Pietro Liuzzo edited this page May 2, 2018 · 336 revisions


NEW version of the GUIDELINES

TEI should always be referred to for definitions and use of elements and attributes.

However TEI is not a prescriptive system or a schema, and the present guidelines describe the declension of the TEI for this project and its purposes. Whenever possible the reason for a particular encoding decision is given. This shall help also to track bugs related to data entry.

The ODD and the schema can be found in the schema repo.

If you think something should be automatised, click here first.

##Table of contents


The root element TEI should contain always an @xml:id and @xml:lang as well as a @type to quickly identify entities with one of the following values:

  • mss for manuscript
  • ins for institutions
  • pers for person
  • place for places
  • work for works
  • nar for narrative units
  • auth for authority files


Remember, in contents put spaces only where you really want them. As much as you want.


<roleName type="title">ʾaṣe</roleName>Galāwdewos 

will always return this result


if you want the space between ʾaṣe and Galāwdewos or any other two words, please, remember to have it. On the other side, the following two are equal in XML

<roleName type="title">ʾaṣe</roleName>                   Galāwdewos

<roleName type="title">ʾaṣe</roleName>


ID and Title

The manuscript ID should be composed of:

  • The abbreviation of the institution holding the manuscript, as it is contained in the insitution ID, written in capitals
  • The abbreviation of the collection to which the manuscript belongs
  • a numeral: BAVet1 The collection and the numeral should be given as they are in the current/most commonly used official shelfmark of the manuscript.

If the manuscript has not yet been catalogued and/or has no shelf mark, an ID has to created individually, following as much as possible the usual process_

  • The abbreviation of the institution holding the manuscript, as it is contained in the institution ID, written in capitals
  • If no more information is available, the ID can be created with a work contained in the manuscript and a numeral, such as: TMChronicle1
    <editor key="DN"/>
    <editor role="generalEditor" key="AB"/>
    <funder>Akademie der Wissenschaften in Hamburg</funder>

Authors and editors are referred to only by @xml:id. A list of ids of project members and editors is here

The title should be intended as a generic title for this manuscript record, which might refer to contents or identifiers. The title should be made of all the normalized titles of the works contained in the manuscript. If there are many different works which would make the title too long, it is possible to create a more generic title where this makes sense.


Any @xml:id within the file can be pointed to with an element ref with a @target attribute. oXygen will not let you enter a double id, and will pull up a list of @xml:id when you attempt to enter @target or @corresp

<ref target="#p1"/>

This example will point to the msPart with @xml:id=‘p1’. I the preview you will see it as ’n. p1’ with a link to the point in the resource you are speaking of. You can also use this to mark up a value, e.g.

this happens also in <ref target="#q14">the final quire</ref>

this will preserve your text. However, please note that multiple references will be splitted in several links, so having something like

this happens also in <ref target="#q14 #q15">the final quires</ref>

will not return the text in the element but the values of the targets.

You can also use @target to point to external resources. In this case, please enter the full url, including http.

If you are adding a reference, please simply mark it up using placeName, persName, etc. If you want to refer to another entity in the project but you don’t want to mark up the entity you can use ref with @corresp as follows. This is intended to be useful especially for pointing directly to a particular part of a manuscript description, e.g. a hand or a quire or a part. This should also be used to refer to specific works.

<ref type="mss" corresp="ESap001">AP-001</ref>
<ref type="work" corresp="LIT4018Mashafa"/>

In @corresp don’t start with a # but use it if you are referring to a specific part of the quoted manuscript.

This manuscript is in the same hand as <ref type="hand" corresp="ESap002#h2"/>

Please specify the @type for each reference. The values of type are

  • mss
  • title
  • mspart
  • author
  • place
  • item
  • hand
  • quire
  • binding
  • work
  • ins
  • pers
  • group
  • deco
  • authFile

To link to a manuscript in vHMML, please, use the provided permalink in the metadata.

 <ref type="mss" target="">

For citations, like Matthew 14, 3 use ref with a @cRef containing the urn which you can get from Perseus catalogue searching for the passage.

 <ref cRef="urn:cts:greekLit:tlg0031.tlg001.perseus-grc1:14.3">Matthew 14, 3</ref> 

You can also point to our texts for example to Numbers 21, 8 like this

 <ref cRef="urn:dts:betmas:LIT2075Number:21.8">Numbers 21:8</ref> 

To mark up a passage in the text which is a literal quotation of another text you might use cit

    <quote xml:lang="gez">
         ለሙሴ᎓ ግበር᎓ ለከ᎓ አርዌ᎓ ምድር᎓ ዘብረት᎓ ኵሉ᎓ ዘነሰኮ᎓ ይርአይ᎓ ኪያሁ᎓ ወይሕየው᎓ ኪዮስ᎓
    <ref cRef="urn:dts:betmas:LIT2075Number:21.8">Numbers 21:8</ref>

Bibliographic References

Bibliographic references are marked up by a bibl element containing at least a ptr element with a @target, as in the following example:

<bibl><ptr target="bm:Villa2013Erma"/></bibl>

If need an indication of exact reference, it should be given in a citedRange element, as in the following example:

<bibl><ptr target="bm:Villa2014Erma"/><citedRange unit="page">5-13</citedRange></bibl>

citedRange can take several values in @unit and can be repeated if there is a need to specify a volume, an entry, etc.

Please, do not use @from and @to in citedRange as this is unnecessary.

Please note that lists of references about a manuscript should be kept with the manuscript and bibliography about a text should be kept with the text, etc. No typed bibliography is entered in the record, only pointers to a common bibliographic database.

The target should be a uniquely assigned tag in the Zotero Group Bibliography EthioStudies. If the item you want to cite does not yet have a unique tag, please assign one with this format and using only latin alphabet letters


Please make sure this contains no special characters, only digits and latin characters. To check that this is unique, after saving the record in Zotero, click on the tag and check that only one item has this tag. If not, change the keyword of the tag just assigned.

When adding records to the Zotero Group Library, please follow the principles at this page.

Referencing the manuscript

locus element can be used almost everywhere to indicate the physical location of an item. It contains always the two attributes @from and to. Please do not use other numbering systems or spaces within these attributes.

<locus from=“5ra" to="43vb"/>

When a line number is given, use:

  • in case of an exact locus, line indicated by @n: <locus target="#1v" n="1"/>
  • in case of a range, lines indicated as a content : <locus from="1v" to="2ra">the text begins on 1v in line 1 and ends on 2ra in line 6</locus> OR with two <locus> elements: <locus from="134rb" n="11"/> <locus to="135ra" n="28"/>

A list of individual references at page level (without line indication) can be given in a single layout with a @target

<locus target="#28v #33r #49r #50v #55v #82r #82v #83r #83v #84v #85r”/>

Please DO NOT use unknow as a value for any of this attributes. Simply do not enter the attribute if its value is not known.

 <locus from="20r"/>

NOT <locus from="20r" to="unknown"/>

A manuscript's existing foliation should not be shifted - it is possible to point to folia not included in the foliation by assigning individual leave numbers. This will lead to the inability to calculate the foliation.

Two folios between <locus target="#66"/> and <locus target="#67"/>, <locus target="#66i"/> and <locus target="#66ii"/>, are so badly mutilated that only small corners - each not containing more than fragments of 4 letters - are preserved which have not been taken into account for the foliation of the manuscript.


The TEI file is in english. Please, make sure you use @xml:lang with the appropriate value for each element containing text in a different language. If the content of the element is mixed, please, use foreign

 <p> <foreign xml:lang="gez">ሄሮድያዳ፡</foreign> <foreign xml:lang="grc">Ἡροδιάδα</foreign> </p>

foreign should also be used for transliterated words:

<p> <foreign xml:lang="gez">salām</foreign> </p>

Named Entities and authority lists

Throughout the document placeName and persName will take a @ref containing the xml:id of the place or person record.

<placeName ref="LOC2969Ensala"/>

repository takes a @ref as well, pointing to the institution id, which will contain further place pointers.

