Tools of Dual Utility: Multimedia Applications for Native American Language Preservation and Teaching

1. Introduction
1.1 Text, Multimedia, or Both?
As hyperliterate academics, it is easy for us to
forget that the basis of language is speech and
gesture. We are accustomed to viewing language
as written text, yet textual conventions only approximate language; they are not language itself.
A more accurate representation of a language
would include audio (speech) and video (gesture).
The analysis and manipulation of language data,
however, require it to be in an abstract form, and
rightly so. This requirement has lead to a bias in
linguistic computing towards texts and against
multimedia. We argue here for a multimedia application which includes speech and gesture and
which is fundamentally based in a fully abstractable text.
This has been a project to create a multimedia
CD-ROM of a Native American language, Rio
Grande Téwa. It is fully cross-platform, with a
user interface and an abstracted expandable
language database. It contains videos of Tribal
storytellers with accompanying text (transcribed
and translated), a pronouncing dictionary, language-learning activities, and many images. The aim
has been to start with speech and gesture (video,
audio), but to still maintain primary focus on the
texts (story transcription, dictionary, language games). These texts, representing organic language,
need to be in a format flexible enough to allow
manipulation of the data by future users. Yet multimedia applications available to date treat text as
an irritating vestigial appendage, usually as graphical images. The problem that we have addressed, then, is how to use the advantages of multimedia yet still have the text in an abstract form.
1.2 The Social Conscience of Computing
Technology
Those of us working with this expensive and (to
most) inaccessible technology have at times feared it elitist, reinforcing inequities in society. Yet
many of us have justified the use of such technology by citing its potential to resolve societal ills:
it is supposed to enhance the collecting, archiving,
processing and teaching of language data.
Recently there has been a wave of interest in
multimedia applications among native peoples of
the Americas as a means of archiving and teaching
indigenous languages and traditions. This is part
of a larger movement that of cultural and linguistic
renewal worldwide.
The goal which drives this project is local empowerment. In the 1990s and beyond, notes linguist
Kenneth Hale, all work done on Native American
languages will be under the auspices, direction,
and control of Native Americans themselves. Native American language rights in schools and government are again being challenged in the United
States today by a small but vocal group of Englishonly advocates. Those Native American juveniles
and adults who were forbidden to speak their
language when growing up now have a keen desire
to learn their own language with the computing
tools they have become accustomed to in the
workplace. With these tools, the speech community has not only a stand-alone CD for languagelearning, but it also has the tools to append information to the language database. Other
language-teaching modules can easily be created.
2. Project Background
Téwa is a Kiowa-Tanoan language spoken in six
Native American communities in New Mexico
and one in Arizona. Through the years these communities, known as Pueblos, have maintained and
transmitted their language orally, often in the face
of active language suppression. It is only since the
1960’s that orthographies were devised for these
Pueblo languages. Only with an orthography is
wide-scale language renewal possible.
This project provides language renewal tools and
curriculum materials for one of these communities, San Juan Pueblo. Once a leader in Native
American bilingual education, the San Juan Pueblo Day School was forced to give up its language
training staff and curriculum during the 1980’s
due to pressure from the Bureau of Indian Affairs.
In 1995, Jacobs, who has worked with the community for 25 years, obtained permission from the
San Juan Pueblo Tribal Council to work with
community members and school staff to develop
computer-based tools for teaching Téwa as a second language. Within months, the Tribe was
awarded funds from the Chamiza Foundation to
develop these pedagogical tools for both children
and adults at three key sites: the Pueblo school, the
library, and the Ohkay Ówîngeh Cooperative.
This language renewal program is The Téwa
Language Project.
At the University of Washington’s Center for Advanced Research Technology in the Arts and Humanities (CARTAH), we have been working on
two aspects of this program. First, we created PC
and Macintosh fonts in the Téwa practical orthography. These are already being used both by
children at the Pueblo school and by adults for
Tribal business. Having Téwa fonts, said one Pueblo member, is “a liberating experience.... At last
I can write in my own language!”
Secondly, we created an interactive CD-ROM for
pedagogical and archival purposes. We are using
multimedia authoring software (ToolBook and Director) to display video, concurrent text, and play
sound files; the CD also includes a queryable
dictionary and numerous language-learning activities.
3. Practical Considerations
3.1 Fonts
The dialect differences between the Rio Grande
Pueblos are reflected in their orthographies. Téwa
uses a practical (as opposed to phonetic) orthographic system introduced in the 1960s by Summer
Institute of Linguistics researchers Randell and
Anna Speirs. Since Téwa distinguishes both tone
and nasalization phonemically, a Latin-script based orthography necessarily includes a large character set. Using Fontographer, we designed and
created Téwa practical orthography fonts for both
Macintosh and PC-Windows platforms. A simple
keyboard-reassigning program allows Pueblo
users mnemonic access to characters not found on
the keyboard.
The text material can be represented in the database as a font specification plus an ASCII code.
