A computational stylistic comparison between English used on Chinese governmental websites and English used on US and UK governmental websites
Jiyaue Wang
LAEL, Lancaster University
English texts on Chinese governmental websites are often criticised for being 'Chinglish' or 'lifeless'. This project investigates how English versions of Chinese governmental websites can improve their stylistic quality. The project is a computational stylistic comparison between English texts on Chinese governmental websites and English texts on UK and US governmental websites. The approach is corpus-based and employs Biber's (1988) multidimensional analysis. A corpus (including two subcorpora) of websites had previously been downloaded using the wget-m method. Perl scripts were used to extract text content from web pages to form a txt file for each website, and word frequency lists and trigrams have also been extracted. Keyword lists for the two subcorpora have been generated based on a COCA word frequency list. Several issues remain to be dealt with before further analysis can be conducted, including: whether it is possible to separate 'real content' from purely repetitive content when data comes from web pages (such as menus, navigation, copyright); the alternatives to manual annotation when this is not a practical option given the massive size of the corpus; and how to identify which features to consider to make the comparison more significant.