     <repository ref="INS0030Addaqaharsi">

If the manuscript is lost and its current repository is not known, write


title takes a @ref as well, pointing to the work id, which will contain the standard edition of the text.

 <title ref="LIT1013ActaSa">

Variant titles, translations of the title, etc. will be contained in the file for that work.

Before creating new entities for PLACES like Stuttgart, Paris, London, etc. which unlike all institutions are very much unlikely to receive specific descriptions by this project, use external links to geonames ids as follows.

<country ref="gn:2017370"/> 
<settlement ref="gn:498817"/>

(same for placeName and region)

to decide whether to create a new place record or not if in doubt, try the following.

  • check if there is a record already in our list
    • if yes, then 1. check if there is also a wikidata item (you can simply google what you are looking for, open the correct wikipedia page you find and hit the wikidata item link on the left.
      • if yes add a @sameAs attribute with the Q id in the url or at the top of the item.
        1. check if there is also a record in
          • if yes add a @sameAs attribute to place linking to such geonames id.
          • if not leave the record
        2. add a ref to our record
    • if not, then 1. check if there is also a wikidata item 1. check if it is in * if yes add the uri in the @ref * if not (but check thoroughly, because it would be an exception) create a new place record

generally, when possible I would always prefer an external link to an authority file for places out of scope which we are not going to describe or give details about. If there is already a record, no need to delete it or update it. Simply add the geonames ID.

adding references to places not in our db:

  • for ancient place prefere pleiades, with namespace pleiades:
  • for places related to coptic literary tradition use PAThs paths:


Dates anywhere in the text quoted can be tagged with precision

<date calendar="ethiopian" when="1888" when-custom="1881" evidence="internal">፲ወ፰፻፹ወ፩ ዓመት፡ ምሕረት፡ </date>

@when always carries the standard ISO date in the gregorian calendar, while the @when-custom (or notBefore-custom, notAfter-custom) can hold the value in the calendar specified by @calendar.

BCE dates are therefore expressed with -: <death when="-0931"/>

date expressing a duration rather than a point in time, might use the @dur following w3c recommendations . A period of 10 years and 5 months would be encoded as

<date dur="P10Y5M"/>
attribute value
@when ISO gregorian date
@when-custom date in specified calendar
@notBefore date range limit
@notAfter date range limit
@notBefore-custom date range limit in specified calendar
@notAfter-custom date range limit in specified calendar
@dur duration
@type type of date (reign, lifespan, etc.)
@calendar specified calendar according to list
@resp with the id of responsible
@cert certainty low, medium or high (default)
@evidence especially for date assigned to something, with set of values

In-text placeName and persName

All information related to places or people should be stored in the respective place or person record. If a long description needs to be provided, then the name of the place might not be omitted.

The manuscript was perhaps found in <placeName ref="LOC2969Ensala">ʾƎnsalāle</placeName> but there is no agreement about this.

There are some standard typologies of persons identified in the record description which can be used.

Information about a place should be entered in the TEI record for that person following the guidelines for a Place entity.

Information about a person should be entered in the TEI record for that person following the guidelines for a Person entity.


Filenames and IDs should correspond EXACTLY and as always they are case sensitive. The main xml:id of a record is directly into the TEI root element.

<TEI xmlns="" xml:id="PRS1009Abakaraz">

roleName and @role

The titles of persons, such as ʾabbā, wayzaro, makʷannǝn etc., should be encoded with <roleName type="title"> at the place in the texts where they occur. In order to be reasonably retrievable, <roleName> always has to be encoded within <persName>. The name of the person can be typed in the <persName>, or, if avoiding typing it is preferable, the <persName> element with the same ID can be repeated immediately after the first <persName> which contains <roleName>:

It lists the books belonging to <persName ref="PRS10796Tasfo"><roleName type="title">ʾabbā</roleName></persName><persName ref="PRS10796Tasfo"/>


It lists the books belonging to <persName ref="PRS10796Tasfo"><roleName type="title">ʾabbā</roleName> Tasfo</persName>

will both be visualized as: It lists the books belonging to ʾabbā Tasfo.

Therefore, if only the title encoded should be visualized (without the name of the person it refers to), it should be marked up in the following way:

he was also appointed as <persName ref="PRS5309hesanMo"><roleName type="title">qāla nǝguś</roleName></persName> of King <persName ref="PRS2334BaedaM">Baʾǝda Māryām I</persName>

The use of <roleName> has to be clearly separated from the use of @role. The first is used to mark up a concrete title mentioned in a text, the second specifies the role the person has in regard to the record in which it is mentioned. This applies for instance to scribes or owners of manuscripts.

If possible, persNames should take a @role where relevant in the description, and then don't need to be listed again in <listPers>:

<item xml:id="e1">
<locus target="#60v #143r #144r #152v #167v #171r #173r #189v #192r #234r #255r"/> 
<desc>The names of the owner <persName role="owner" ref="PRS11297ZenaG"/> and his wife <persName ref="PRS11298SeyonM"/> are mentioned several times.</desc>

If there is no sensible place within the description, then the information on the role can also be encoded in <listPers> in particDesc.

Entities ID structure

  • Manuscripts
    • EMML2312
    • ESam002 (Ethio-Spare)
    • BLadd100 (British Library, fondo, numero)
    • BNFabb3 (Bibliotheque nationale de France, fondo, numero)
  • Works = LIT1001Homili (LIT + incremental number + first word of title)
  • Places = LOC2969Ensala (LOC + incremental number + first word of place name)
  • People = PRS1001AbbaTam (PRS + incremental number + first word of name)
  • Institutions = INS0030Addaqaharsi (INS + 4 digit number + initials)
  • Art Themes = AT1005Crucifixion (AT + incremental number + title of art theme)

To create a minimal manuscript record for referencing follow these simple guidance. For Places, Institutions, Works and People, look at the dedicated page in the side bar.

Manuscript Structure

Reading of TEI guidelines is highly recommended. It also says

Since each component of such a composite manuscript will in all likelihood have its own content, physical description, history, and so on, the structure of msPart is in the main identical to that of msDesc, allowing one to retain the top level of identity (msIdentifier), but to branch out thereafter into as many parts, or even subparts, as necessary.

This means that for manuscripts made of only one part we will use only msDesc, but in any case where a structure in codicological units occurs we will use msPart. msPart can contain other msPart. msFrag can be used when there are at least two parts of one manuscript kept in different repositories.

As for issue 368 we have decided to use msFrag and msPart as follows in cases where leaves or quires have been moved. The Ms containing currently the unit, will have a msPart with it, with an appropriate description and pointer to the manuscript where the unit comes from. The Ms which has lost a unit will instead have a msFrag pointing to the Ms where the unit is now. Please, do not confuse this with cases where leaves or quires are lost.

General Structure of Manuscript Records (Simple and Composite)

It is recommended when starting encoding manuscript records to read these guidelines alongside with example records, for example of the BAV or BDL collections.

Each manuscript record should contain in <msDesc> a description of the manuscript's intellectual and physical content, its history and the source of the record. The intellectual content of the manuscript is encoded in <msContents>, divided into as many (nested) <msItem>s as needed. The physical description of the manuscript is encoded in <physDesc>. <history> contains all available information on the manuscript's provenance and further history. The source of the record is encoded in <additional>.

The order for these elements should be for simple manuscripts:

  • <msIdentifier>, see below
  • <msContents>
  • <physDesc>
  • <history>
  • <additional>

Each codicological unit of a composite manuscript is a <msPart>. Since the codicological units of a composite manuscript might have different physical aspects and origins, all the above mentioned elements can and should be used as needed within each separate <msPart>. However, the shared aspects of the manuscript - for example parts of its history, or the description of the binding - can also be encoded for the entire manuscript. Therefore, the <msDesc> of a composite manuscript might have the following structure:

  • <msIdentifier>, see below
  • <physDesc> Containing the physical description pertaining to the entire manuscript
  • <history> Containing information on the origin and provenance pertaining to the entire manuscript
  • <additional> Containing the source of this record
  • <msPart xml:id="p1"> Description of the first codicological unit, possibly including again <physDesc>, <history> and <additional> as pertaining only to this codicological unit
  • <msPart xml:id="p2"> Description of the second codicological unit, possibly including again <physDesc>, <history> and <additional> as pertaining only to this codicological unit
  • <text> Containing any transcribed text of any <msPart>

Each <msPart> contains the description of its intellectual content in <msContents> as well as all the above elements, limited now to the description of this particular <msPart>. If no further information is available for example on the history of a particular <msPart> or if no other source than the one already specified in the first <additional> element concerning the entire manuscript has been used, it is not necessary to repeat these elements.


