START Conference Manager    

On the promise and pitfalls of repurposing existing language technologies for endangered language documentation

Emily Prud'hommeaux, Robert Jimerson, Richard Hatcher, Raymond Ptucha and Karin Michelson


Categories

category:  Poster
Session:  6 December Session P6: North American Languages Poster Session

Additional Fields

 
Abstract:   Like many indigenous languages of North America, the Iroquoian language Seneca is endangered, with fewer than fifty living native speakers. Although descriptive grammars of Seneca exist, there are few texts and recordings available to support immersion programs and other revitalization efforts. In a collaboration between university researchers and Seneca community members, we are working to produce textual and audio documentation of the Seneca language using both existing toolkits and custom architectures. We find that while some toolkits yield promising results, the morphological complexity of Seneca and the variable quality of the available recordings present challenges for deploying one-size-fits-all solutions.

 
Resume:   Dah ne:' dih tša'deyo'dëh oya' yei' niyöëdza'geh niodiwënö'dës koh neh onödowa’ga:’ gawënö’ agaiwahdö't so'jih gao' wis niwashëh niënödih ahsöh deodišnye'öh onödowa’ga:’ gawënö’. Gwaheh ha'deyöh gayadö' niyo'dëh onödowá’ga:’ gawënö’ koh neh dohga:’ah niyoh gawënöhdas ogwenyöh aonödesdë' adiyë'he't onödowá’ga:’ gawënö’. Dah ne:' hae'gwah dogëh dwadade'gë:' tënödeyësdahgwa’geh hënöjëönyanih koh neh onödowa’geonö' deodiyenö'. Dah ne:' hae'gwah hodihšöniaje' gayadöshäse:' watšowih na'ot gawënöhdas gayë' hadiyä'ta' dejaöh yesta' koh neh gadogëh neh nigayeëh gahšönih. Dah ne:' wa'agwaiwaho' neh dohga:’ah niyoh yesta' agwas wadesta' gwaheh so'jih ha'deyöh gayë' sgawënot koh neh dewenö:' niyo'dëh gawënohdas da'aöh ahsehšöni' gwisdë' neh ogwenyöh agaya'daei' gagwegöh koh neh agaiwaeis.

File(s)

[Paper (PDF)]  

START Conference Manager (V2.61.0 - Rev. 5964)