introduction to the encoded books.


There are two types of book meant for private worship, the book of hours and the prayer book. In a book of hours the prayers are always preceded by a Calendar, The Hours of the Virgin, The Seven Penitential Psalms and a Litany, whereas the prayer book generally does not contain these items. The textual programme of a prayer book from the late Middle Ages has a standard form: at the beginning of the book there are prayers to God the Father, the Holy Ghost and the Trinity. They are followed by prayers to the suffering of Jesus, his five wounds, his limbs, his resurrection and the prayers of the communion. The book continues with prayers to the Virgin, to her sorrows and joys, to her mother S. Anne and, in the end, prayers to a variety of saints and angels. The selection and the arrangement of the prayers vary in each book depending on the user for or by whom it was made ; every extant Danish prayer book is unique.

Prayers are accompanied by presentations, either preceding or following the prayer text. These rubrics are the ‘user’s guide’ to the prayers. They state the purpose of the prayer and give instructions on how, when and how often the reader shall pray. They also inform the reader of the effects of the prayer, sometimes by providing an example, and very often they promise an indulgence if these recommendations are followed.

The prayer texts themselves are formal in a highly structured way comparable to an epistolary form, each block having its own place in the form and being constructed by variable formulas and collocations. The predominant stylistic figures are apostrophe and tautology (pleonasm), the frequent use of adjectives in the superlative and epithets.

Marine Jespersdatter’s prayer book

The prayer book known as Marine Jespersdatters bønnebog, with the number AM 421 12mo in The Arnamagnæan Collection at the University of Copenhagen, is a piece of unique Scandinavian cultural heritage. It is a parchment manuscript from around 1500 in the size of a very small pocket book, 103 x 68 x 47 mm and 120 leaves. The origin of the prayer book is unknown, but it was presumably produced for a Danish noble woman named Marine Jespersdatter as stated in a prayer for the owner [6].

This case study provides a wonderful opportunity to integrate different scholarly fields in the research of one of the main religious genres of the Middle Ages. The prayer book contains the only known medieval Danish translation in verse of the Latin compassionate hymn Stabat mater dolorosa. This poem has given the book a place in the history of Danish literature, though it did not lead to any analysis of the whole textual complex. The book is also remarkable for its 48 fully coloured miniatures illustrating the prayers. No other Scandinavian prayer book containing a series of pictures illustrating the prayer texts has been preserved, and an iconographic examination of these pictures and the relationship between the pictures and the text is therefore essential.

The majority of the miniatures in AM 421 12mo follows the programme found in the European book of hours : Scenes from the life of Mary and Jesus beginning with the Annunciation and ending with the Coronation of the Virgin and the Last Judgement, illustrate the prayers in the central part of the book. This central part, a Psalter of the Virgin and of the Holy Spirit contains 34 prayers to Jesus and Mary, placed on the facing pages with the prayer text on the right-hand side and an accompanying illustration on the left-hand side. There is a close relation between the text and the depiction ; the rubric literally refers to the subject of the image which depicts a central theme in the prayer text. Thus each prayer is reproduced both in a painted and in a written version. In addition to this narrative image cycle which combines the pictures traditionally illustrating the Hours of the Virgin (Infancy and Passion cycle), the Hours of the Cross, the Hours of the Holy Spirit and the Seven Penitential Psalms there are a number of devotional images of the Trinity and the Virgin.

There are other good reasons to choose this manuscript as an object of study. Prayers are the largest group of extant Danish texts from the Middle Ages. An examination of the texts as they have been put together in each book, focusing on the visual and physical aspects of the artefacts as well as the intellectual contents, will define the style and the language of the prayers and throw light on the religious practice and mentalité of the 16th-century devoted laywoman. Most of the prayer books are connected with women. Owned by women, written for women, and sometimes also written by women.

The TEI / XML encoding

The markup of a primary source preserves the visual information of the manuscript page, it brings new information in showing how the encoder has interpreted a text and its intellectual content and it lays the groundwork for processing the encoded documents for display, i.e. the editions.