Each msPart, msItem and msFrag should have an @xml:id with a letter (respectively p, i or f) and a progressive number. For nested msParts numeration should restart and use a full stop for the second level.

<msDesc xml:id="ms">
     <repository ref="INS0030Addaqaharsi"/>
  <msPart xml:id="p1">
    <msPart xml:id="p1.1">
    <msPart xml:id="p1.2">
    <msPart xml:id="p1.3">
  <msPart xml:id="p2">

msItems will have an id formed of the msPart, underscore and the item id, like p.1.1_i1

<msPart xml:id="p1.1">
    <idno>  </idno>
    <msItem xml:id="p1.1_i1">
       <msItem xml:id="p1.1_i1.1">
       <msItem xml:id="p1.1_i1.2">
    <msItem xml:id="p1.1_i2">

More than one identifier for a manuscript or a part of manuscript

If there are other possible identifiers, add them with the following elements Inside msIdentifier

     <repository ref="INS00000030Addaqaharsi"></repository> 

Into msPart

There should be a msIdentifier for the manuscript as a whole, with a repository[@ref]. If this contains msParts then each of these only needs to contain msIdentifier with an idno

A @class can also be used in msItem with the abbreviated name of the unit type as in Andrist et. al. 2013. [TBD]

The description of the manuscript part is contained in physDesc, the events which affected the history of the manuscript can be added in history while additional informations about the manuscript (e.g. surrogates, further bibliography, etc.) have a specific element for them: additional.

Manuscript Contents

The contents of each manuscript part should be given into the smallest possible unit (msPart) to which they refer to, into an msContent, but they can be added at any level in a hierarchy of parts.

       I) Gädlä Kiros (fols. 5ra-43rb)
      II) Mälkəˀa Kiros (fols. 43va-50va)
     III) Mälkəˀa Kiros (fols. 50vb-51va)
    <msItem xml:id=“p1.1_i1">
       <locus from=“5ra" to="43vb"/>
       <title type="uniform" ref="LIT1488GadlaQ"/>
       <textLang mainLang="gez">Ge'ez</textLang>
    <msItem xml:id="p1.1_i2">
       <locus from="43va" to="50va"/>
       <title type="uniform" ref="LIT0000MalkaK”>Mälkəˀa Kiros</title>
       <textLang mainLang="gez">Ge'ez</textLang>


title should point with a @ref to the identifier of the Work in the Authority list. In the Text file alternative titles in different languages should be given.


msContents can contain a summary of the contents, but these can be structurally described with nested msItems.


msItem will hold information about contents of the manuscript.

 <msItem xml:id=“p1_i5”>
     <textLang mainLang="gez">Ge'ez</textLang>
     <locus from="1ra" to="182vb">1ra-182vb</locus>
     <title type="Complete">I. Sinodos</title>
        <locus from="1ra" n="1"/>
     በስመ፡ አብ፡ ወልድ፡ ወመልድ[sic]፡ ወመንፈስ፡ ቅዱስ፡ አሐዱ፡ አምላክ። ንቅድም፡ በረድኤተ፡ እግዚአብሔር፡ በሠናይ፡ ኅብረት፡ ወበጽሕፈተ፡ ሲኖዶስ፡ ዘአብው፡ ሐዋርያት፡ ቅዱሳን፡ ዘሠርዑ፡ ለአርትዖ፡ ቤተ፡ ክርስዲያን፡ ቅድስት። አሕብሮተ፡ ሲኖዶስ፡ ዘአበው፡ ሐዋርያት፡ ኅሩያን።
     <foreign xml:lang="en">Index of the Statutes or Commandments of the Apostles): In the name of the Father, and the Son, and the Holy Spirit, one God. With the help of God, we begin with the good collection, with the writing of of Sinodos of fathers, the Apostles, which they established for the guidance of the holy Church. The gathering of the Sinodos of the fathers, the chosen Apostles.</foreign>

One msItem can have more than one locus, for instance in composite mss:

<msItem xml:id="p1_i8">
<locus from="94v" to="105"/>
 <locus from="107"/>

If the text of an incipit, explicit, or addition which is quoted in the catalogue record contains text in verse, do use lb with a progressive number in @n for each line.

<incipit xml:lang="gez">
    <lb n="1"/>ኦእግዚአብሔር፡ ጥበበ፡ ጠቢባን፡ ከሃሊ።
    <lb n="2"/> እስከ፡ ለዓለም፡ እምጥንት፡ እንተ፡ ኢትበሊ።
    <lb n="3"/> ድኩመ፡ ሕሊና፡ ገብርከ፡ ወፈራሃ፡ ልብ፡ ኢተሃባሊ።
    <lb n="4"/> እስመ፡ ኀቤከ፡ እግዚኦ፡ ጸሎተ፡ አድኅኖ፡ እጼሊ።
    <lb n="5"/>  አድኅነኒ፡ አድኅነኒ፡ እምኵሉ፡ ቀታሊ።

title will point to a univocally identified work, for msItems which correspond to an identifiable work or can be identified as a work. The id of the work in the authority list file will be given in a @ref

  <title type="complete" ref="LIT1468GadlaK">Gädlä Kiros</title>

you can use @type to give information about the completeness of the text. Allowed values are

  • Complete
  • Incomplete
  • ...

rubric contains the text of any rubric (heading or subheading) attached to a particular manuscript item.

Incipit and Explicit

incipit and explicit will contain a locus and the original text directly within the element. Their language should always be specified, even if identical to the one given in textLang. Translations of the text will have to be into a foreign specifying the language of the translation. More than one translation can be added.

Within the text all TEI markup can be used to indicate mistakes, lacunas, editorial interventions, etc.

The use of explicit follows the TEI definition: "the closing words of the text proper, exclusive of any rubric or colophon which might follow it"

Subscriptions stating that the text is completed, which have not been added later to the text, are encoded as <explicit type="subscription">:

 <explicit type="subscription" xml:lang="gez">ተፈጸመ፡ ነገር፡ ዘጦቢት፡ ወጦቢያ፡ ወልዱ፡፡ 
 በረከቶሙ፡ ትኲን፡ ምስለ፡ ንጉሥነ፡ ኢያሱ፡ ለዓለመ፡ ዓለም፡፡ አሜን፡፡</explicit>

Similarly, supplications written at the same time as the main text should be encoded, depending on their location in the text, as <incipit type="supplication"> or <explicit type="supplication">.

Musical notation

Information about musical notation can be given in <notatedMusic>

 <desc>Extensive musical notation.</desc>
msItem for contents of a work

N.B.: You can nest into a msItem also other msItems which describe part of a work. If the content is partial, you can use @defective with the value true. Default is false and is not encoded.

<msItem xml:id="p1_i5">
 <msItem xml:id="p1_i5.1"  class="content">
       <locus from="1ra" to="2vb">fols. 1ra-2vb</locus>
       <title type="uniform" xml:lang="la">Index of the Statutes or Commandments (təˀəzazat) of the Apostles </title>
       <textLang mainLang="gez">Ge'ez</textLang>
<msItem xml:id="p1_i5.2"  class="content">
       <locus from="3ra" to="40vb">fols. 3ra-40vb</locus>
       <title type="uniform" xml:lang="la">Statutes or Commandments (təˀəzazat) of the Apostles </title>
       <textLang mainLang="gez">Ge'ez</textLang>
Lists of similar Manuscripts

A listBibl containing bibl elements for bibliographical references can be given in each msItem. References to similar manuscripts can instead be given with a refs as described above. Note that informations about these manuscripts should be stored with the manuscripts for coherence, thus a pointer to the manuscript file will be given and that will contain the bibliographical references, date, etc.

