Slide 1
Slide 1 text
Language Documentation and
Description: And What Comes After
Claire Bowern,
Yale University:
[email protected]
Cundeelee Wangka Mission
Slide 2
Slide 2 text
Chirila community/outreach projects
o “Unofficial” partnerships with language centres to make primary
materials more available
o Public media work (e.g. articles for Conversation)
o Grammar bootcamps for language
documentation
Slide 3
Slide 3 text
What is a “Grammar Boot Camp”?
o Several undergraduate students work together (and with
professor)
o Work on someone else’s fieldnotes (+ texts) intensively
o Regular consultation with communities and linguist
o Write a sketch grammar and make other materials
o in a month!
Slide 4
Slide 4 text
Yale Bootcamps:
o 2014: Tjupan
o 2015: Ngalia
o 2016: Cundeelee Wangka
and Kuwarra
o 2017: Noongar
Kado Muir, Sue Hanson, and Andy Zhang (2014)
Slide 5
Slide 5 text
Bootcamp language locations
Slide 6
Slide 6 text
Boot camp origins
o Collaboration between Sue Hanson (then Wangka Maya, now
Goldfields Language Centre, Western Australia).
o Sue working with communities in WA
o Claire looking for documentation experiences for students in the
US
Slide 7
Slide 7 text
Why write a grammar in a month?
Slide 8
Slide 8 text
Why write a grammar in a month?
o Benefits to Communities
• Materials on language more available
o Benefits to Linguistics
• Learn about more languages
• [All the reasons we do language documentation]
o Benefits to Students
• Intense Training, research experience.
• CV – book publication and experience.
• Working with real materials, not ‘cleaned up’ data
• Chance to learn firsthand about ethics and language partnerships
Slide 9
Slide 9 text
Ngalia Case Study
Slide 10
Slide 10 text
The 2015 bootcamp team
Slide 11
Slide 11 text
What we did
o Worked from previously
recorded materials:
• 1000+ sentence dictionary
• texts
• short learner’s guide
o Materials collected primarily
by Sue Hanson and Kado
Muir for Ngalia.
o Other bootcamps included
language data from additional
sources
Slide 12
Slide 12 text
What we did
o Assigned topics to students [e.g. ‘how do locatives work?’]
o Regular small deadlines
o Daily meetings
o Read grammars of related languages and typological materials
Slide 13
Slide 13 text
What we did
o 2-3 day topic cycle, example:
o Monday, AM meeting:
• Assign topics
o Monday, PM meeting:
• First analysis
• Workshop with all on analysis
o Tuesday, PM meeting
• Write-up
• Feedback
• Revise for Wednesday
o Wednesday, AM :: New topic
Slide 14
Slide 14 text
What we did
o Skype sessions with speakers,
where possible
o Clarification questions with Sue
Geraldine Hogarth and Luxie Hogarth-Redmond, Kuwarra
(c) Goldfields Language Centre
Slide 15
Slide 15 text
Results
o Grammar
• 130 page [sketch] grammar
• Covers all topics you would expect in a sketch
• Lots of examples, aiming for clarity, avoiding too much terminology
• Accepted for publication by Asia Pacific Linguistics
• Reviewer comments being addressed at present
Slide 16
Slide 16 text
Results (2)
o Lots of work on the Ngalia dictionary
• spelling standardization (e.g. tj ~ ty)
• entries from words in examples
• added more examples
• added paradigm information
• edited glosses
o Materials that form the basis of other materials about the
language
• e.g. lessons
Slide 17
Slide 17 text
Quality?
o Faster ≠ Better
o Collaborative work, catching errors
o “Accessible but some gaps” is much better than “for the future”
o Time to go back and revise
o Advantages to staying focused
o More time ≠ Always Better
Slide 18
Slide 18 text
Bootcamp (interim) Conclusions
o Boot camps are not a general model of grammar writing
o But very useful for certain types of materials
o Valuable for top students who are training to become linguists
o Good for ‘long-distance’ collaboration
Slide 19
Slide 19 text
Linking Documentation and
Description
Slide 20
Slide 20 text
Himmelmann (1998)
Slide 21
Slide 21 text
No content
Slide 22
Slide 22 text
“
o Document ‘linguistic behaviors’
o Separate the primary data from the analytical
results
o Pay attention to reproducibility, accountability
o Design a multifunction corpus
o Focus on collection of data by linguists: distinction
between field data and other data
22
Slide 23
Slide 23 text
Thoughts on Documentation
Slide 24
Slide 24 text
What are we documenting?
o Mostly not working with ‘linguistic behaviors’
o Working with language communities where the
linguistic ecology has shifted.
• Young people’s varieties
• recent diglossia
• refugee or diaspora communities
• ...
o That’s ok!
o These also need documenting.
Slide 25
Slide 25 text
Data is linked to results
o The primary data are inextricably linked to the
analytical results
o Working from documentary materials themselves
o Much reclamation work, particularly from older
sources (or where primary material isn’t
available)
o H98 has rightly focused us on the importance of
getting data collection right, but beware of
ignoring valuable data sources just because they
weren’t collected under a H98 gold standard.
Slide 26
Slide 26 text
Corpus building
o Documentation >> Description isn’t my
experience
o ‘who collects it’ isn’t as important as
what’s in it
o A multifunction corpus is good for some
things, but not for others.
o Corpus best measured not by size, but as
the data that lets you take the next step
Slide 27
Slide 27 text
How a corpus is compiled
Slide 28
Slide 28 text
The documentary/descriptive spiral
▷ plan <>
▷ record exploratory
data <>
▷ analyze <>
▷ expand <>
▷ reanalyze <>
Slide 29
Slide 29 text
Conclusions
Slide 30
Slide 30 text
Conclusions
o H98 has focused us on what linguists collect.
o But what they do with it is important too.
o The grammar boot camps are one way of
• increasing available material
• training students
o Once the material is collected, that’s just the
beginning.
o When planning documentation,
Slide 31
Slide 31 text
Acknowledgments:
o Sue Hanson and Kado Muir
o 4 years of bootcampers!
• Matt Tyler, Andy Zhang, Anaí Navarro, Ryan Budnick, Sasha Wilmoth, Akshay Aitha,
Sarah Mihuc, Sarah Babinski, Kate Mooney, Omar Agha, Tom McCoy, Joshua Martin
o NSF Grant BCS-1423711 “Typology as a Window on Prehistory”
o NSF grants BCS-0844550 and 1423711
o Contributors to Chirila
• 1000+ Aboriginal people
• 100+ linguists
• 50+ student researchers