An XML markup based on the TEI markup scheme describes structured documents in a way that makes them interchangeable, and it makes it possible to combine transcribed text and metadata about the transcribed sources themselves, the transcriptions, the digitization and the publication. In the text header of an XML document, the encoder describes the principles of the encoding along with a description of the encoded manuscript source. The MASTER standard is an extension to the TEI in the encoding of manuscript descriptions and manuscript catalogues, and it is a most suitable way to encode a manuscript description in a highly structured way. The document type definition is an adaption of the TEI DTD’s.[7]

Complex manuscript material and historical texts, such as medieval prayer books, are a challenge to standard XML markup, because an encoding of each hierarchy the encoder observes in the text may result in the overlapping of the hierarchies in the documents. If you want to publish a ‘whole book edition’ which presents the many different aspects of the manuscript such as the physical structure, the textual structure, rhetorical devices and vocabulary, extensions of the existing standards will be necessary.

The principles of the encoding

The encoding on the source level provides an encoded diplomatic transcription of the text with a codicological markup concerning the layout and a markup of the intellectual content as regards the textual structure of the prayer book as a whole and the rhetorical structure of the individual prayers.

1. The type of transcription

Since digital facsimiles of the manuscript pages form the base reproduction of the text, and a fully diplomatic transcription (a facsimile print) which reproduces all the different characters and signs of abbreviation is not necessary. The level of transcription, however, is diplomatic in the sense that it reproduces the source text closely.

The special characters ‘æ’ and ‘ø’ are encoded as numeric character entities, æ and ø.

Abbreviations are expanded by using the <expan> element for the expansion with an attribute value ‘abbr’ for the abbreviation itself <expan abbr="&bar;">. Information about the abbreviation is given in form of an entity which reflects the name and function of the sign in question, thus the superscript horizontal bar used as a general mark of contraction is named &bar;, the nasal stroke for ‘n’ or ‘m’ &nbar;, the final 3-like sign for ‘et(h)’ or ‘m’ &et;, the superscript vertical z-like stroke for ‘-er’ or ‘-e-’ &er;, etc. [8].

The transcription is graphemic which means that variants such as ‘i’ and ‘j’ (‘long i’ / ‘i’ with a descender), ‘r-rotunda’ and ‘r’, ‘low s’ and ‘long s’, ‘y’ and ‘ÿ’ are transcribed as ‘i’, ‘r’, ‘s’ and ‘y’. A graphemic transcription facilitates an examination of the vocabulary, and if any reader wishes to know which form was used in the manuscript, he or she can consult the images.

Editorial emendations of obvious scribal errors in the primary source such as misspellings and omissions are tagged, and in each case the original reading is supplied in the form of the attribute value: <corr sic=""> , <supplied reason="omitted">. In cases where the transcriber wishes to indicate a suspicious reading without correcting it the <sic> element and the ‘corr’ attribute are used : <sic corr=""></sic>.

2. The codicological markup

Manuscript quires, page and line boundaries are marked up using the empty elements for quire, page and line breaks <qb/>, <pb/>, <lb/>. It is possible to align the transcription and the facsimiles by assigning an ID value to the <pb/> element

<pb n="56r" id="AM421_56r"/>.

The highlighting of the text has a decorative and mnemonic function. Thus the rubrication, decorated or coloured initials, initial letters filled or stroked in red, underlining and line-endings are marked up using the element for highlighted text with the rendition attribute : <hi rend="">. Shifts of the scribal hands are marked up with references to the description of the writing in the header of the document

<handShift new="hand2" old="hand1"/>.

The encoding of these categories registers the visual programme of the prayer book and different scribal habits.

3. The markup of the intellectual content

Different categories belonging to the intellectual content are marked up, among others the division of the text and the hierarchical and rhetorical structure of the prayer. The text as a whole is divided into textual items marked as <div>’s, <div type="psalter">, <div type="prayer">, <div type="prayer in verse">, <div type="psalm">, <div type="litany">, an ‘id’ attribute identifying the textual division and a ‘type’ attribute giving the information of the type of prayer, and each item is given a supplementary title in which the identity of the worshipped figure and the subject of the prayer are registered. This encoding makes it possible to search for and locate types of prayers and titles. A prayer consists of the prayer text and the preceding and / or following rubric <div type="rubric">, <div type="final rubric">, and a collect, an antiphon or a verse may finish the textual unity  <div type="collect">, <div type="antiphon">, <div type="verse">.

