A team of four undergraduates, including him, are gathering words from native speakers and will make that corpus-with audio versions -available through a web application. So far, it has completed work on two languages-Ladakhi and Mao Naga. More are in the works.
Vikalp, originally from Chennai, speaks Telugu, Tamil, Kannada, Hindi, Urdu, Punjabi, English and Sarazi (or Saradzi)-spoken in one district in Kashmir-and possesses a "workable understanding" of Persian and Sanskrit. Initially, he was thinking large scale-"We wanted to cover South Asia."
As mentor and coordinator of BTech in Humanities, Sukrita Paul Kumar's job was to keep ambitions realistic. An editor for People's Linguistic Survey of India, she knew just how massive an undertaking this project was. The students found out soon enough-spadework alone took a semester; the questionnaire took three months. Typically, this sort of exercise would claim large chunks of funds and field visits. The team found ways around both.
Delhi's 'melting-pot' status helped. "There are speakers of 80 northeastern languages in Delhi," says Vikalp. They found some through friends and student associations. Containing over 2,600 English words (covering 30 topics) and phrases in English, the questionnaire is circulated among native speakers of a language for the closest equivalents in it.
In September 2013, members of Ladakhi and Kargil student associations participated in what Vikalp calls a "rapid vocabulary collection workshop". In about four hours, 2,500 words in Ladakhi were "collected"-"enough for a basic dictionary"-and recorded. He found speakers of Dhatki, from in Sindh in Pakistan, at the South Asian University in Delhi.
Technology allowed Vikalp to cross borders. He contacted a speaker of Khowar (from Chitral, near Swat Valley) in Islamabad through Facebook. Words are "collected" by email and recordings, by instant messenger, Whatsapp.
"When we have about five languages," says Kumar, "We can go public." She's also considering letting future batches of students pick up where the current leaves off, adding to the number of languages.
But the app isn't another online dictionary. It has songs, subtitled videos and indicates the geographical spread of a language. "There aren't equivalents for all English words. In Sarazi, there's 'here', 'there', 'yonder' and 'out-of-sight' instead of 'front', 'back' etc," explains Vikalp, "Some languages have words for 'uphill', 'downhill', 'upper-stream' and 'lower-stream', others don't." "You can see how geography influences language formation," adds Kumar.
Himanshu Patel and Vivek Shekhar worked on "geography, culture and politics" for the first semester. The 'tech' team-Himanshu and Leelambar Soren-had to teach themselves Flash from internet tutorials; help was also sought from linguistics departments within and outside DU.
In his fourth semester, Vikalp is taking a few courses in linguistics from the university department-BTech in Humanities runs in the meta-college system allowing him to pick what he likes. A bachelor's degree isn't offered in it at any college.