The Encoding of Time in Manuscripts Transcription: Toward Genetic Digital Editions

Writing is a process that occurs in time. This simple and
obvious consideration involves many issues both from
a theoretical and practical point of view. A critical evaluation
of timing is really crucial in the case of modern authors'
autograph draft manuscripts because different layers of
corrections, deletions and additions can give insight into an
author’s way of working, key to the interpretations of his/her
works and the evolution of the author’s Weltanschaung, as
highlighted in genetic criticism over the last few years.1
Medieval manuscripts copied by one ore more scribes are also,
of course, the result of a process that occurred in time, but the
different kind of authorship involved in such cases seems to
involve a difference in the evaluation of the cultural weight of
recorded variants.
When a scholar inspects a written text, especially a manuscript,
s/he has in his/her hands the final result of that process and can
choose whether to approach it from a
codicological/documentary point of view, “photographing” the
resulting product, or from a genetic point of view, trying to
describe the teleological flow of authoring. The first approach
is more typical of scribes’ copies, the latter of autographic
(draft) manuscripts.
Practically, time based editions are really difficult to represent
in printed works, because of the bi-dimensionality of paper
sheets.
On voudrait représenter dans la bidimensionalité des pages un
processus génétique dont on s’est pourtant appliqué à montrer
que sa proprieté est d’ajouter à l’écrit, qui est bidimensionel,
une troisième dimension, qui est celle du temps!
(Gresillon 1994:121)
In a digital framework bi-dimensionality can be overruled by
a hypertextual/multimedia approach that can allow the creation
of a more flexible context for the presentation of a genetic
edition. The problem of a time based encoding has been
discussed in several circumstances (i.e. TEI Manuscripts SIG
Meeting Report 01, Vanhoutte 2002 that suggests the usage of
the markup solution employed in the transcription of speech
for the purpose), but – up till now – a coherent encoding model for such editions has not been proposed. And for a good reasons:
while it is possible to describe a relative-timed process, it is
very complicated, if not impossible, to draw a general
absolute-timed framework.
Analysis and time-based encoding of
authorial interventions
Let us consider a couple of examples taken from page 3595
of Zibaldone of Giacomo Leopardi.
Fig. 1
In this line we can detect two corrections: 1) deletion of the
line under “si” and 2) deletion of “dotti” corrected into “denti”.
We can try to draw the timing of the creation of the segment
as follow:
Time 1: writing “che e’ si rechi a dotti”
Time 2: deletion of “dotti”; consequent writing of “denti”
Time 3: underlying of “si”
Time 4: deletion of the line under “si”
Time 5: underlying of “rechi” and “denti”
Time 6: writing of “l’un d’essi cibi”
Other timetables are also possible, but let’s assume this is the
more probable one. The text can be encoded using a TEI-based
mark-up with the help of a new global attribute (@time),
intending that it’s just an example to help the conceptualization
of the problem: <seg time="1">che e' <del
type="underline deletion" time="4">
<hi rend="underline" time="3">si</hi>
> </del>
<hi rend="underline" time="5">rechi a'</hi>
<del type="overstrike" time="2">dotti</del>
</seg>
> <seg time="2">
<hi rend="underline" time="5">denti</hi>
</seg>
<seg time="6">l'un d'essi cibi</seg>
Such transcription tries to model the real flow of writing, but
such a model may not be workable. In fact, it will fragment the
flow of the plain writing in potentially infinite pieces. To
simplify it, we can assume as Time 0 (default) the time of the
normal plain writing flow, timing just editorial interventions.