<div type="psalter" id="MJ20">

   <div type="prayer" id="MJ20-2-14">

      <head type="supplied">A prayer to the Virgin for the Visitation</head>

      <div type="rubric"></div>



The « rhetorica divina » [9] of the prayer text is encoded using the same subdivision element <div>, which may include other <div> elements, and the « type » attribute identifies the rhetorical content of the subdivision

<div type="prayer" id="MJ20-2-14">

   <div type="exordium">

      <div type="invocatio"></div>

      <div type="narratio"></div>


   <div type="petitio">



Other rhetorical devices, such as tropes and figures, are marked by the use of the <seg> element for textual units or segments. The notion of  ‘segment’ is broad: figures of speech <seg type="figure"> and other rhetorical devices <seg type="praise"> as well as user guiding content in the rubrics <seg type="instruction">. The attributes ‘type’ and ‘subtype’ provide the category of the segment <seg type="figure" subtype="apostrophe">, <seg type="figure" subtype="pleonasm">, <seg type="figure" subtype="tautology">, <seg type="trope" subtype="metaphor">,  <seg type="rhyme" subtype="alliteration">, <seg type="rhyme" subtype="assonance">.

Quotations <q> and titles of works <title> occurring in the source text are tagged, and to specify the language of the title or quotation the ‘lang’ attribute is used

<q lang ="LAT">, <title lang="LAT">. Personal names are identified and given a ‘key’ attribute with a code value referring to a separate bibliographic file of names where the persons in question are described

<name type="person" reg="Maren Jespersdatter" key="MarJes01">.

An example of the markup on the source level

The Danish text

English translation

[A prayer to Mary for the Visitation]


1    O Alsom helligsthe

O, most holy,

2    Iomffrw Maria bene=

Virgin Mary, praised

3    didhet oc welsigneth

and blessed

4    ware thw som saa ytmyghe=

are you, who most humbly

5    lighe wylle søghe thyn ffren=

would seek your kinswoman

6    kæ sancte Elisabeth oc gene=

Saint Elizabeth and at once

7    sthen glædes barneth sancte

the child Saint John rejoices

8    Iohannes i syn modhers lyff

in his mother’s womb

9    Oc thw aff stwr glædhe oc

And you from great joy and

10  gwdz nadhe dichtedhe then loff

the grace of God composed the praise

11  sang Magnificat O werdige

Magnificat O worthy

12  Ionffrw Maria werdis tyl ath

Virgin Mary be pleased to

13  søghe meg i myn døtz stwnd

seek me in the hour of my dead

14  Amen Item her skwlle i læsse

Amen Item here you should read

15  .v. Aue Maria for Ion

five Ave Maria in front of the

16  frw Mries (sic) billithe som hwn

Virgin Mary’s image as she

17  soghede syn frenke elizabeth

sought her kinswoman Elizabeth

The XML file

<?xml version="1.0"?>

<!DOCTYPE TEI.2 SYSTEM "C:/TEI/dtd/tei2.dtd" [

<!ENTITY % TEI.verse "INCLUDE" >

<!ENTITY % TEI.transcr "INCLUDE">

<!ENTITY % TEI.names.dates "INCLUDE">

<!ENTITY % TEI.figures "INCLUDE">

<!ENTITY % TEI.corpus "INCLUDE">

<!ENTITY % TEI.linking "INCLUDE">


<!ENTITY % TEI.extensions.ent SYSTEM


<!ENTITY % TEI.extensions.dtd SYSTEM


<!ENTITY bar "">

<!ENTITY nbar "">

<!ENTITY et "">

<!ENTITY er "">

<!ENTITY is "">

<!ENTITY dis "">

<!ENTITY isup "">

<!ENTITY hyphen "=">

<!ENTITY xp "">

<!ENTITY linefiller "">

<!ENTITY versemark "">

<!ENTITY paragraphmark "">


<?xml_stylesheet href="file:///C|/master/transcription_amh.css" type="text/css"?>





