The following inquiry, which came to me privately a few weeks ago,
may be of general interest, so with the permission of the inquirer, I
am posting it to TEI-L, together with my reply.
First, the inquiry, from Rafal T. Prinke of the University of Poznan:
Date: Wed, 04 Jan 95 17:19:29 CET
From: Rafal Prinke <RAFALP@PLPUAM11>
Subject: Conference report & a question
To: "C.M. Sperberg-McQueen" <[log in to unmask]>
While I think I understand how to mark up names (with roles, id's,
pointers, etc.) and dates (with various time frame references), I
have a problem finding out how to deal with events. The <event> tag
seems to be reserved for annotating records of spoken language. Or
perhaps it is legal TEI to use the same in such contexts as:
king Alfred's ascension to the throne
If so, then I suppose all other TEI mechanisms can be applied to
it, using pointers, roles played by participants of the event, etc.
And my reply:
Thanks for your note and your inquiry about tagging events. If you
don't object, I would like to forward your inquiry and portions of this
reply to the list TEI-L, as I think the question is of general interest.
Your first instinct is correct, and the existing EVENT element is
only legal if the base tag set for spoken texts is selected, and it is
designed to be used in the restricted sense described in that chapter.
Tagging events of the type you mention (e.g. the coronation of King
Alfred), or events of a typical type (the Buddha goes into a city to
beg for food) poses an interesting problem in the application of the
Guidelines, for which so far several possible solutions have been
found; I'm not sure any is wholly satisfactory (but then I'm not sure
any one solution can be wholly satisfactory for all users).
1 use the RS element, on the grounds that what is being tagged is a
reference (or 'referring string') to an event of a particular type.
This works reasonably well for many purposes, and may be the best
approach for most cases. Some people are unsettled (as I am, a bit)
by using RS for events, clauses, or passages of narrative, rather
than for noun phrases or pronouns: the tag was originally proposed
as a generalization from specialized tags for PERSON, PLACE, and ORG
which were not limited to proper nouns but could encompass any
reference to objects of those types. The definition of the element,
however, does not restrict RS to noun phrases, and so this may be
a misplaced and erroneous objection.
2 use FS elements (feature structures) to describe an event or a type
of event in any degree of detail desired, and link the feature
structures to the actual passages of text using the normal mechanisms
(e.g. the ANA attribute on the passage of text, tagged as P or as
SEG or as whatever, or a LINK element).
This has the advantage that you can define as elaborate a structure
of information as you like, and are not limited to the information
structure provided by the pre-defined attributes of RS. The structure
you define can be enforced, in theory, or at least documented, by a
feature system declaration. The disadvantage is that at present these
are among the least widely used, and thus least widely understood,
portions of the Guidelines, and for some purposes they seem rather
cumbersome, with more layers of indirection than one could wish.
3 Extend the Guidelines with a new element of your own (preferably
with a name other than EVENT, to avoid confusion with the standard TEI
element of that name), adding attributes to capture the characteristics
of the event which you want to record (type, date, duration, agents,
patients, instruments, causes, effects, ... ?)
Without better knowledge of your texts and goals, it's hard to give
firm advice, but the simplest way to begin is probably to use the RS
element at least to begin with; after encoding a number of texts, the
precise shortcomings of RS for your particular application will be
clear, and you will be in a better position to design a good feature
system or new SGML element.
And now, the question for TEI-L: is anyone using any other TEI
mechanisms to tag events?
-C. M. Sperberg-McQueen
ACH / ACL / ALLC Text Encoding Initiative
University of Illinois at Chicago