The encoding of Ibsen's manuscripts for "Emperor and Galilean"

poster / demo / art installation
  1. 1. Hilde Bøe

    Henrik Ibsen's Writings

  2. 2. Ingvald Aarstein

    Henrik Ibsen's Writings

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

The encoding of Ibsen's manuscripts for "Emperor and


Henrik Ibsen's Writings


Henrik Ibsen's Writings


University of Tübingen







Henrik Ibsen (1828-1906) left a manuscript material that covers approximately
20.000 pages. All of these manuscripts are now being prepared for
publishing. For the drama texts there are prompt copies; parts; printer's
copies; final fair copies; copies of printed editions revised by hand and
used as printer's copy for the next edition; drafts that range from ideas
for the whole play on a piece of paper to complete fair copies that have
been revised so thoroughly that they can no longer be considered to be fair
copies. Then there are drafts and fair copies of the poems, and finally
Ibsen's letters.
Our situation is surely not unique. Projects that are publishing the complete
writings of a modern writer are sure to encounter the same range of
manuscripts, and will also, when using the TEI
Guidelines, encounter phenomena in the manuscripts that are not
covered by chapter 18 "Transcription of Primary Sources". The Guidelines state that their "recommendations are not
intended to meet every transcriptional circumstance likely to be faced by
any scholar" (quoted from the introduction to chapter 18). Still we would
find it useful if the Guidelines did include
recommendations for some of these transcriptional circumstances, or if a
documentation of best practise was included. For this reason we would like
to present the problems we have encountered and the solutions we have chosen
in the encoding of Ibsen's manuscripts for Emperor and
Galilean (1873). We feel that it is appropriate and useful to
present the whole process of encoding a complex manuscript, and the
questions this process has raised.

The manuscripts for «Emperor and Galilean»
The planning and writing of Emperor and Galilean
took almost ten years. The play was published in 1873, and consists of two
dramas, both with five acts; the first edition covers 512 pages. The content
of the drama is the life of the Roman emperor Julian the Apostate. Though
later not thought to be one of Ibsen's best plays, Ibsen himself considered
Emperor and Galilean to be his most important
There are six manuscripts for Emperor and Galilean.
Four of these (NBO Ms. 8to 1937a, NBO Ms. 8to 1937b, NBO Ms. 8to 1937c, NBO
Ms. 8to 2821:14) are small fragments containing drafts and notes. These are
all at the National Library in Oslo. Then there is the printer's copy (Kgl.
Bibl. Kbh. Collin 262, 4to, II, 3), which is at the Royal Library in
Copenhagen. The last one, also at the National Library in Oslo, and the one
we would like to present here, is NBO Ms. 4to 1111.
NBO Ms. 4to 1111 consists of notes, drafts and unfinished fair copies, bound
by the library in one large volume of about 700 pages. Smaller notes and
fragments have been placed first in the volume, larger drafts and temporary
fair copies have been placed at the end. This organization is new, and has
probably been done during the binding. Earlier it has been ordered
differently; a red pagination (supposedly done by a librarian) is sign of
this earlier order. In the volume the red pagination is no longer in its
original order. Ibsen himself has left us few clues as to how the material
should be organized; there are few dates and only some parts are paginated.
We do not consider the current organization of the manuscript as final,
since the volume has been made after Ibsen's own time. Our first task is
thus to organize the manuscript according to our project's principles. We
have decided on chronology/genesis, where this is possible to decide. When
we have no direct clues in the material as to what belongs where we will
have to make qualified guesses based on content analysis.
The next question is the question of typology. This is important for
searchability and for the possibilities of organizing the electronic
edition. As said, NBO Ms. 4to 1111 consists of a number of different types
of manuscripts. There are two sides to this. First we have to decide how
detailed the typology should be. Secondly we have to find appropriate terms
for the types of material in the manuscript (cfr. facsimile below). There is
not yet a fixed typology in this area. Scholars are not always using terms
like notes, drafts, fair copies etc. in the same manner. This we think holds
for both Norwegian and international scholars. The TEI
Guidelines keep themselves out of this discussion (for obvious
reasons) and leave the question of typology to their users. Our project has
developed a typology that fits the Ibsen material. This typology might be of
interest to others as well. (The facsimile below contains a working
manuscript with drafts of speeches, and notes in prose.)