             <title>AM 421 12mo: A XML encoded transcription of 45r, the 3rd prayer of a psalter of the Virgin and the Holy Spirit, a prayer to Mary for the Visitation.</title>

             <respStmt><resp>Transcribed from the manuscript and marked up by</resp></respStmt>



         <publicationStmt><p>Unpublished demonstration file</p></publicationStmt>



               <msIdentifier n="KK2501" id="AM12_421">

                  <country reg="DK">Denmark</country>


                  <repository>Det Arnamagn&#230;anske Institut</repository>

                  <idno>AM 421 12mo</idno>








         <div type="psalter" id="MJ20"><pb n="45r" id="AM421_45r"/>

            <div type="prayer" n="20.1.3" id="MJ20.1.3">

               <head type="supplied">A prayer to the Virgin Mary for the Visitation</head>

               <div type="exordium">

                  <div type="invocatio"><p><lb n="1"/><seg type="figure" subtype="apostrophe"><hi rend="3">O</hi> <hi rend="stroked in red">A</hi>lsom helligsthe <lb n="2"/><hi rend="stroked in red">I</hi>omffrw Maria</seg>

<seg type="praise"><seg type="figure" subtype="pleonasm">b<expan abbr="&bar;">e</expan>n<expan>e</expan>&hyphen; <lb n="3"/>didh<expan abbr="&et;">et</expan> oc welsigneth</seg> <lb n="4"/>ware thw</seg></p></div>

                  <div type="narratio"><p>som saa ytmyghe&hyphen; <lb n="5"/>lighe wylle s&#248;ghe thyn ffren&hyphen; <lb n="6"/>k&#230; <name type="person">sancte <hi rend="stroked in red">E</hi>lisabeth</name> oc gene&hyphen;<lb n="7"/>sth<expan abbr="&er;">e</expan>n gl&#230;des barneth <name type="person">sancte <lb n="8"/><hi rend="stroked in red">I</hi>ohannes</name> i syn modh<expan abbr="&er;">e</expan>rs lyff <lb n="9"/><hi rend="stroked in red">O</hi>c thw aff stwr gl&#230;dhe oc

<lb n="10"/>gwdz nadhe dichtedhe th<expan abbr="&er;">e</expan>n loff <lb n="11"/>sang <title>Magnificat</title></p></div>


               <div type="petitio"><p><seg type="figure" subtype="apostrophe"><hi rend="stroked in red">O</hi> werdige <lb n="12"/><hi rend="stroked in red">I</hi>onffrw <hi rend="stroked in red">M</hi>aria</seg> wer<expan abbr="&dis;">dis</expan> tyl ath

<lb n="13"/>s&#248;ghe meg i my<expan abbr="&nbar;">n</expan> d&#248;tz stwnd

<lb n="14"/><hi rend="stroked in red">A</hi>me<expan abbr="&nbar;">n</expan></p></div>

               <div type="finalrubric"><p><hi rend="red">Item <seg type="instruction">her skwlle i l&#230;sse <lb n="15"/>.v. <title>Aue Maria</title> for Ion<lb n="16"/>frw <sic corr="Maries">Mries</sic> billithe</seg> som hw<expan abbr="&nbar;">n</expan>

<lb n="17"/>soghede syn frenke <name type="person">elizab<expan abbr="&et;">eth</expan></name></hi> <pb n="46r"/></p></div>






The metrical markup

The Danish version of the Latin hymn Stabat mater dolorosa has been an object of great interest to scholars in the field of verse and style history. There are other prayers in verse in the prayer books, and a metrical markup may identify the poetic tradition in the prayers and the relation of the pre-reformation prayer to the subsequent post-reformation prayer, to psalms and ballads.

       The metre of Hos Korsens Træ is a standard medieval metre, the iambic dimeter, consisting of four metric accents, the fifth line consisting of three iambs:

 - + | - +  and  - - + | - + | - + -.