If you want to specify within the file a relation between a bibliographic entry in the catalogue list and one of the similar manuscripts, assign an @id to the ref pointing at the manuscript and give a @corresp to the bibl of the catalogue bibliography.

    .... <ref type="mss"  corresp="TuebingenAeth7" xml:id="TuebingenAeth7"/>
    <ref type="mss" corresp="BNabb65" xml:id="BNabb65"/> ...

<listBibl type="catalogue">
    <bibl corresp="#BNabb65">
        <ptr target="bm:Abbadie1859Cat"/><citedRange unit="page">75-79</citedRange>
    <bibl corresp="#TuebingenAeth7">
        <ptr target="bm:Ewald1843Tuebingen"/>
        <citedRange unit="page">179-180</citedRange>

Relations can be listed as well for manuscript in a specific listRelation. This will contain relation elements with @name taken from a list of values in the in the work in progress schema and can be updated.

listRelation is allowed EVERYWHERE.

For example, within a msItem such relations and a description of them should be given within the element filiation

<filiation> This text describes in general the relations to other surviving manuscripts
    <relation name="saws:isRelatedTo" active="ESam002" passive="ESam003">
         <desc>It is always a nice idea to provide a description of this relations with a sentence or two.</desc>

This relations can be much more precise than a simple list of ref elements stating similarity, and this is therefore the recommended way to encode similar manuscripts. multiple relations can be stated if needed, for example if a manuscript A is a copy from two different manuscripts B and C of which one is certainly a translation from a coptic manuscript D, one might have a relation between A and B, A and C, C and D. or parts.

In the @active and @passive attributes, please don't use # as this hinders the possibility to use further specific ids within the record, and to point to a specific msItem, e.g.

It is preferable to have a separate <relation> for each relationship, even if the same relation exists between one and several more items.


































  • ecrm:P57_has_number_of_parts

inside any msItem in the description of the contents a colophon with @xml:id=‘coloph’ can be used to enter the text of the Colophon and a description. This will be entered with a special element colophon within the msItem

  <colophon xml:lang="gez" xml:id="coloph1” >
    <locus from="49rb" n="28"/>
    በአስተሐምሞ፡ በትእዛዘ። እግዚአብሔር፡ ገበርክዎ፡ ለዝንቱ፡ ነገር ።

Bilingual colophons can have a foreign element with @xml:lang=‘...’ of the second language used (see e.g. ESakm010).

Extra information from the catalogues

Information given in the catalogues about manuscripts belonging to other traditions, i.e. Arabic, should not be included in our description.


note can contain any other information which cannot be better structured in the description. It can be used within most elements of a manuscript description.

<colophon xml:lang="gez" xml:id="coloph1”>
   <locus target="#92r"/>
   ተፈጸመ ፡ በዝየ ፡ ኦሪት ፡ ሰመንቱ ፡ ብሔር ፡ በሰላመ ፡ እግዚአ ብሔር ።
   <note>Written in hand <ref target="#Hc" />, cp. also Additiones <ref target="#a1"/>, <ref target="#a4"/>, <ref target="#a8"/>, <ref target="#a10"/>, <ref target="#a12"/>. 
   The note is written over erased punctuation sign./note>


The physical description of a Manuscript or manuscript part is contained in an element called physDesc.

Object Description

objectDesc contains all the information about the object as such. Within the @form you can give one of the following values:

  • Codex
  • Scroll
  • Leporello
  • Leaf
  • Other
Support Description

Into supportDesc a series of information about the material support are given.

  <material key=“Parchment”/>

In material/@key enter the a value from the Writing Material taxonomy proposed from the schema. oXygen should prompt the values from the schema.

extent contains information about the global dimensions of the object. "Outer" indicates the external dimensions, including the board, the cover, and the spine. "Leaf" indicates the size of the leaf. The leaf from which the measurements have been taken can be added in a <note>:

  <measure unit="leaf">183</measure>
  <measure unit="leaf" type="blank">3</measure>
  <measure unit="quire">25</measure>
  <dimensions type="outer" unit="mm">
  <dimensions type="leaf" unit="mm" xml:id="leafdim">
<note corresp="#leafdim">
Data on leaves dimensions 
taken from <locus target="#6r"/>.

Measurements in dimensions should always be given in millimiters.

You can also specify which leaves are blank, adding a locus immediately after the relative measure.

collation can be used to give textual information about how the leaves are arranged.

       <item xml:id="q1">
           <dim unit="leaf">8</dim>
           <locus from="1r" to="8v"/>
3, stub after 6
6, stub after 3
       <item xml:id="q2">
           <dim unit="leaf">10</dim>
           <locus from="11r" to="20v"/>

Give a @xml:id to each item in the list to be able to refer to this unit, using “q” and a consecutive number. You can also nest lists, following the same id pattern used for msPart for example.

If only general information on the collation is provided and no quire structure is given, this information can be stated in a <note> inside <collation>:

The textual sequence of the manuscript is disturbed, see <ref target="#ms_i1.1"/>.

Quires with an uneven dim[@unit] will ALWAYS need clarification of the phenomenon occurring as below. No item view will be loaded if only the following is given

<item xml:id="q2">
           <dim unit="leaf">9</dim>

you need at least the following

<item xml:id="q2">
           <dim unit="leaf">9</dim>
1 added

To notate singletons and stubs, please use the following pattern as in Ethio-Spare

6, stub after 3

Added leaves, where no stub is visible might be recorded as

1, no stub

where the number before the comma is the folio number of the singleton.

Also, you can use with care for consistence the following values

3 replaced, 4 missing, 5 missing, 6 replaced, 9 added 

During preview, this will be analyzed to produce diagrams of each quire and formula notations. Added leaves are shown with hyphens, stubs are not shown, replaced leaves are dot-hyphen lines.

Sometimes you might need to describe the position of a quire using locus elements. please, do this in a note, as the locus element is here reserved for the quire extent.

When a Quire and not just a leaf is added into another quire at some point, you can follow this example

<item xml:id="q11" rend="withInsertion" corresp="#q12">
     <dim unit="leaf">4</dim>
      <locus from="92r" to="99v"/>
       X (<note>between <locus target="#94">94v</locus>
       and <locus target="#99"> 99r</locus>: <ref target="#q12">the
     second codicological unit</ref></note>) 
<item xml:id="q12" rend="inserted"  corresp="#q11">
     <dim unit="leaf">4</dim>
    <locus from="95" to="98"/>
    XI (s.l.: 2, stub after 4; 1, stub after 4)

Here @rend is used in quires and takes two values (free), "inserted in" or "with insertion". The value of the locus in the item, needs to have the extremes of the quire, as in <locus from="92r" to="99v"/> this remains true, and is then clarified in the note as necessary. You can use @corresp to express the linked quire. This assumes that the two quires are described separately and that they are put in some order. You might also decide to put the inserted quire at the end of your list, then using @corresp is very important.

Different numerations of quires are encoded as follows

<item xml:id="q26" n="25"> <num value="20">፳</num> <dim unit="leaf">8</dim> <locus from="195r" to="202v"/> </item>
  • In item/@n we encode the progressive number of quires, assigning "a" as value for initial guard leafs and "b" for final guard leafs. This information was given in Roman Numerals in DomLib, but needs to be in arabic numerals here.
  • The numeration given by the scribe is copied and annotated with a simple num[@value] element.
  • The number of leafs in the quire is given in dim
  • locus provides as usual the extremes.

Inside collation you can use signatures for quire marks and decoration of quiremarks.

   <signatures> Decorated <note>Simply decorated quire marks are
                written in the middle of the upper margin of fols. 3r, <locus target="#11r"/>, 19r, 27r.</note>

You might also use foliation with a list of items in it to describe different numbering systems used.

<item xml:id="fol1">
    <locus from="1r" to="98v"/>
    arabic numbers on the top right corner

<item xml:id="fol2">
    <locus from="1r" to="45v"/>
    ethiopic numbers on the left right corner

<item xml:id="fol3">
    <locus from="1r" to="102v"/>
    quire numbers on the bottom right

You can also add a condition element containing a full markedup description of the condition of the manuscript. Do indicate any way a general value from the State of Preservation Taxonomy for indexing purposes.

