LINGUIST List 7.465

Tue Mar 26 1996

FYI: Endangered languages documentation mailbox

Editor for this issue: Ljuba Veselinova <lveselinemunix.emich.edu>


Directory

  1. Dietmar Zaefferer, FYI: Endangered languages documentation mailbot

Message 1: FYI: Endangered languages documentation mailbot

Date: Tue, 26 Mar 1996 17:59:39 +0100
From: Dietmar Zaefferer <ue303bhsunmail.lrz-muenchen.de>
Subject: FYI: Endangered languages documentation mailbot


 Announcing LDUL: The Language Documentation Urgency List
 ========================================================

Abstract
========

LDUL is an automatic mailbox and database for the collection and
retrieval of information on how urgently the individual languages of
this world are in need of documentation. The aim is to help in the
decision of where to focus fieldwork and in the writing of proposals
for fund raising purposes.

Overall documentation urgency (DU) is measured as the average of six
special documentation urgencies in phonology, morphology, lexicon,
text corpus, syntax, and semantics/pragmatics. These in turm are
measured as overall degree of endangerment times special degree of
documentation need, where endangerment is the inverse of estimated
language vitality and special documentation need is the inverse of the
estimated sufficiency of existing special documentation.

The language vitality score is calculated from eight different factors
such as age of youngest speaker, number of speakers, percentage of
monolingual speakers etc. (for details cf. the comments to the
demoquestionnaire.)


Background
==========

Most linguists know that the number of human languages in use is rapidly
decreasing. Estimates of the rate of disappearance vary between 12 and 50 a
year. This means that up to 10% of the linguistic heritage of humankind
will be irretrievably lost by ten years from today and that by the end of
the coming century, less than 1000 of the current over 5000 languages will
still be alive. It seems obvious that where it is not possible to save
these languages, they should at least not die without leaving a trace, that
is without being documented in a satisfactory way.

This is not to deny that the task of saving the peoples who are
sometimes endangered together with their language is much more
important; it's just that this task is not specific to linguists!
Those who wish to participate in helping endangered peoples may
contact organizations like - Survival International, 310 Edgware Road,
London W2 1DY, UK;
 phone ++44-71-2421441, fax ++44-71-2421771, or - Gesellschaft fuer
bedrohte Voelker, Postfach 2024, D-37010 Goettingen;
 phone ++49-551-499060, fax ++49-551-58028,
 e-mail: gfbv-germanyoln.comlink.apc.org


Current activities
===================

The international community of linguists is not unaware of this situation.
To mention just a few activities:

- in January 1991 a special symposium entitled "Endangered Languages
and their Preservation" was held at the Annual Meeting of the LSA,
- in 1991 a volume on "Endangered Languages", edited by R.H. Robins and
 E.M. Uhlenbeck, appeared (Berg, Oxford/New York), 
- in March 1992 "Language" published a collection of essays in its
vol. 68, and
- in August 1992 the XVth International Congress of Linguists devoted a
 plenary session to this topic, both with the same title as the
 Robins/Uhlenbeck volume,
- in October 1992 the working group 'endangered languages' of the DGfS
 (German Linguistic Society) published an "Informationsbroschuere zur
 Dokumentation von 'Bedrohten Sprachen'",
- in August/September 1993 the University of Cologne hosted a summer
school on language description and fieldwork,
- at its September 1993 meeting, the Linguistic Association of Great
 Britain (LAGB) had a special session on Endangered Languages,
- in July 1994, a workshop on Language shift and maintenance in the
 Asia-Pacific region was held at La Trobe University, Melbourne,
- on the 7 Sep 1994 Endangered-Languages-L electronic forum at ANU was
 established (http://coombs.anu.edu.au/CoombsHome.html),
- at the January, 1995, LSA Meeting there was an organized session called
 'Field reports/Endangered Languages',
- in February 1995, there was a conference on Endangered Languages at
 Dartmouth college, and for
- on April 21, 1995, the University of Bristol held a seminar
 on the conservation of endangered languages.
- November 18-20, 1995, there was an International Symposion on Endangered
 Languages at the newly founded ICHEL (International Clearing House for
 Endangered Languages, U of Tokyo, http://www.tooyoo.L.u-tokyo.ac.jp/).


What else can we do?
====================

Still, there remains a lot to be done. One central thing to do is FUND
RAISING: We have to talk politicians, institutions, responsible into
giving money not only for the preservation and documentation of
species of birds and insects, but also of cultures and languages. Note
that biologists (three joint societies) are demanding $3 billion a
year (sic!) for a documentation of biodiversity called Systematics
Agenda 2000 (Nature, vol. 368, 3 March 1994, p.3). How much do we need
for a documentation of glotto- and ethnodiversity?