The schedule will then be modified as follows:
Time 0: writing “che e’ si rechi a dotti”
Time 1: deletion of : “dotti”
Time 0: writing of “denti”
Time 2: underlying of “si”
Time 3: deletion of the underline under “si”
Time 4: underlying of “rechi” and “denti”
Time 0: writing of “l’un d’essi cibi”
A further simplification is also possible: assuming that – in
genetic criticism terms – so-called “writing variant” (deletion
of a single word substituted by another that immediately follows
on the same line) occurs during the normal writing flow, the
following model can be drawn:
Time 0: writing “che si rechi a dotti”; deletion of : “dotti”;
writing of “denti”; writing of “l’un d’essi cibi”
Time 1: underlying of “si”
Time 2: deletion of the underline under “si”
Time 3: underlying of “rechi” and “denti”
This will be the consequent new encoding: che e' <del
type="underline deletion" time="2"> <hi
rend="underline" time="1">si</hi></del>
<hi rend="underline" time="3">rechi a'</hi>
<del type="overstrike">dotti</del> <hi
rend="underline" time="3">denti</hi> l'un
d'essi cibi
The last possibility should not imply that any inline correction
is to be considered as done in Time 0, but just the one followed
by the correction. In fact, in the case of a deletion of an adjective
or of any other word not essential from a syntactical point of
view, the correction can occur in any time.
Fig. 2
This passage can be timed as follow:
Time 1: writing “che di due Eroi a”;
Time 2: deletion of “a”;
Time 3: writing of “quanto si voglia pari di”
Time 4: interlinear addition of “o più” after “Eroi”
Time 5: deletion of the addition
Time 6: interlinear addition of “o più” after “due”
Or, in the simplified version, as follow:
Time 0: writing of “che di due Eroi a”; deletion of “a”;
writing of “quanto si voglia pari di”
Time 1: interlinear addition of “o più” after “Eroi”
Time 2: deletion of the addition
Time 3: interlinear addition of “o più” after “due” The encoding model (simplified version): che di due <add
place="intralinear" time="3">o più</add>
Eroi <del type="overstrike">a</del> <del
type="overstrike" time="2"><add
place="intralinear" time="1">o
più</add></del> quanto si voglia pari di
By applying different stylesheets to the encoded texts, it will
be possible to show all the different stages and to give the user
the possibility of browsing among them.
Relative or absolute?
The two above examples occur on the same page: shall we
then consider Time 1 of the first example the same of
Time 1 of the second example? The answer should be: no; very
little can be said about the timing of editorial/authorial
intervention in two different segments. The possibility of
establishing an absolute timing for correction is applicable only
where we have strong palaeographic evidences or authorial
declarations dating or describing a revision.
For instance, we can imagine an author used to typewrite his/her
texts and then to correct them by hand: in this case the
assumption of an absolute time is possible. But as different
layers of hand corrections can also occur, there will be, in this
case also, the necessity of considering relative-timed
interventions. This situation can be represented in encoding
distinguishing absolute and relative timing for instance by the
application of two different attributes, i.e. @timeRel and
@timeAbs In a digital framework an incorrect consideration
of time as absolute or relative can bring to display texts that
never existed. Let us imagine for a moment that we build an
XSLT based tool able to display at a time either Time 1, Time
2 or Time 3 etc. variants: the results would be the display of
variants that might have lived in different moments.
I think that, in absence of explicit authorial declarations or of
palaeographic evidence, the only possible display would be to
show timing of variants segment by segment, i.e. to give
evidence just of relative timed corrections.
In the presentation I will present some examples from different
authors (Giacomo Leopardi, Jane Austen, Virginia Woolf and
some others) to examine the complexity of representation of
temporal factors in digital genetic editions.
1. The question is too complex just to try to give some basic
references; anyway Gresillon 1994 will offer a good starting point.
Bibliography
Grésillon, A. Eléments de critique génétique. Lire les manuscrits
modernes. Paris: Presses Universitaires de France, 1994.
Sperberg-McQueen, C. M., and Lou Burnard, eds. TEI P4:
Guidelines for Electronic Text Encoding and Interchange, XML
Compatible Edition. Oxford: TEI Consortium, 2002. <http:
//www.tei-c.org/P4X/index.html>
TEI Manuscripts SIG. Meeting Report 01. Nancy: ATILF,
November 8th 2003 . <at http://www.tei-c.org.uk
/Activities/SIG/Manuscript/mssigr01.xml.I
D=body.1_div.3>
Vanhoutte, Edward. "Putting Time Back in Manuscripts. Text
Ontology, Critique Génétique and Modern Manuscript." Paper
presented at ALLC/ACH 2002 Tübingen: University of
Tübingen, 25 July 2002. 2002. <http://www.uni-tueb
ingen.de/cgi-bin/abs/abs?propid=93>