The NCCU Corpus of Spoken Chinese: Mandarin, Hakka, and Southern Min

Abstract

In Taiwan, most people speak Mandarin, Southern Min, or Hakka. Not only are the three Chinese dialects undergoing linguistic changes, but the population of Southern Min and Hakka is also diminishing. The NCCU Corpus of Spoken Chinese is thus a project of language documentation whereby open online access to Mandarin, Hakka, and Southern Min data is provided for non-profit-making research. As a language documentation project, the NCCU spoken corpus focuses on collecting and archiving spoken forms of various types. It consists of three sub-corpora, namely the Corpus of Spoken Mandarin, the Corpus of Spoken Hakka, and the Corpus of Spoken Southern Min. The three corpora share a common scheme for the collection of spoken data, mostly in the form of spontaneous face-to-face conversations. The infrastructure of the corpus is designed in a simple yet user-friendly way, so that data can be processed efficiently in the database, and users can browse the spoken data directly from the web. We hope that our work can encourage more people to engage in building up spoken corpora from different perspectives and for different purposes.