The stanzas belong together in pairs, and the rhyme scheme is aaBBC  ddeeC. The stanzas and the verse lines are tagged as numbered line groups <lg n="1"> and lines <l =n"1">. The ‘type’ attribute identifies the line groups as stanzas <lg type="stanza">, and the attributes ‘met’ and ‘rhyme’ register the metre and the rhyme scheme of the stanzas <lg met="iambic dimeter" rhyme="aaBBC">, while the ‘real’ attribute on the <l> element specifies deviation from the metrical pattern defined by the ‘met’ attribute.

The Danish text

English translation

1    HOss korssens tree

1    At the cross tree

      meth sorg oc wee

      with sorrow and woe

      stood cristi moder

      stood Christ’s mother

      med graadzens floder

      in floods of tears

      hennes søn paa korssyt hengde

      her son was hanging on the cross


2    Aff sywck oc graad

      From sighs and tears

      som ffør war spaad

      as previously predicted

      meth stor wfryd

      in great distress

      stack drøwelzens spyd

      stuck sorrow’s spear

      hennes hierthe thet tha trengdhe

      her heart it tormented

The encoded text

<fileDesc><titleStmt><title>AM 421 12mo: A metrical encoding of 30 recto, l.1-8, the two first stanzas of Hos korsens tr&#230;, the versified Danish version of the Latin hymn Stabat mater dolorosa.</title>

<respStmt><resp>Transcribed from the manuscript and marked up by</resp></respStmt><editor>Anne Mette Hansen</editor></titleStmt></fileDesc>



<pb n="29r" id="AM421_29r"/>

<div type="prayer of contemplation in verse" n="15" id="MJ15">

<head type="supplied">Hos korsens tr&#230; ( Stabat Mater)</head>

<div type="rubric">

<p><hi rend="red ">Item her effther skriff&#173;<lb n="13"/>wes een b&#248;n aff Iomfrw <lb n="14"/> maria I hwilken b&#248;n <lb n="15"/> wor herres ih<expan abbr="&bar;">es</expan>u <expan abbr="&xp;">crist</expan>i d&#248;d oc <lb n="16"/> pynelss&#230; y hwffkommess <lb n="17"/> och he<expan abbr="&nbar;">n</expan>nes <sic corr="hwffswalelss&#230;">hwffslalelss&#230;</sic></hi></p>


<div type="prayer text">

<pb n="30r" id="AM421-30r"/>


<lg2 n="1" type="stanza" met="iambic dimeter" rhyme="aaBBC">

<l n="1" rhyme="a"><hi rend="3_red">H</hi><hi rend="stroked in red">O</hi>ss korssens tree &versemark;</l>

<l n="2" rhyme="a">m<expan abbr="&et;">eth</expan> <seg type="figure" subtype="tautology">sorg <lb n="02"/>oc wee</seg> &versemark;</l>

<l n="3" rhyme="B" real="_+_+_">stood <expan abbr="&xp;">crist</expan>i mod<expan abbr="&er;">er</expan> &versemark;<lb n="03"/></l>

<l n="4" rhyme="B">med <seg type="trope" subtype="metaphor">graadzens floder</seg> &versemark;<lb n="04"/></l>

<l n="5" rhyme="C" real="__+_+_+_">hennes s&#248;n paa korssyt hengde &versemark;<lb n="05"/></l>


<lg2 n="2" rhyme="ddeeC">

<l n="1" rhyme="d"><hi rend="2_blue">A</hi>ff <seg type="figure" subtype="tautology">sywck oc graad</seg> &versemark;</l>

<l n="2" rhyme="d">som ff&#248;r <lb n="06"/> war spaad &versemark;</l>

<l n="3" rhyme="e">m<expan abbr="&et;">eth</expan> stoor wffryd &versemark;<lb n="07"/></l>

<l n="4" rhyme="d" real="_+__+">stack <seg type="trope" subtype="metaphor">dr&#248;welzens spyd</seg> &versemark;</l>