Another thing is to MOTIVATE linguists to do the necessary field work
once the money is available. This has to include a change in hiring
politics and CV-evaluation: Time spent for field work should count as
favorable for a candidate, and not the opposite.

And a third thing is to provide the motivated and funded linguist with
the necessary INFORMATION on where to go first, since he is in the
situation of a fire fighter when fire is all over the place. And this
is where the LINGUIST discussion list (with over 7200 subscribers,
more than 1.4 per existing language) and similar ones come into play.


The contribution of the Language Documentation Urgency List (LDUL)
==================================================================

The world-wide computer networks and especially the LINGUIST LIST have
turned the world's community of linguists (at least its electronically
accessible part, but via them a lot more) into a global village. And
if the inhabitants of this village join forces, it should be easy to
solve the third problem mentioned above using the pot luck party
method: Everybody who knows about a language in need of proper
documentation or in the process of disappearing throws his knowledge
into a pool called LDUL. This is an automatic electronic mailbox and
database with the following address:

 ldulcis.uni-muenchen.de

If you want to know more about LDUL, simply a message to this address
with the following entry under "Subject" (the message body may be
empty, or, if your mail system doesn't tolerate this, contain
anything, it will be ignored):

 about LDUL

and you will receive the information you are reading right now. If you
just want a short overview of supported mail commands, send a message
(again with empty or trash body) with the following entry under
"Subject":

 mail commands

If you think you have discovered a bug in the program, send a
corresponding message (no empty body!) with the following entry under
"Subject":

 bug


Contributing to LDUL
====================

If you want to contribute, send a message to the LDUL address with the
following entry under "Subject":

 send demoquestionnaire

The system will mail you a copy of a completed questionnaire with
annotations that specify the different questions.

When you've done this you will hopefully want to complete an empty
questionnaire, and you can get one by sending another message under
"Subject":

 send questionnaire

The system will send you by return mail a copy of a blank
questionnaire.

Once you have completed a questionnaire (please read the annotations
to the different questions that come with the demoquestionnaire
carefully!), write

 deposit questionnaire

into the subject field and mail the completed questionnaire to the
same address.

The way it is treated there is the following:

If the language code on the questionnaire you have completed is
identical with the language code in a questionnaire already on file,
your contribution is added to that file, else a new file is opened. So
it's the language code that counts for the identification of a
language and not its name(s), since there are too many ambiguous
language names! If you deposit a questionnaire without the language
code, LDUL will add it for you, if your language name is unambiguous,
else it will complain.

For each vitality factor you have either checked one of the five
values (there is a minimum, a maximum, and three intermediate
degrees), or the option 'unknown'.

The weights of the first vitality factor (age of youngest speaker) are
0, 4, 8, 12, and 16 points, the weights of the other seven factors are
0, 3, 6, 9, and 12 points each. The total vitality score therefore
ranges from 0 (worst case, lowest vitality) to 100 (best case, highest
vitality).

If you are not in a position to give information on all the factors,
the vitality score will be an interval rather than a point value.
Suppose all other factors add up to 20, but the language attitude is
unknown. Then the vitality score will be the interval between 20
(worst case, very negative attitude) and 32 (best case, very positive
attitude). If another questionnaire on the same language contains
information about language attitude, your 'unknown' contribution is
simply ignored.

The sufficiency (quantity and quality) of existing documentation in a
given domain such as phonology is given a score of 1, if it is very
high (completely sufficient), and 0, if it is completely unsufficient,
with the three obvious intermediate values.

The special Documentation Urgency or DU scores are calculated as (100
- vitality score) * (1 - special documentation sufficiency score).
For instance the text corpus DU score will be 100 (the maximum) just
in case vitality is 0 and the text corpus documentation sufficiency
score is 0 as well. The overall DU score as the average of the special
DU scores will therefore be 100 (the maximum) just in case vitality is
0 and all special documentation sufficiency scores are 0 as well.

If you check only 'unknown' in either the vitality or the
documentation section, you will receive a corresponding error message:

 The error occurred while parsing :
 No vitality information specified in Section 2 !

or

 The error occurred while parsing :
 No documentation information specified in Section 3 !


Consulting LDUL
===============
If you want to consult the list, you have several choices.