<condition key=“deficient”/> 

value of @key is the corresponding id from the state of preservation taxonomy proposed from the schema.

Layout Description

layoutDesc contains the description of how the text is laid out on the page.

Each layout information is given in a layout element therein.

For example for measurement a series of specific purpose attributes and elements can be used

<layout columns="1" writtenLines="18">
   <dimensions unit="mm" rend="pricked ruled">
      <dim type="intercolumn">15</dim>
   <dimensions type="margin">
      <dim type="top">35</dim>
      <dim type="bottom">65</dim>
      <dim type="right">40</dim>
      <dim type="left">15</dim>

If the layout measurement refer to a specific part of the manuscript a locus element can be used within layout to provide the range or a @corresp with the id of the msPart, msItem or quire in the collation, whichever might be relevant.

  <layout corresp="#p.1.1_i1" columns="1" writtenLines="18">


  <layout columns="1" writtenLines="18">
     <locus from=“1r” to=“20v”>

The leaf from which the measurements have been taken can be indicated in a <note> inside <layout>:

  <layout columns="2" writtenLines="26">
                           <locus from="119ra" to="132ra"/>
                              <note corresp="#textarea #margin">Data on text area and margin dimensions taken from <locus
                           <dimensions unit="mm" xml:id="textarea">
                           <dimensions type="margin" unit="mm" xml:id="margin">
                              <dim type="top">15</dim>
                              <dim type="bottom">30</dim>
                              <dim type="right">20</dim>
                              <dim type="left">5</dim>
                              <dim type="intercolumn">10</dim>
  • @columns contains the number of columns on a page.
  • @writtenLines contains the number of lines on a page.

The first dimensions contains the information about the text area. A @rend can be used to list separated by spaces other visible characteristics. Allowed values are:

  • pricked
  • ruled

The second dimension with @type=‘margin’ contains the measures of the margins.

Within a layout you can also use ab as in the following example

 <ab type="ruling" subtype="pattern">Ruling pattern: 1A-1A/0-0/0-0/A.</ab>
 <ab type="pricking" subtype="pattern">Primary pricks are partly visible (s., e.g., fol. 37v).</ab>

Allowed values of @type here are

  • ruling
  • pricking
  • punctuation
  • cruces
  • CruxAnsata
  • ChiRho
  • coronis
  • ...

Allowed values of @subtype here are

  • pattern
  • usage
  • executed
  • dividers
  • ...

ab can contain locus if you need to state precisely where an event occours.

<ab type="punctuation" subtype="dividers">
       <item xml:id="div1" corresp="#p.1">Colon</item>
       <item xml:id="div2" corresp="#p.2">
           Five red and black dot vertical colon. 
           <locus from="30v" n="3"/>
           <locus from="30v" n="11"/>
           <locus from="31r" n="14"/>
           <locus from="31v" n="10"/>
Hands Description

Into handDesc you can have as many handNote as you wish.

<handNote script="Ethiopic" xml:id=“h1” corresp="#p1.2 #p1.3 #a4”>
  <seg type="script"> 15th Century square script?</ab>
  <seg type="ink">Black, red</seg>
  <seg type=“rubrication”>nomina sacra</seg>
  <persName role=“scribe” ref="PRS1691Alib"/>
  <date>Mid 15th - mid 15th-cent.</date>
  <desc>Regular, by a well-trained hand</desc>

Give an @xml:id to each handNote to able to refer to it in other parts of the record. You can also use @corresp to point to msParts or msItems in which this hand is used. @script should take one of the values in the Writing Systems Taxonomy proposed from the schema. You can always also add a locus within handNote to determine where the hand is used. This can overlap with the corresponding parts without any problem.

Inside handNote you can also list abbreviations

 <list type="abbreviations">
       <abbr>ፍ፡</abbr> for <expan>ፍቱሐ፡</expan> (e.g., <locus target="#36ra #36vb #39ra"/>)
        <abbr>ይ፡ ሕ፡</abbr> for <expan>ይበል፡ ሕዝብ፡</expan> (e.g., <locus target="#40va"/>)
Decoration Description

decoDesc contains a description of the various decorative elements found in a manuscript. If multiple elements are present you can provide general observation in summary using it directly into decoDesc. Within this text simply use ref to refer to each decoNote or locus to refer to a folio. If there is a note about all miniatures and a note about all drawings of a certain type, then the sentence in description will point to the decoNotes involved.

            The manuscript’s decorative features are typical of the period. It includes the Eusebian Apparatus 
            (3+7 arrangement) set under colourful stylized arches and followed by a Tempietto. As the pages 
            before the beginning of the Gospel of Mark <locus target="#82v"/>, Luke <locus target="#127v"/>, 
            and John <locus target="#203v"/> contain inventories and notes it is possible that they were originally 
            intended to receive <ref type="authFile" corresp="AT1013EvaPortrait"/>. The miniatures are painted in 
            the <term key="FirstSolomonicStyle"/> and recall examples produced between  
            <date notBefore="1300" notAfter="1450"/>. The decorations appear to have been paintedby the same artist.

N.B. summary can also be used to provide information about the decorative features of the manuscript that cannot be given elsewhere. In the above example, the editor of the decoDesc believes that the manuscripts offers evidence that is relevant to the Art Theme AT1013EvaPortrait even though the theme is not included in the work in question.

Each decorative element in the manuscript is listed in a decoNote as follows:

<decoNote type="miniature" xml:id="d1">
     <locus target="#1r"/>
     <desc>drawing of a cross; black ink.</desc>
<decoNote type="frame" xml:id="d2">
     <locus target="#41r"/>
  <desc>wide, coloured (red and black), horizontal ornamental band over col. 
       a-b, at the incipit of text I.-3), with three lateral pendants reaching l.6 
       whose the central one decorated with a cross; the horizontal band extends to 
      the top of the page(interlaces, crosses).

Available values of @types for decorations are

  • frame
  • miniature
  • band
  • punctuation
  • drawing
  • border
  • diagram
  • frieze
  • illustration
  • initial
  • map
  • marginal
  • mixed
  • other
  • ornamentation
  • paratext
  • printmark
  • publishmark
  • secondary
  • tooling
  • unspecified
  • unwan
  • vignette
  • engraving

Include always a locus[@target] to identify the position of the drawing.

There should be a ATID for each subject represented as miniature/drawing in the manuscript/item when possible (there should also be specific ATIDs for subjects that cannot be confidently identified (e.g. Saints on Horseback, Holy Men, Group of Figures). These latter IDs are to be used only when identification is not possible by means of attributes and/or captions and/or iconography and/or context. For consistency, when editing miniature and drawing @types in <desc> the please try to follow, when possible, the following principles. 1. Order: main subject and related iconographic elements; main iconographic features from left to right and from top to bottom; and secondary elements and background. 2. Generic References and Bibliographic References are given in the ATID, include only references to: works referring directly to this particular set of miniatures; or works relevant to a particular feature of the specific Art Theme that are not included in the general ATID bibliography. 3. Images depicting groups of individuals/saints will have a specific ATID but, when possible, each individual should be identified with the appropriate PersName or persongrp record. 4. Images depicting generical geographical features (e.g. trees; mountains) can be encoded using keywords in the Place types section. Towns, areas, etc. can use PlaceName. If a geographical reference is consistently present in a Theme, then consider including it in the ATID rather than in the DecoNote. 5. Keywords are valuable tools that you can use to provide further information about such features as the style and iconographic elements of a theme (see relevant Wiki section).