While not as elegant as a 16-bit (Unicode) representation, the installation of a font in Windows or
on the Macintosh is simple, and ASCII text works
with any application, while Unicode is still not
available in most applications. As this is a practical orthography to be used by novice and experienced computer users alike, we were compelled to
use schemes that could easily be applied to a wide
variety of applications.
3.2 Media Objects in a Database
While storing ASCII text in any database is easy,
the storage of audio, video, and pictorial material
presents some complex problems. First, these objects tend to be very large and therefore cannot be
stored in normal fields. As a result, there is no
indexing scheme available for the data; we cannot,
for example, mark video frames in any standard
database. Also, formats for different media types
tend to vary on different platforms. It is fairly
simple to convert between formats. But if the
objects are stored in the database, we must decide
whether (1) to store all possible formats, or (2) to
devise a scheme that converts binary objects on
the fly. Another problem is the unavailability of
cross-platform database file formats that support
multimedia objects. This problem alone is enough
to make both of the above approaches unfeasible.
One solution has been to store only references to
multimedia files in the database. This presents
some new problems and doesn’t solve many of the
problems of an embedded system, but it is certainly feasible. Neither does this approach address any
scheme of indexing in the files themselves, nor
does it solve problems of differing media formats
across platforms. Referencing files presents new
problems, including the management of the data
files, the verification of the data files, and the
construction of naming schemes the are independent of any operating system. Because it is
only ASCII text that is being stored, from the
perspective of the database itself implementation
is not problematic. Its advantages include: the
database stays a small size; cross-platform compatibility of the database is proven; it is somewhat
easier to manage files as opposed to managing
large binary objects in a database; and any large
or universal formatting changes can be done independent of the database (e.g. converting stereo to
mono audio files).
We’ve chosen to use an XBASE file format, dBase, because we’ve found suitable readers for the
XBASE format for both Mac and PC. Since the
file format is very common, it is a simple matter
to build reader applications on any platform. The
files originate as data entered into a specially-designed ToolBook application, which are then exported to a dBase file format. Once we have the
database well defined, we will move to a ToolBook
front end to the dBase file and enter data directly
to the database. ToolBook’s facility for this is
provided in a DLL that is part of the ToolBook
product. On the Macintosh platform, we are using
a product called FileFlex that is made available as
an XLIB that, when incorporated into Director,
allows Director to act as a front end to the dBase
file. We chose ToolBook as the primary editing
tool because its relationship to dBase is well
known by members of the team. We have been
able to edit the database from ToolBook on a PC
and immediately view the results in Director on a
Macintosh.
72
3.3 The CD-ROM
We will demonstrate a CD-ROM with the database information organized and optimized to be
cross-platform. Along with the database, we will
present supporting applications and examples of
language study curricula. The dictionary is independent of any application or platform, being abstracted into a dBase file with suitable references
to audio and visual media. The organization of the
media files is contingent on compatibility within
any one medium. Some file formats will play on
both platforms and so will be held in a common
directory. Others may need different formats; the
original format will be converted to one suitable
for the remaining platform and both files will be
placed on the CD. The reader applications on the
CD will need to be aware of all these quirks and
discrepancies so that they can interpret the media
references in the database in a meaningful manner.
These considerations limit the amount of information we can place on the CD. Video consumes
large amounts of space. It becomes imperative that
any video we use not be duplicated, so we must
convert our videos to QuickTime files, as this is
the most stable cross-platform standard for video.
Still, the size of the videos makes it impossible to
include more than five or six extended (six minute
or more) videos. This is unfortunate, as this type
of material, typically story-telling, conveys best of
all the gesture aspects of the language. Obviously,
much of the material that is currently on video
cassettes held in private collections will only be
available for general distribution once higher-density media becomes available.
Audio files must be dealt with in a similar manner,
but it is easier to control the size of the files.
Mostly, audio files will be used for pronunciation
of the words in the dictionary, and while these will
be short, there will be a large number of them. In
order to avoid duplication, we will need to use a
single format that both platforms can read. Illustrations and texts can be treated in a similar
manner.
The main database itself will not be particularly
large. The final version of this database will probably not be larger then two megabytes. The fields
for each record in the database are: a Téwa headword: a unique identifier; English gloss; example
sentence; English gloss of the examples sentence;
reference to the audio file; semantic category; and
variant lexemes.
Certain possible fields we have deemed unessential to the purpose of the primary database, but as
having potential interest to some users. We will
build secondary, related databases as time allows
and include them on the CD-ROM. Examples of
these databases include an index of video frames
that illustrate words and an index of where words
are used in the included texts.
References
Hale, Kenneth. 1972. A New Perspective on American Indian Linguistics. In Alfonso Ortiz, ed.,
New Perspectives on the Pueblos, Albuquerque N.M.: University of New Mexico Press:
87-133.
Leap, William L. 1988. Indian Language Renewal. Human Organization, Vol. 47, No. 4:
283-291.
Martinez, Esther. 1982. San Juan Pueblo Tewa
Dictionary. San Juan Pueblo, New Mexico:
SJP Bilingual Program.
Speirs, Randall. 1968. Linguistic Aspects of Rio
Grande Tewa (Ph.D. dissertation). Ann Arbor,
MI: University Microfilms