Speaking in support of clear use of language, and as someone who has spent 
the past thirty years learning (or trying to learn, to the best of my 
ability) how to date medieval manuscripts, I add to the previous comments:

The values of dated, datable and dated by the cataloguer are crucial to 
the field of paleography; hiding them behind words that require definition 
to be understood by people in the field is pointless.  As an indicator of 
the centrality of the issue and of the standard use of the terms "dated" 
and "datable," it will be sufficient to go to the website of the Comite' 
International de Pale'ographie Latine at:
 	http://www.irht.cnrs.fr/cipl/Cipl.htm
and click on the link "Catalogues de manuscrits date's."  Since 1953 when 
the Comite' first proposed this goal, eleven countries have to date 
produced over fifty volumes of plates of dated and datable manuscripts to 
serve as touchstones for the huge majority of manuscripts that are, 
instead, dated by the cataloguer.  "Dated" and "Datable" (in whatever 
language) is always part of the title of these volumes.

Because of the centrality of the issue, the TEI Working Group on med/ren 
manuscript description proposed an attribute, dateAttrib, on 
<msDescription> and on <msPart> with stated values "dated," "datable," and 
"unknown" in order to unequivocally recognize these categories.  We 
retained the generic TEI attribute, evidence, with values "internal," 
"external," and "attributed" on <origDate> where the discussion about the 
datability (and various levels thereof) is presented, potentially at 
least, in some length and with as much nuance as the situation might 
require.

Which brings me to Issue # 2:  "conjecture" vs. "attributed."  Whether the 
cataloguer of medieval manuscripts works from paleographic knowledge, 
art-historical knowledge, textual knowledge, knowledge based on the origin 
of certain medieval collections, or, more likely, from a combination of 
all of the above, the manuscripts acquire a date based on expertise, not 
on guesswork.

One would not say that an art historian "conjectures" that a certain 
painting is by Degas; one would say that the attribution to Degas of a 
certain painting has been made by so-and-so.  Equally, I and those who 
catalogue medieval manuscripts do not "conjecture" or guess at a date, we 
assign or attribute it.  The choice of words speaks to the level of 
professional knowledge and experience involved.

For the DTD and the documentation proposed by the TEI working group on 
manuscript description, and a discussion of the terminology "dated," and 
"datable," and the term "attributed," please see:
 	http://www.scriptorium.columbia.edu/
and click on "Technical Information"; the left navigation bar offers 
"Description DTD" with the relevant pages.

\Consuelo Dutschke

Consuelo W. Dutschke, Ph.D.
Curator, Medieval and Renaissance Manuscripts
Rare Book and Manuscript Library
Columbia University
535 W. 114th Street
New York NY 10027

phone: 212-854-4139     fax: 212-854-1365     e-mail: [log in to unmask]
---------------------------------
On Mon, 14 Mar 2005, M. J. Driscoll wrote:

> There was quite a lot of discussion about dating in the MS-
> task-force, but in the end it was decided, by default
> admittedly, to leave things as they were in MASTER, i.e.
> with two attributes, certainty (high | medium | low) and
> evidence (internal | external | conjecture). One suggestion
> was to replace "conjecture" with "attributed", which for
> some reason sounds less subjective, another was that
> evidence should be dropped and dating (or dateAttrib)
> added, with possible values (dated | datable | undatable |
> unknown). The argument for not doing this is that
> evidence="internal" means dated, evidence="external" means
> datable, evidence="conjecture/attributed" means undatable:
> there are, after all, pretty much only these three
> alternatives and what you call them doesn't really matter.
> Discussion concerning the nature of the evidence should
> come in the body of the <origin> element, it seems to me,
> rather than in attribute values. But I'm (almost) always
> willing to listen to counter-arguments.
>
> Matthew
>
>
> M. J. Driscoll
> Den Arnamagnæanske Samling
> Nordisk Forskningsinstitut
> Københavns Universitet
> Njalsgade 136
> DK-2300 København S
---------------------------------------------------
>
>> When it comes to assign dates to something (date, origDate,
>> @notBefore, @notAfter), so far P5 allows only for the
>> ambigious attribute evidence for explanation. The
>> documentation offers internal, external and conjecture as
>> possible values. What about the differentiation  between the
>> "why" and the "how"? How can I make explicit both the source
>> and the "kind-of" of my dating? Something like "script",
>> "explicitly dated", "finding from watermarks" additionally
>> to "internal" and "external"? Did someone ever feel the need
>> for another attribute like source or reason or the like? If
>> so, what did you do? Did you use mainly other values within
>> evidence? Should the documentation be "updated" to give
>> hints to use the attribute this way as well? Should be
>> introduced another attribute?
>>
>> Torsten
>>
>> -----
>> Torsten Schassan (Projekt Handschriftendatenbank/MASTER)
>> Herzog August Bibliothek, Postfach 1364, D-38299
>> Wolfenbuettel Tel.: +49-5331-808-117, [log in to unmask]