Compiling a corpus of South Asian online Englishes: A report, some reflections and a pilot study

Abstract

Abstract In this research article we introduce the South Asian Online Englishes (SAOnE) corpus representing four South Asian countries, i.e. Bangladesh, India, Pakistan, and Sri Lanka, and two native English-speaking countries, i.e. the UK and the USA. We have used semi-automatic and manual methods to collect data from three internet registers, i.e. newspaper comments, web forums and tweets, and a collection of internet sub-registers which we label as blogs and websites. Additionally, we have collected text messages using online freelance hiring platforms from each of the South Asian countries mentioned above. Each register category in the corpus consists of approximately 1 million words per register per country, except text messages, which contains around 500,000 words per country and only includes the four South Asian countries. We have verified the origin of website and blog links, authors of Twitter, and where possible of commenters and web forum users to make sure that only local content of each country is included. The corpus features some indigenous language content, which is tagged. In addition to the description of this dataset, we also present a pilot study analysing three discourse particles, namely na, neh, and yaar. The discourse particles na and yaar are native to Hindi/Urdu, while neh is based on a Sinhala negation marker. Our analysis indicates that na and neh have similarities in terms of their position in the clause/utterance. However, neh is confined to Sri Lanka while the Hindi/Urdu based discourse particles are also used in our Twitter data from Sri Lanka and Bangladesh. The use of these discourse particles in Bangladeshi tweets shows the influence of Indian culture through Bollywood celebrities. Of the Hindi/Urdu discourse particles yaar and na, yaar is preferred in Pakistan while na is preferred in India; additionally, yaar is used at the start of the clause more often in our Pakistani data. Lastly, we discuss the implications of the pilot study, the advantages of the type of data used for the pilot study, and future research directions.