1. If you are interested in a specific language, say Lisu, send a
message to the same address as above with the following entry under
"Subject":

 info on Lisu

Then LDUL will mail you the set of questionnaires that have been
completed with Lisu in the list of names and aliases. The subject line
of the message will look like follows:

Subject: Re:info on lisu ( TLC=lis )

The (TLC=lis) information is important here, since there is another
language named Lisu with a different three letter code (TLC=tkl), and
if information is on file on that language as well, it will also be in
your mail. So if you are in doubt about the identity of the language
you are inquiring about, please consult the "Ethnologue Database" to
find out among other things the three letter code of your language,
e.g. via the World Wide Web:

http://www.sil.org/ethnologue/ethnologue.html
 or
http://www-ala.doc.ic.ac.uk/~rap/Ethnologue/
 or via Gopher:
gopher://sil.org/11/gopher_root/ethnologue/

2. If you want to know which languages have been treated so far, send
the subject entry

 get languages

to LDUL, and it will mail you the current alphabetical list of names
of languages about which information is on file.

3. If you want to know the current overall documentation urgency (DU)
ranking of the languages that have been treated so far, send the
subject entry

 get overall DU ranking

and LDUL will send you a list of the languages on file, ranked
according to their overall DU scores.

4. If you want to know one of the current special documentation
urgency (DU) rankings of the languages that have been treated so far,
send one of the subject entries

 get phonology DU ranking
 get morphology DU ranking
 get lexicon DU ranking
 get text corpus DU ranking
 get syntax DU ranking
 get semantics/pragmatics DU ranking

to LDUL, and it will send you a list of the languages on file with
their appropriate special scores, ranked according to these scores.

5. If you want to know the complete current statistics, send the
message

 get statistics

to LDUL, and it will mail you its complete statistics in its current state:
- the number of completed questionnaires,
- the number of languages treated,
- the average number of completed questionnaires per language,
- the list of languages ranked according to their overall DU score,
 giving for each language
 - the language code,
 - all language names collected so far,
 - the number of completed questionnaires for this language,
 - the overall documentation urgency score for this language together with
 - the special urgency rankings of its aspects (phonology etc.),
 with the average value first followed a maximum and a minimum value,
 (the latter two matter only for interval values and can be neglected
 if they coincide)
 - the language's vitality score,
 - its endangerment score,
 - its overall documentation score, and
 - its overall documentation need.

Since this file will soon be rather bulky, be careful with this command.


Spreading the word
==================

Those who don't have access to email but might have relevant knowledge
should be addressed by snail mail, fax, phone or face-to-face and
should be encouraged to pass on the news in order to achieve some kind
of snowball effect. Information on paper questionnaires will have to
be transferred to the computer by somebody who does have access to
email.


!!!!!!!!!!!!!!!!!!!!!! WARNING !!!!!!!!!!!!!!!!!!
=================================================

There is no such thing as a perfect questionnaire and it should go
without saying that no linguist wishing to do fieldwork should base
the choice of his language exclusively on the LDUL data.

One factor that had to be neglected in the design of the questionnaire
is the degree of relatedness of the language in question to the 'next'
well- documented language: the lower this degree, the higher the
documentation urgency.

Another thing is that the program cannot resolve contradicting
information on the same language. It will rather compute the average
scores, e.g. if one contributor thinks the quantity and quality of the
documentation on Tsachila phonology is medium (.5) and another one
thinks it is low (.25), it will come up with a score of .375.

The whole enterprise is a first and as such an experiment, but I think
it is worth trying. Let me conclude with a somewhat pathetical appeal:

 Linguists of all countries, unite
 and try to save what can be saved
 of what you love enough to devote your professional life to:
 contribute to LDUL!


Dietmar Zaefferer
Institut fuer Deutsche Philologie Phone: +49 89 2180 2060 (office) or
Universitaet Muenchen +49 89 2180 3819 (office)
Schellingstr. 3 +49 89 36 66 75 (home)
D-80799 Muenchen Fax: +49 89 2180 3871 (office)
Germany Email: ue303bhsunmail.lrz-muenchen.de


Acknowledgements
================

I am indebted first of all to "jollycis.uni-muenchen.de" alias
Patrick Stein, who did the programming, then to Franz Guenthner for
organizational support, and last but not least all those who commented
on earlier, more naive versions of the questionnaire (none of whom may
be held responsible for remaining shortcomings of it):

- Leila Behrens
- Bernard Comrie
- Bill Croft
- Christian Lehmann
- Hans-Juergen Sasse
- Christel Goldap
- Gunter Senft
- Vladimir Tourovski
- Nancy Dorian
- Martin Haspelmath
- Barbara F. Grimes

PD Dr. Dietmar Zaefferer
Institut fuer Deutsche Philologie phone: +49 89 2180 2060 (office) or
Universitaet Muenchen +49 89 2180 3819 (office)
Schellingstr. 3 +49 89 36 66 75 (home)
D-80799 Muenchen fax: +49 89 2180 3871 (office)
Germany email: ue303bhsunmail.lrz-muenchen.de
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue