Jean Okimâsis, Falene Karey-McKenna and Katie Schmirler hard at work on itwêwina.
Members of the Cree Literacy Network including Dorothy Thunder, Arok Wolvengrey, Arden Ogg and Jean Okimâsis, have been proud to collaborate with Antti Arppe and his computational linguistics team at the University of Alberta since 2013 in the SSHRCC-funded project, 21st century Tools for Indigenous Languages. At the third annual Computational Workshop for Indigenous Languages (CWIL) meeting in Edmonton this May, it was exciting to see concrete progress. (The web page associated with the project can be found here: http://altlab.artsrn.ualberta.ca/)
One portion of the overall project is reflected in the Ā icon, now proudly installed on the left edge of the Cree Literacy Network’s blog page (with thanks to CWIL collaborator Ryan Johnson for installation help). It reflects a Plains Cree reading-assistance tool called itwêwina (Plains Cree for ‘words’). The tool is still very much under development: If you give it a try, you can legitimately boast that you’re contributing to its alpha testing.
The tool is designed to allow readers to click on Cree words and get an immediate reading of that word – a dictionary entry and analysis. You can find a walk-through via screen shots at http://altlab.artsrn.ualberta.ca/wp-content/uploads/2015/08/crk_itwewina_1508091.pdf or you can try it for yourself right here, by alt-clicking on Cree words (in text – not in pdf or image files) to try it yourself. Words like itwêwina and nêhiyaw. (If you’d like to try a bigger sample, Solomon Ratt’s stories in y-dialect are good examples. The th-dialect is not as well represented just yet in the dictionary.)
In addition, you can visit the itwêwina home page at http://itwewina.oahpa.no/ and search for Cree words using the search box. One of the coolest things about this tool is that it will (eventually) show you *all* of the forms for a given noun or verb.
The nêhiyaw basis of this tool is the Wolvengrey Dictionary, supported by the morphological work of Jean Okimâsis, and foundational work by H.C. Wolfart and the late Freda Ahenakew. The computational basis is the work of http://giellatekno.uit.no/ with a significant foundation of comparable tools for Saami. Project collaborators include ALTLab at the University of Alberta, First Nations University of Canada, and Giellatekno at UiT: The Arctic University of Norway.
Like all new technology, it’s wonderful when it works. And it works wonderfully on words that are correctly spelled in SRO – and that are in the existing (but growing) Plains Cree (y-dialect) dictionary that runs in the background. It also allows users to click through to additional dictionary and verb paradigm data so they can explore a given word in depth.
When it doesn’t work (and it is still in testing, so there will continue to be gaps for some time), the errors you see are being automatically recorded and reported to a log file that will contribute to making the tool work better in time. So please consider this your invitation! Click around and give it a workout!