lt;div type="workingManuscript">

<div type="act" n="1-3">

<sp who="JULIAN">


<sp who="GALLUS">

<spOpener><speaker>Pr. G.</speaker></spOpener>


<sp who="JULIAN">

<spOpener><speaker>Pr. J.</speaker></spOpener>





<lb/><figure type="bar"></figure>

<div type="notes">

<div type="act" n="1-4">







Then there is the question of the basic encoding of this material. We have
decided on a structure of groups of texts (cfr. the TEI
Guidelines 7.3). We also use unnumbered <div>
elements rather than numbered ones since we have found the same types of
texts on different levels (cfr. the TEI Guidelines
7.1.3). At the more detailed level of the encoding, we will have to consider
for instance what encoding is appropriate for text fragments that later turn
up as parts of speeches. How should these be encoded, as speeches or prose?
They usually lack speakers and stage directions. To what extent should we
link these fragments to their later use in speeches?
It will be possible to link the manuscript material both internally and
externally (to the rest of the text witnesses to Emperor
and Galilean), and it will also be possible to organize the
material according to other principles than ours. As shown in the encoded
example above we supply the outer <div> level
(type="workingManuscript/notes") with a type attribute holding the
appropriate value for the text type. The inner <div> level
(type="act") contains type and n attributes. The value in the n attribute
records which part of the drama and which act (within this part) the text
belongs to. The type attribute records which textual level (e.g. "act") this
piece of text belongs to. With this information we can for example
reorganize the material according to text type or according to the
chronology of the first print. We can also present comparisons of manuscript
and e.g. the first print through linking.
In addition to the questions mentioned above, there are a few minor problems
to solve. First, some of the bits and pieces of the manuscript are
unfinished; the text ends in the middle of a sentence. We will of course
check that the missing parts really are missing, and put them in the correct
place if they are located somewhere else in the manuscript. Secondly, we
have encountered what is for us a new phenomenon. There are a few lists
(cfr. facsimile below) in this manuscript, listing either sources (with page
numbers) that Ibsen has read and then used in his writing, or listing
themes/scenes (also with page numbers) that he wants to include. As he has
checked his list he has crossed out (some, but not all of) the items. How
should we record this? These deletions seem not to be standard manuscript
deletions, yet we consider using the <del> element with the
rend and type attributes to record the rendition and the unusual purpose of
the overstrike.

<del type="checkList" rend="overstrike">...</del>
The manuscript also includes graphics, illustrations (cfr. facsimile below)
and calculations (this manuscript reveals that Ibsen among other things
invested some of his money in train stocks, cfr. facsimile below). There is
also writing upside down on some pages (cfr. facsimile below). So far we
have decided that upside-down writing and calculations should be recorded in
the appropriate part of the <teiHeader>. Graphics and
illustrations are encoded as usual as <figure>.
This is work-in-progress which hopefully by summer 2002 will be well on the
way. We are very interested in hearing what experience others have with
similar questions, and we hope that our work may be of interest to others
working in this field.

Ibsen, Henrik: Emperor and Galilean. The National Library of Norway Ms. 4to
1111. Reproduced from the database at Centre for Ibsen Studies
(), University of Oslo.


Guidelines for Text Encoding and Interchange

Chicago and Oxford
Text Encoding Initiative

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

Conference Info

In review

"New Directions in Humanities Computing"

Hosted at Universität Tübingen (University of Tubingen / Tuebingen)

Tübingen, Germany

July 23, 2002 - July 28, 2008

72 works by 136 authors indexed

Affiliations need to be double-checked.

Conference website:

Series: ALLC/EADH (29), ACH/ICCH (22), ACH/ALLC (14)

Organizers: ACH, ALLC

  • Keywords: None
  • Language: English
  • Topics: None