<l n="5" rhyme="C" real="__+_+_+_">he<expan abbr="&nbar;">n</expan>nes <lb n="08"/> hierte th<expan abbr="&er;">e</expan><expan abbr="&et;">t</expan> tha trengdhe &versemark;</l>









The coming stage of the encoding will show how the concurrent markup of the normalized level and the lemma level may interact. The source level markup outlined in this paper, is, along with a lemmatization and a normalization of the original reading in the texts, the markup that is necessary in order to investigate, both as single objects and as a corpus, the style and language of the prayers, the structure and textual programme of the texts, the iconographic programme of the painted prayers, the relations between the decorative features and the written text, and finally the use of the books.

[Consulter également le résumé en français rédigé par Élisabeth Lalou]

[1]. Karl Martin Nielsen (ed.), Middelalderens danske Bønnebøger I-V. København, Gyldendal, 1946-1982.

[2]Private devotional literature in the 16th century : The digitisation of Danish prayer books in The Arnamagnæan Collection and in the Library of Karen Brahe is the working title of a PhD project funded by the Danish Research Council for the Humanities. The project concerns the digitization of Danish national literature and scholarly editing and aims to lay the methodological and practical groundwork for digital editions of Danish literature from before 1900 developing a standard for encoded transcriptions of Danish primary sources and suggestions to the processing of the documents. It also concerns a new integrational philological approach and editorial theory.

[3]. XML is the World Wide Web Consortium’s markup language. It is extensible and it is independent of applications and platforms which means that XML marked documents can be moved between computers and processing systems. A markup language is a way of describing document structures, it is a set of codes which are used to identify or ‘tag’ certain features in a text. An XML document consists of ‘elements’ and ‘entities’. An element consists of to markup ‘tags’, an opening tag <note> and a closing tag </note> and the text between the two tags, the element name tells that the content of the element is of a certain category, e. g. a note, <note>this is a note</note>. Elements can have ‘attributes’ and ‘values’ applied refining the meaning of the element, <note type=»endnote»>this is a note</note>, to be more precisely, an endnote. ‘Empty elements’ have no text between the opening and closing tag, e. g. <pb/> marks a page break in the source.

[4]. The Text Encoding Initiative (TEI) is an international scheme for standard markup of electronic texts. The 2001 version of the encoding reccomendations is TEI P4 Guidelines for Electronic Text Encoding and Intercharge : XML_compatible edition, edited by C. M. Sperberg_McQueen and Lou Burnard, the TEI Consortiu. http://www.tei-c.org/P4X/.

[5]. MASTER (Manuscript Access through Standards for Electronics Records) is an international European Union funded project whose goal is to define and implement a standard for manuscript descriptions using SGML/XML. The Reference Manual for the MASTER Document Type Definition : Discussion Draft, edited by Lou Burnard for the MASTER Work Group, is at http://www.hcu.ox.ac.uk/TEI/Master/Reference.

[6]. We are not certain who Marine Jespersdatter was. It has been suggested that she was a daughter of the nobleman Jesper Friis of Lyndby and Hesselager on Fyn. If this uncertain identification with Mette Jespersdatter Friis is correct, she was married to Claus Urne of Hindemae on Fyn in his first marriage, and died childless (Claus Urne died in1531).

[7]. The current version of the TEI P4 DTD including the additional tagsets for physical transcription, language corpora, names and dates and the TEI.extensions.dtd to allow the MASTER <msDescription> in the <sourceDesc> element in the <teiHeader>.

[8]. The names employed for the entities are in accordance with the recommendations of M. J. Driscoll, « Encoding Old Norse / Icelandic Primary Sources using TEI-Conformant SGML », Literay and Linguistic Computing, Vol. 15, no. 1, Oxford, University Press, 2000, p. 81-91.

[9]. See Eckart Conrad Lutz, Rhetorica divina : Mittelhochdeutsche Prologgebete und die rhetorische Kultur des Mittelalters. Berlin, Walter de Gruyter, 1984.


[Consulter également le résumé en français rédigé par Élisabeth Lalou]

© CNRS - IRHT 2004 - Contacts - Infos légales - Rédaction