A Corpus-Based Study Of Urdu-English Code-Switching On Pakistani Facebook Pages: Distinguishing Borrowings From Switches

This paper aims to study Urdu-English code switching in Pakistani Facebook comments by using a corpus-based study to get an idea of the types of English words used in Roman Urdu and whether they serve as established borrowings or as chosen switches. As the use of multilingual communication on social media continues to grow, Pakistanis often use both Urdu and English to talk about day-to-day issues, share opinions and communicate in different social settings. The data used in this study consists of 118 comment lines from 12 posts taken from three Pakistani Facebook pages: foodpanda, Jazz and the Pakistan Cricket Board (PCB). The corpus comprised 2,363 words and was cleaned, classified, analyzed for word frequency, and visualized with the help of Python libraries such as pandas, NLTK, and matplotlib. The English content words in mixed comments were coded into six meaning areas: technology and telecom, service and complaints, money and shopping, sport, general everyday words, and evaluation and emotion. The study also explored the difference between established borrowings, chosen switches, and words that fall between these two categories. The results show that technology and telecom words (32.0%) are the most common, followed by service and complaints (18.3%) and money and shopping (16.1%). In addition, 52.3% of the English content-word tokens were categorised as established borrowings, 39.2% as falling between borrowing and switching, and 8.5% as chosen switches. The results also indicate that technology-related words are mainly established borrowings, while English words used for evaluation and emotion are more often chosen switches. These patterns are interpreted through the Matrix Language Frame model and translanguaging, which show that English words in Urdu-English digital communication serve both naming and expressive functions. The results indicate that Urdu-English mixing on Pakistani Facebook pages is not a replacement of Urdu by English, but a flexible multilingual language practice. The findings, however, are limited to the corpus used and may not be generalized to all Facebook users in Pakistan. Keywords: Urdu-English code-switching, Roman Urdu, corpus linguistics, lexical borrowing, Matrix Language Frame model, translanguaging