Miniatures may include: a) legends that illustrate the general subject matter of the miniature; b) shorter captions, a short sentence or a word, that identify some of the Persons/Objects/Actions in the miniature. Legends should be provided at the end of <desc> inside the corresponding <decoNote> and use <q> captions should be associated to the relevant iconographic element using <foreign>:

<decoNote type="engraving" xml:id="d2">
 <locus target="#125v"/>
  <desc> The <ref type="authFile" corresp="AT1005Crucifixion"/>.
   At the sides of the cross are the Two Thieves, identified by captions as 
   <foreign xml:lang="gez">ፈያታይ፡ ዘየማን፡</foreign> and 
   <foreign xml:lang="gez">ፈያታይ፡ ፀጋም፡</foreign>. The legend at the top of the
   folio presents the
   <q xml:lang="gez">ነዋ፡ በግዑ፡ ለእግዚአብሔር፡ ዘያአትት፡ ኀጢአተ፡ ዓለም።</q>

Manuscripts that contain exclusively miniatures and no text receive the keyword miniatureCollection. Any other keywords relevant to these miniatures, such as for example Hagiography, can be assigned freely.


Always in physDesc, you can add a additions element, which will contain a list of additions. Each item will have a structure as follows. additions should be used only for elements added later to the text! A general note on all additions can be added before list within a note element.

<item xml:id="a1">
    <desc type="acquisition">Acquisition note.</desc>
    <locus target="#40vb" n="30"/>
    <q xml:lang="gez">አጥረይክዋ፡ አነ፡ ገበረ፡ መድኅን፡ በመን{ፈ}ሰ፡ ጥበብ፡</q>
    <q xml:lang="en">I acquired it (this book), Gäbrä Mädḫən, through the spirit of wisdom.</q>
 <item xml:id="e1">
Crude recent notes in pen and in
                pencil throughout the manuscript, above and below the text (e.g. <locus target="#1rv #2r #4r"/> Crude drawings in
                pencil, mixed with crude notes, on <locus target="#45r"/>.</item>

@xml:id will use “a” and a consecutive number. In case of other contents (varia) this will have an id with “e”.

If the language of the addition is known but no excerpt is copied, use an empty q as in the following example from BLadd16187.

<q xml:lang="am"/>

You can specify the type of addition in the desc element, using one of the value provided for the @type.

If among these also marginal additions or additions on guard leafs need to be added, add to item also a @rend with value ‘marginal’ or ‘guardLeaf’.

If the type of document is known then there should be a narrative entity to which a @ref can be given.

<item xml:id="a2" xml:lang="am" ref="NAR0001gwǝlt">
   <locus target="#3v"/>
   <p>Land donation (gwǝlt) of <persName ref="PRS8344sahlaDe">Śāhlā Dǝngǝl</persName> (A.D. 1832-40, and again 1841-55) to the priests of <placeName ref="LOC2550Darasg">Darasge</placeName>, in Amharic. </p>

If a work with its own work ID is added later to a manuscript, it will be added as an addition specified with the type "GuestText" having its work ID in the description:

<item xml:id="a10">
                  <locus from="115v"/>
                  <desc type="GuestText">
<title type="Complete" ref="LIT3957SalotaB">ጸሎተ፡ ቡራኬ፡ ዘቅዱስ፡ ባስልዮስ።</title>
                  <q xml:lang="gez">
                    <desc>The following is the incipit:</desc>
                 ኦሥሉስ፡ ቅዱስ፡ አብ፡ ወወልድ፡ ወመንፈስ፡ ቅዱስ፡ ባርክ፡ ዲበ፡ ሕዝብከ፡ ፍቁራን፡ ክርስቶሳውያን፡ በበረከት፡ ሰማያዊት፡

The distinction between additions (or better additiones) and Varia comes from the Ethio-SPaRe guidelines. _Additiones _are very often written in a secondary hand, and they are later than the main text(s) of the Ms. They mark the so-called "secondary life" of the manuscript as the latter is used as a repository of documentary information, i.e. a true archive. _Additiones _ that can be tyipically found in a Ms. are the following: "AcquisitionNote", "Admonition", "CalendaricNote", "CommemorativeNote", "Comment", "Condemnation", "Correction", "DonationNote", "Doxology", "Excerpt", "Excommunication", "Exhortation", "FoundationCharter", "Genealogy", "Gloss", "GuarantorsList", "Headband", "Inventory", "Invocation", "LandGrant", "Letter", "MagicFormula", "MagicText", "OwnershipNote", "Poem", "PoemArke", "PoemQǝne", "PoemSalam", "ProtectivePrayer", "PurchaseNote", "ReceiptNote", "Record", "RecordDistribution", "RecordGuarantors", "RecordLitigation", "RecordReconciliation", "RecordTransaction", "ScribalNoteAssigningLang", "ScribalNoteBequeathing", "ScribalNoteCommencing", "ScribalNoteCommissioning", "ScribalNoteCompleting", "ScribalNoteOrdering", "ScribalSignature", "ScribalSupplication", "StampExlibris", "Subscription", "Supplication", "Tailband", "TransactionNote", "Unclear" or "foundation". These value are provided for the @type in the desc element. When describing an addition, please

  • language (if different from Ethiopic: in Amharic, in Tǝgrǝñña);
  • hand (by a recent/old hand, by the same hand as that of the main text, by the same hand as that of another addition, etc.); “in a secondary hand” should be avoided: it is obvious;
  • ink (in recent ink, hardly legible because of the faded ink, etc.).
  • readability (hardly readable, washed/scratched out etc.);
  • some additiones can be rhymed.

Under Varia, Ethio-SPaRe used to list the following phenomena, which simply concern the copying process or are more trivial than additiones:

  • the title of the work written on the front board or cover, sometimes with blue or black paint or ink;
  • the internal number of the book in the monastic library;
  • the midpoint in the "Psalms of David" (in Ps. 77, after the word በደመና፡);
  • the presence of stamps of the owning institution;
  • part of the line is taken up or taken down (e.g., in the Psalters), ff. ....
  • glosses written in the margins of ff. ....
  • the so-called tämälläs-sign, in the shape of a cross (as +) or two perpendicular lines (as ┴ or├), indicating the place where the omissions (written in the upper or lower margin) belong: ff. __.
  • omitted letters, words, lines or passages inserted interlineally, in the margins, at the beginning/end of the line (ff. ...).
  • erasures: ff. ....
  • corrections written over erasures, in the main hand or in a later recent hand: ff. ....
  • erasures marked with thin black and red lines: ff. ....
  • corrections immediately indicated by the scribe through thin series of dashes above and below the letters or the words: ff. ....
  • spaces for rubricated names left unfilled: ff. ....
  • unfilled spaces marked with thin black lines: ff. ....
  • cues for the rubricator written in the margin or in the interocolumn (e.g., small digits): ff. ...
  • erased names or names secondarily written over erased names (typically in the supplication formulas): ff. ....
  • the scribe reduced the size of his handwriting to accommodate the end of the line: ff. ....
  • the scribe used compressed script ("compressed" means two lines in a space originally meant for one single line), often to accommodate lines written over erasure: ff. ....
  • monthly or daily readings are usually indicated by writing the names of the months and weekdays in the upper margin of the folio, sometimes within frames.
  • pen trials, notes crudely written in pencil or pen;
  • devotional images sewn between ff. ....
  • scribbles and doodles.
Binding Description
   <binding contemporary="false" xml:id="binding">
     <decoNote xml:id='b1'>Wooden boards (recently) covered with
                reddish-brown tooled leather. </decoNote>
     <decoNote xml:id='b2'>Four pairs of sewing stations.</decoNote>
     <decoNote xml:id='b3' type="bindingMaterial"><material key=“wood”/></decoNote>
     <decoNote  xml:id='b4' type="Endbands">...</decoNote>
      <decoNote xml:id='b5' type="Other">
          Parchment guards (possibly of the same
                origin as <locus target="#183"/>) were used to sew together <locus from="1" to="2"/>.

binding can take @notBefore and @notAfter and takes a @contemporary with the following values:

  • false
  • inapplicable
  • unknown
  • true
  • partly

oXygen should prompt the values from the schema.

decoNote elements in Binding Description MUST have a @xml:id starting with the letter 'b' followed by a progressive number, e.g. b1, b2, b3.

Some of these decoNote can take a special form (which was established to host the data coming out of EthioSpaRE).

