[tei-council] Report on Vilnius meeting

[Matthew James Driscoll]

TEI Place-name meeting, Vilnius, 23-24 February

In attendance were: Lou Burnard, Matthew Driscoll, Øyvind Eide, Richard
Light, Tadeusz Piotrowski, Tatjana Timchenko and Sebastian Rahtz.

The first day we were at the Institute of Lithuanian language
(http://www.lki.lt/indexeng.php), part of the Lithuanian Academy of
Sciences, and on the second day the Faculty of Philology
(http://www.flf.vu.lt/) at Vilnius University; we are grateful to both
institutions for acting as hosts.

1. Names and nyms

We began by tackling one bit of business outstanding from the Personography
meeting last year in Oxford, viz. the development of a mechanism for
pointing from actual name instances to the canonical form of that name, thus
addressing the needs of onomasticians (who are interested in names per se),
in addition to those of prosopographers (who are interested in people). If,
for example, the names Tony Blair and Tony Benn occur in a text it should be
possible to tag them in such a way that they both point to information about
these respective gentlemen and are flagged as instances of the name "Tony".
It should further be possible to indicate that "Tony" is a (pet-)form of
"Anthony", which is itself a member of a family of names containing forms
such as "Antonio", "Anton", "Antoine" and so on. We propose doing this by
means of a new element, called <nym>, which contains the definition for a
canonical name or name-part of any kind. In addition to global attributes
and those inherited from att.typed it can take a @parts attribute, which
points to constituent nyms. The attribute @nymKey is available on any
element which is a member of the att.naming class in order to point to the
nym with which it corresponds. Thus, to take our example, the name "Tony
Blair" in running prose could be tagged as follows:

<persName key="#ACLB">
    <forename nymKey="#ANT1.1.1">Tony</forename>
    <surname>Blair</surname>
</persName>

The @key attribute on <persName> would point to a <person> element giving
information about Tony Blair, while @nymKey on <forename> points to the
relevant <nym>. A <nym> element may also combine a number of other <nym>
elements together, where it is intended to show that one is a pet form or
diminutive of another, or that different nyms are to be regarded as variants
of the same base nym, using the <form> element (from the model.entryParts
class); orthographic variants are dealt with using <orth>, while <etym> can
be used for information on the origin of the name:

<listNym>
    <nym type="base" xml:id="ANT1">
        <form xml:lang="la">Antonius</form>
        <etym>From the Roman family name <mentioned>Antonius</mentioned>,
which is of unknown, presumably <lang>Etruscan</lang>, origin. It has been
commonly, but incorrectly, associated with <lang>Greek</lang> <mentioned
xml:lang="gr">&#x03B1;&#x03BD;&#x03B8;&#x03BF;&#x03C2;</mentioned>
<gloss>flower</gloss>, which resulted in spellings with th in some
languages.</etym>
        <nym type="lingVar" xml:id="ANT1.1">
            <form xml:lang="en">
                <orth>Anthony</orth>
                <orth>Antony</orth>
            </form>
            <nym type="dim" xml:id="ANT1.1.1">
                <form xml:lang="en">Tony</form>
            </nym>
        </nym>
        <nym type="lingVar" xml:id="ANT1.2">
            <form xml:lang="it">Antonio</form>
            <nym type="dim" xml:id="ANT1.2.1">
                <form xml:lang="en">Tonio</form>
            </nym>
        </nym>
        <!-- other variants of Antonius -->
    </nym>
    <!-- other nyms -->
</listNym>


This mechanism could also be used for place-names and place-name elements
(e.g. thorp, caster).

2. Place-names

Our principal task at this meeting was to develop mechanisms for encoding
place-names, analogous to those which were developed for personal names at
the meeting in Oxford last year, which would allow for the recording of
abstracted information about a place, such as map coordinate, GIS
information etc., as well as variant forms of the name, in different
languages (e.g. Praha, Prague, Praga) and/or different forms over time (e.g.
Lundunum, London). On the analogy with <person>, we propose a <place>
element, which will usually contain at least one, and possibly several,
<placeName> elements, followed by one or more <location> elements to provide
geographical and/or geo-political information about the location of the
place. The existing <locale> element is available to provide a brief
informal description of the nature of a place. In addition, three new
elements have been proposed: <placeTrait>, <placeState> and <placeEvent>,
which all have similar content models to their counterparts within <person>.

To take a fairly simple example:

<place xml:id="IS">
    <placeName xml:lang="en">Iceland</placeName>
    <placeName xml:lang="is">Ísland</placeName>
    <location type="lat-long">65 00 N, 18 00 W</location>
    <placeTrait type="area">103,000 sq km</placeTrait>
    <placeState type="gov" notBefore="1944">Constitutional
republic</placeState>
    <placeEvent type="political">Previously part of the kingdom of
<placeName key="#DK">Denmark</placeName>, Iceland became independent
on <date value="1944-06-17">17 June 1944</date>.</placeEvent>
</place>


3. Events

There was also some discussion on the feasibility (and desirability) of
developing a generic tagset for encoding assertions about events, although
no real conclusion was reached.

M. J. Driscoll