Our language our flower: Day 1 of ELIIP

This blogpost comes to you from Salt Lake City at the University of Utah, thanks to the Center for American Indian Languages which is co-hosting a Workshop on Endangered Languages Information and Infrastructure (ELIIP) project with Linguist List(organised by Lyle Campbell, Helen Aristar Dry, Anthony Aristar). It’s intended mostly for the specialist, but there’s an interesting push to reach out to the general public- if they don’t understand what we do, they won’t support it. Cute and less cute facts help in conveying this – more on this later.
A thousand flowers on endangered languages are blooming on the web, from Wikipedia to blogs on particular languages to the language resources catalogued by libraries. Helen Aristar Dry suggested that users want to view the whole flowerbed from a convenient vantage point. That’s the II of ELIIP: do we need a comprehensive catalogue/database/website/portal of endangered languages?
So suppose Jane LUser does a google search on the web for ‘Ossetian language’.

Top hits

• Wikipedia
• WorldLanguage: the ultimate language store
• Omniglot
• Accredited language services

Not until page 2 do I get:

• OLAC, the major US-based data harvester on languages. This links to Ethnologue, and to Linguist List, and then to OLAC’s own report, which contains information on language resources including the WALS Online Resources for Ossetic, and links back to LINGUIST List Resources for Osetin

I do NOT get in the first two pages to

• the major European language documentation resource CLARIN’s link to the TITUS Ossetian corpus
• Ethnologue run by SIL which has information on speakers, dialects, geography and is the closest thing we have to a worldwide resource.
• The World Atlas of Language Structures Online , which has lots of comments on typological features, and references to documentation on the language (but not to the TITUS corpus).
• Linguist List site by subject language, for Ossetian which gives 1 Linguist, Erschler, David, Independent University of Moscow found, and that There are materials linked to Avestan, Yagnobi, which are closely related to the language you selected. You can view this information by clicking on their name above.
• the Linguist List site by language or family, which gets various names (Osetin, Ossete, Ossetic, Ossetian), ISO code , family information and links to
  • Multitree family trees
  • a map (LL-MAP) which then links to
    • the description in Ethnologue,
    • any interlinear texts in the Odin database ( We’re sorry, no records for language code “OSS” can be found in ODIN)
    • The World Atlas of Language Structures Online
• UNESCO Interactive Atlas of the World’s Languages in Danger (which has it under Ossetic only)

So the lack of a simple portal with a high page rank is why linguistics departments get rung up by people looking for basic information. We need a portal because:

• speakers of the languages want to get stuff on them
• so do researchers
  • We don’t know which languages are endangered, what interesting typological traits they may have, what projects are underway, who works on them
• Existing data structures on languages aren’t integratable into other data structures on languages

We do need a window on the flowerbed. A similar metaphor is the ‘virtual language observatory’ – the CLARIN project’s label for their resource discovery portal. BUT both metaphors obscure other important factors – speakers live their languages rather than observe them (the languages are their flowers), and speakers and researchers want to contribute to the assembly and verification of resources (?cultivating the flowers?).
The workshop’s task then is working out the structure of the flower bed.

• What’s in it? (we have huge wishlists, we’re torn between accuracy and presenting competing hypotheses)
• How do people learn about it? (E.g. high page rank, simple interface and RSS feeds updating you when a new resource is added to languages you follow)
• Is it run as a data manager or a data harvester and aggregator?
• And whatever way, how do people contribute material? WALS has a nice feature whereby users can comment on any data-point – and they receive LOTS of comments, many of them useful.
• How is it moderated and verified?

In between working groups looking at this ideas, looking out the window at the high snow caped mountains, we heard about all sorts of interesting ideas, projects, and tools, both in the program and in breaks. Here are some that struck me.
Not all apparently healthy languages are safe as Arienne Dwyer (University of Kansas, Committee on Endangered Languages and Their Preservation (CELP)) showed. There are around 10 million speakers of Uyghur. But recent changes in government education policy aimed at increasing access to the dominant language, Chinese (Putonghua) have changed Uyghur from language of instruction to a subject language. This devalues Uyghur and is likely to lead to reduction in use. She and colleagues have been preparing instructional materials in Uyghur.
How the context of the speakers shapes the work (contra a ‘ Noah’s Ark approach to language documentation’) was discussed by Tony Woodbury with respect to Chatino speaker-linguists work on the importance of land-related kowledge and of verbal art.
Alice Harris gave a lovely paper on Exuberant exponence in Batsbi describing psycholinguistic field testing of 40 native Batsbi speakers (their average age was 67 and they constitute 20% of all speakers – it’s endangered!) to see if having lots of exponents of the same class marker helped processing and word recognition. Short answer: No.
A nice example of a web dictionary is the Dictionary of the Archi (Daghestanian) Language (sounds and pictures) organised by
Greville G. Corbett and Marina Chumakina.
Finally, access to information on languages isn’t enough – Carol Genetti and Margaret Florey are proposing a> consortium on training in language documentation and conservation. Brian Joseph gave a neat description of a capstone general education unit on Language endangerment and language death– so no knowledge of linguistics assumed.
More tomorrow.

Follow