New Ways of Analyzing Variation Asia-Pacific Region
2016.4.22-24, in Chiayi, Taiwan
How fast would it be? -Observing the distributions of emerging words through Twitter
Kota Hattori¹, Shinsuke Kishie², Takashi Kirimura³, Yukako Sakoguchi⁴, & Nanami Shiokawa⁵
Tokushima University¹²⁴⁵, University of Tokyo³
The present study examined whether we could observe how newly emerged words from a local dialect spread using Twitter. Specifically, we observed how koyan, which means ‘not come’ , has spread so far in Japan. We collected data between November, 2012 and December, 2015.
We cleaned up our data with R and manual checks, leaving approximately 47,000 tweets. The results demonstrated that the word has been heavily used in Osaka area, where the word
emerged. The results also demonstrated that the word has been spreading to adjacent areas as well as distant areas.
Conducting Research on the Geographical Linguistics by Utilizing the Data Comprising Twitter Postings
Shinsuke Kishie¹, Shuichi Matsunaga², Takashi Kirimura³, Shin Abe⁴, Kota Hattori⁵, Yukako Sakoguchi⁶
TokushimaUniversity¹⁵⁶,Jumonji University²,The University of Tokyo³,Nagoya University of Foreign Studies⁴
Social Networking Sites (SNS) have become a common tool for our daily communication. They have also become a powerful tool for researchers to investigate various issues in language variation. (e.g., Mocanuetal., 2012; GonçalvesandSánchez, 2014)
Geo-tagged SNS data may not reflect dialectal variations within a language well, because the SNS data’s mobility may lead to mismatch between the distribution of the geographical data and the location of where a speaker’s dialect is spoken.
In this study, we focused on a specific word, ‘driving school’, whose dialectal variations are already known by former survey data (Fig.1), and we examined whether dialectal variations in Twitter data correspond with our survey data.