  • The first decoNote, with @xml:id='b1' should be a general note and remain untyped.
  • Inside the special decoNote[@type=‘bindingMaterial’] you can add the element material with a @key equal to one of the values of the corresponding taxonomy from the schema. oXygen should prompt the values from the schema.

Other defined types for decoNote are

  • Headbands
  • Tailbands
  • Other

You can always further specify this information within one of the decoNote elements within binding.


you can use sealDesc and seal for seals reported in the catologue description. This should only be used for seals made out of wax and not be confused with stamps and ownership marks, which are encoded under "extras".

  <seal xml:id="s1">
    <p><locus target="#3r"/> ....</p>


Events in the history of a manuscript part can be described in history. These events are grouped in:

  • acquisition
  • origin
  • provenance
  • summary
  <origPlace><placeName ref="LOC4705MayRub"/></origPlace>
     <origDate notBefore="1872" notAfter="1876">Mention of <persName ref="PRS10378Yohanne">King Yoḥannәs IV</persName> (<date>1872-89</date>)
     on <locus target="#43r" n="8" corresp="#coloph"/>, and of <persName ref="PRS2223Atnatewo">Metropolitan ˀabunä ˀAtnatewos</persName> (in tenure <date>1869-76</date>) on <locus target="#43r" corresp="#coloph"/>. Besides, the colophon refers to the year of John the Evangelist.</origDate>

origDate contains as iso values in @notBefore and @notAfter (or only one of them, if that’s all we have) the range of dates in the Gregorian calendar for the production of the manuscript.

If a precise date is know, please use @when instead of @notBefore and @notAfter.

If the date is documented in the manuscript add @evidence='internal-date'. You can also use this attribute in a date element , e.g. marking up the colophon text. If the date is known by another source, such as the date of the reign of a ruler or paleography, please choose the relevant value prompted by the schema for @evidence.

It is possible to add more than one <origDate> if alternative datings for the manuscript are provided in the catalogue or literature. In these case the source for each date should be specified in a <note> inside the respective <origDate>. The person responsible for the dating can be indicated with @resp inside <origDate>:

<origDate notBefore="1540" notAfter="1699" resp="PRS4805Grebaut PRS9530Tisseran"> 
<note> Date according to <bibl><ptr target="bm:GrebTiss1935Codices"/> <citedRange unit="page">187-190</citedRange></bibl>. </note> 
<origDate notBefore="1550" notAfter="1599" resp="PRS3182ContiRo"> 
<note>Date according to <bibl><ptr target="bm:ContiRossini1927Inediti"/> <citedRange unit="page">516</citedRange></bibl>.</note> 

Please, note that almost anywhere text can be annotated, for example with a date element which can take the same attributes as described above. If marking up a sentence, for example "in the early 1920s-1930s" there is no need to remove the text just mark it up, as there is no way to consistently output all options for a date value.

in the <date from="1620" precision="low" to="1635">1620s - early 1630s</date>


This is used for any other information related to the manuscript. Note surrogates where for example microfilms or photographs should be recorded.

        <listBibl type="catalogue">
         <bibl><ptr target="EMML"/><citedRange unit="volume">V</citedRange><citedRange unit="page">354-356</citedRange></bibl>
     <custEvent type=“restoration” subtype=“modern”/>
   <bibl corresp="EMML1234"/>
   <bibl corresp="ES4567"/>
    <bibl>Ms. previously unknown</bibl>

For custodial history events use @subtype with one of the suggested values from the corresponding taxonomy. oXygen should prompt them from the schema.

Here the msDesc or msPart ends. Check that all is well nested and that all elements up to fileDesc are closed.

Profile Description


Important people related to the manuscript can be listed in particDesc with a list of pointers to names with a specific role. Allowed value of role here are

  • donor
  • scribe
  • other
  • ...

Here it is enough to enumerate the scribe, owner, donor and patron, not everyone mentioned in the manuscript. If this information has already been stated elswhere in the description - which is preferable - it doesn't need to be repeated.

   <person><persName ref="PRS2232AubertR" role="donor"/></person>
   <person><persName ref="PRS2233Audon" role="scribe"/></person>
   <person><persName ref="PRS2233Audon" role="other"/></person>

You can also be more precise and assign a @corresp to the persName if for example a person has commissioned only a specific part.

  <person><persName role="patron" ref="PRS9170TaklaIy" corresp="#p1"/></person>


These keywords are used for the whole record. You can point use one or more of the values provided for @key. You can assign as many keywords as needed. This means that it is not necessary to limit the keywords to the principal text, but keywords for any additional or minor texts contained in the manuscript should also be assigned. Always assign a period keyword, using the keywords referring to Ethiopian literary periods only in records actually belonging to the Ethiopian tradition. The keywords "ChristianLiterature", "IslamicLiterature" and "BetaEsraelLiterature" should be assigned to all records belonging to these traditions.

<term key=“ChristianLiterature”/>
<term key=“Gon”/>
<term key=“Liturgy”/>


within langUsage all languages used in the file should be defined. It is not a problem if one of these statements is not relevant to the current record, but if another language is used, it should be defined.

    <language ident="en">English</language>
    <language ident="it">Italian</language>
    <language ident="gez">Gəʿəz</language>

If you add somewhere e.g. <title xml:lang="am"/>, there should be a language element in langUsage like the following

<language ident="am">Amharic</language>

the values of @ident here are restricted, so you should not be able to add e.g. amh as value for @ident and consequently use it in the file as value of @xml:lang.

Regardless of the validity of this transliteration, Gəʿəz is always referred to by "Gəʿəz" to ensure consistency.


It is very important, at the end of each section of work of improvement of a record to record a changein the revisionDesc element of the teiHeader. In the attribute @who enter your id, in the @when enter the date of the change, in the format YYYY-MM-DD. Inside the element, briefly describe the changes you have made.

   <change when="2010-12-04">registered</change>
   <change who="DN" when="2015-09-11T19:18:15.444Z">last edited</change>
   <change who="PL” when="2016-03-10T17:51:23.806Z">transformed from mycore to TEI P5</change>

When you are done with a file, please, add a change element with content completed. If you have reviewed and corrected a file which is now ready and published, please add a change with content reviewed


To link to an external set of images, use the @facs attribute. In msDesc/msIdentifier/idno add a @facs containing the base url.

<idno facs=""/>

A link to a digital copy of a manuscript can be added to the main identifier:

<idno facs="">BnF Éthiopien 150</idno>

then in locus together with the other attributes add again the @facs containing only the reference to the first page in the range indicated exactly as it appears in the url for that page (so, not the folio number!).

    <locus target="#30v" facs="f62"/>
    <locus target="#64v #65v #110r" facs="f134 f136 f225"/>
    <locus from="31" to="33" facs="f62"/>
    <locus from="31" facs="f62"/>

Links to images hosted on our server in the additional hard drive, should use the relevant part of the path, i.e. BMQ/003/BMQM-003 for example. if the number of the photos are separated with two underscores, then KAE/011/KAE-011_

Links to the images of the Vatican Library MSS are now supported, by adding in the idno/@FACS the manifest.json url which can be found in the digivatlib on the left info page, e.g. to add link to loci, add @FACS attributes. Please do not forget that the folio number (in the following example 3r) does not necessarily corresponds to the URL number (here 0006).

<msItem xml:id="ms_i1.1.1">
<locus target="#3r" facs="0006"/>
<title type="complete">Maqdǝma wangel (Introduction to the Gospels)</title>

Some Manuscripts have images published which are not accessible via IIIF, please, in these cases add inside TEI, between teiHeader and text a facsimile element

        <graphic url="[mets]"/>

You can also add the exact link to a page in a @facs in each pb inside the transcription of the manuscript.

@corresp and @ref

  • if you point to an internal id, use #
  • if you point to an external id don't use #. That is why witness does not take that, as @active, @passive inside relation should not. In fact you want to be able to point to ids within the records you are pointing to.


TEI can express much more precisely where the uncertainty about a certain information provided is using the element certainty.

     <title type="complete" ref="LIT1716Kidanz"><certainty locus="value" match="../@type" cert="low"></title>

instead of

     <title type="complete" cert="low" ref="LIT1716Kidanz"/>

will express that the uncertainty is about the completeness of the work referenced, not about the work generally. Please refer to the TEI guidelines for further guidance and example of how to use this element. Note that an alternative for the aspect of the markup in question can be given in assertedValue:

<persName ref="PRS11436Cyriacus" role="owner"><certainty locus="value" match="../@role" assertedValue="other" cert="low"/>

Text Encoding

TEI records are for Text encoding. We record in Manuscripts records the Transcription of the text on the catalogued manuscript, when available, whilst the edited text or the theoretical structure of it should be recorded into the Work Entity. Text is entered in the body of a TEI file under into a div[@type='edition'] . Each text part should be contained into a div. divs should always be well nested one into the other. These will always take a @subtype='textpart' and a @subtype with the desired value. There should always be also a @n identifying for example chapter 3 as well as a @corresp pointing to the appropriate msPart or msItem

<div type="textpart" subtype="chapter" n="7" corresp="#ms_i3" xml:id="chapter7" xml:lang="gez">
      <locus target="#103"/>
      <l n="1">በቀዳማዊ፡ ዐመተ፡ መንግሥቱ፡ ለበልጣሰር፡ ንጉሠ፡ ከለዴዎን፡ ሐለመ፡ ዳንኤል፡ ሕልመ፡ ወርእየ፡
        ርእሶ፡ በውስተ፡ ምስካቡ፡ ወጸሐፈ፡ ሕልሞ። </l>
      <l n="2">አነ፡ ዳንኤል፡ ሐለምኩ፡ ወርእኩ፡ አርባዕተ፡ ነፋሳተ፡
        ነፍኁ፡ ውስተ፡ ባሕር፡ ዐቢይ፡ </l>
      <l n="3">ወዐርጉ፡ አራዊት፡ አርባዕቱ፡ ዐበይት፡ እምባሕር፡ ኅቡረ፡ ወቀዳማዊ፡ </l>
      <l n="4">ከመ፡ አንበሳዊት፡ ወባቲ፡ ክንፈ፡ ወክንፊሃኒ፡ ከመዝ፡ ንስር፡ ወርኢኩዋ፡ ውስከ፡ ተመልኀ፡ ክንፊሃ፡
        ወተንሥአት፡ ወቆመት፡ በእግረ፡ ሰብእ፡ ዲበ፡ ምድር፡ ወልበ፡ ሰብእ፡ ተውህበ፡ ላቲ።</l>

Note that after each word there should be a separator or punctuation without space.

locus can be used but will be printed as (Excerpt from... ). If the text contains biblical verses we will use l as in the example above. Also lb can be used to mark linebreaks. Be carefull: ab should be only in the lowest level div nesting!

Text Diplomatic Transcription features


Text Structure

Use an empty pb for each page break. this takes a @n which will have the folio number, e.g. <pb n="41r"/> If there are columns, use the element cb as in this example <cb n="a"/>

Hand Shifts

Where a new hand begins insert an element handShift with a @new pointing to the id of the hand in handNote

 <handShift new="#h2"/>

Additions, notes, etc.

For text related additions and marginal notes not reported under additions, use add:

 <add place="top" hand="#h5">ገድለ፡ ላሊበላ፡ ዘቅዱስ፡ መድኃኔ፡ ዓለም፡</add>

The @place has specific values, as the schema will tell you. Please read carefully when choosing.

Always point to the handDesc xml:ids of each handNote in a @hand attribute if you know who added a specific text or letter.

In text

You might have to encode excerpts or text. Please see BLorient719.xml for an example including some of these markup, or BDLaethe3.xml .

description tag usage current preview
red inks hi[@rend='rubric'] text red
later correction add[@hand, @place] /text/
text deleted removing material and rewritten subst/del[@rend'erasure'] and add[@place'overstrike'] the top layer of the parchment was removed and the text was written again. {text}
marginal signs emph [word]
a sign, pointing to the place where the text should be added <gap reason="illegible" unit="chars" quantity="1"/> you can vary unit and quantity. for characters, you will get a plus sign for each missing character +
spaces left blank for rubrication, never filled <space reason="rubrication" unit="chars" quantity="1"/> (@quantity @unit left for rubrication)
underlined text hi[@rend="underline"] underlined
unclear letters or words unclear word/letter?
undeciphrable words orig text grey
Words or lines skipped in omitted by the editor in transcription gap[@reason='omitted'] .....
Text omitted from the edition by the editor, for whatever reason (brevity, context, language, etc.) gap[@reason='ellipsis'] (...)
Lost text gap[@reason='lost'] used to indicate a material loss or a portion of text that is to be hypothesized for material reasons ("it was there!"), but which is lost for loss of the material [- ca. @quantity @unit -]
text supplied by editor supplied[@reason='lost'] are used to indicate some portions of text that were restored on the basis of philological and textual considerations [the supplied text]
text supplied by editor supplied[@reason='omitted'] are used to indicate some portions of text that were integrated on the basis of philological and textual considerations <the supplied text>
text supplied by editor supplied[@reason='undefined'] please do not use this unless really necessary there are other nice options. [the supplied text (?)]
text supplied by copist in the manuscript ቃል፡ ዘ<add place="inline"><supplied reason="omitted">ማ</supplied></add>ርቆስ፡ ፼ወ፯፻ see above for supplied and below for add <the omitted text> with a tooltip giving info on the position
correction choice[sic[@resp] and corr[@resp]] this will print the corrected version in bold and a tooltip which reports the original version and the author of it clicking on the correct form will toogle the other
written upside down hi[@rend='reversed']
written in negative writing hi[@rend='negative']
text erased (scratched with a knife or similar) del[@rend=erasure] 〚deleted text (when still legible)〛
text crossed out with red or black ink, a line or dashes del[@rend=strikethrough]
text marked with dots above or below the line del[@rend=exponctuated]
text deleted or washed out, the material is untouched del[@rend=effaced]
text added above add[@place=above] all these will also have a popup indicating where the text was added
text added below the line add[@place=below]
text added at the bottom add[@place=bottom]
text added within the body of the text add[@place=inline]
text added interlinear add[@place=interlinear]
text added at the left add[@place=left]
text added in the margin add[@place=margin]
text added mixed add[@place=mixed]
text added on the opposite, i.e. facing, page add[@place=opposite]
text added on the other side of the leaf add[@place=overleaf]
text written over previously deleted text add[@place=overstrike]
text added at the right add[@place=right]
text added at the top add[@place=top]
text added at an unspecified place add[@place=unspecified]

TEI is made for this and there are a number of other things which can be much better and more precisely encoded.

When the record does not yet contain any text transcription, but information about some of the above listed phenomena is available, this should be stated temporarily in a note with a commentary for future changes in the respective msItem:

<note>This prayer is written in negative writing. </note> <!--Information belongs to text transcription-->

Note, however, the different usage for the information about rubrication.

Critical Apparatus

Typically a Work file, but also manuscript transcriptions can have critical apparatus.

We follow the TEI guidelines for Location Referenced External Apparatus when adding apparatus notes to manuscript transcriptions. @wit shall contain the id of associated to the manuscript in witness. This is usefully also the abbreviation of the manuscript to be used in the apparatus.

Follow instead for critical edition (i.e. text in a work record) the inline parallel segmentation method, pointing to the ids of witnesses in that record.

a sample file is the Chronicle of Galawdewos

To separate different types of apparatus notes, use the attribute type.

Transliteration Principles

For an overview of transliteration principles, please see: [Transliteration Principles]

summary on the use of @ref and @corresp

internal references

  • <ref target="#q12">
  • <ref type="quire" corresp="#q12">
  • @corresp="#id" (# is the id anchor) points always to a @xml:id WITHIN THE DOCUMENT

otherwise if @corresp is used to point to an element inside another document with an id, as for witness and div in a work record, then it does not need an initial #

element @ref @corresp
ref X
witness X
div X
title X
persName X
placeName X
origPlace X
repository X
country X
settlement X
region X
Clone this wiki locally
You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.
Press h to open a hovercard with more details.